FLOW-3D POST 2023R2는 EXODUS II 형식을 기반으로 하는 완전히 새로운 결과 파일 형식을 도입하여 더 빠른 후처리를 가능하게 합니다. 이 새로운 파일 형식은 크고 복잡한 시뮬레이션의 후처리 작업에 소요되는 시간을 크게 줄이는 동시에(평균 최대 5배!) 다른 시각화 도구와의 연결성을 향상시킵니다.
FLOW-3D POST 2023R2 에서 사용자는 이제 선택한 데이터를 flsgrf , EXODUS II 또는 flsgrf 및 EXODUS II 파일 형식으로 쓸 수 있습니다 . 새로운 EXODUS II 파일 형식은 각 객체에 대해 유한 요소 메쉬를 활용하므로 사용자는 다른 호환 가능한 포스트 프로세서 및 FEA 코드를 사용 하여 FLOW-3D 결과를 열 수도 있습니다. 새로운 워크플로우를 통해 사용자는 크고 복잡한 사례를 신속하게 시각화하고 임의 슬라이싱, 볼륨 렌더링 및 통계를 사용하여 보조 정보를 추출할 수 있습니다.
FLOW-3D POST 의 새로운 EXODUS II 파일 형식으로 채워진 화장품 크림 모델의 향상된 광선 추적 기능의 예
새로운 결과 파일 형식은 솔버 엔진의 성능을 저하시키지 않으면서 flsgrf 에 비해 시각화 작업 흐름에서 놀라운 속도 향상을 자랑합니다. 이 흥미로운 새로운 개발은 결과 분석의 속도와 유연성이 향상되어 원활한 시뮬레이션 경험을 제공합니다.
FLOW-3D 2023R2는 2방정식(RANS) 난류 모델에 대한 동적 혼합 길이 계산을 크게 개선했습니다. 거의 층류 흐름 체계와 같은 특정 제한 사례에서는 이전 버전의 코드 계산 제한기가 때때로 과도하게 예측되어 사용자가 특정 혼합 길이를 수동으로 입력해야 할 수 있습니다.
새로운 동적 혼합 길이 계산은 이러한 상황에서 난류 길이와 시간 척도를 더 잘 설명합니다. 이제 사용자는 고정된(물리 기반) 혼합 길이를 설정하는 대신 더 넓은 범위의 흐름에 동적 모델을 적용할 수 있습니다.
적절한 고정 혼합 길이와 비교하여 접촉 탱크의 혼합 시뮬레이션을 위한 기존 동적 혼합 길이 모델과 새로운 동적 혼합 길이 모델 간의 비교
정수압 초기화
사용자가 미리 정의된 유체 영역에서 정수압을 초기화해야 하는 경우가 많습니다. 이전에는 대규모의 복잡한 시뮬레이션에서 정수압 솔버의 수렴 속도가 느려지는 경우가 있었습니다. FLOW-3D 2023R2는 정수압 솔버의 성능을 크게 향상시켜 전처리 단계에서 최대 6배 빠르게 수렴할 수 있도록 해줍니다.
압축성 흐름 솔버 성능
FLOW-3D 2023R2는 최적화된 압력 솔버를 도입하여 압축성 흐름 문제에 대해 상당한 성능 향상을 제공합니다. 압축성 제트 흐름의 이 예에서 2023R2 솔버는 2023R1 버전보다 ~4배 빠릅니다.
FLOW-3D 의 압축성 제트 시뮬레이션의 예
FLOW-3D 2023R2 의 새로운 기능
FLOW-3D 소프트웨어 제품군의 모든 제품은 2023R2에서 IT 관련 개선 사항을 받았습니다. FLOW-3D 2023R2은 이제 Windows 11 및 RHEL 8을 지원합니다. Linux 설치 프로그램은 누락된 종속성을 보고하도록 개선되었으며 더 이상 루트 수준 권한이 필요하지 않으므로 설치가 더 쉽고 안전해집니다. 그리고 워크플로를 자동화한 분들을 위해 입력 파일 변환기에 명령줄 인터페이스를 추가하여 스크립트 환경에서도 워크플로가 업데이트된 입력 파일로 작동하는지 확인할 수 있습니다.
확장된 PQ 2 분석
제조에 사용되는 유압 시스템은 PQ 2 곡선을 사용하여 모델링할 수 있습니다. 장치의 세부 사항을 건너뛰고 흐름에 미치는 영향을 포함하기 위해 질량 운동량 소스 또는 속도 경계 조건을 사용하여 유압 시스템을 근사화하는 것이 편리한 단순화인 경우가 많습니다. 우리는 기존 PQ 2 분석 모델을 확장하여 이러한 유형의 기하학적 단순화를 허용하면서도 현실적인 결과를 제공했습니다. 이는 시뮬레이션 시간과 모델 복잡성의 감소를 의미합니다.
FLOW-3D 2022R2 제품군 출시로 Flow Science는 FLOW-3D 의 워크스테이션과 HPC 버전을 통합하여 단일 노드 CPU 구성에서 다중 노드에 이르기까지 모든 유형의 하드웨어 아키텍처를 활용할 수 있는 단일 솔버 엔진을 제공합니다. 노드 병렬 고성능 컴퓨팅 실행. 추가 개발에는 점탄성 흐름을 위한 새로운 로그 형태 텐서 방법, 지속적인 솔버 속도 성능 개선, 고급 냉각 채널 및 팬텀 구성요소 제어, 개선된 동반 공기 기능이 포함됩니다.
통합 솔버
우리는 FLOW-3D 제품을 단일 통합 솔버로 마이그레이션하여 로컬 워크스테이션이나 고성능 컴퓨팅 하드웨어 환경에서 원활하게 실행했습니다.
많은 사용자가 노트북이나 로컬 워크스테이션에서 모델을 실행하지만, 고성능 컴퓨팅 클러스터에서도 더 큰 모델을 실행합니다. 2022R2 릴리스에서는 통합 솔버를 통해 사용자가 HPC 솔루션의 OpenMP/MPI 하이브리드 병렬화와 동일한 이점을 활용하여 워크스테이션과 노트북에서 실행할 수 있습니다.
증가하는 CPU 코어 수를 사용한 성능 확장의 예OpenMP/MPI 하이브리드 병렬화를 위한 메시 분해의 예
솔버 성능 개선
멀티 소켓 워크스테이션
다중 소켓 워크스테이션은 이제 매우 일반적이며 대규모 시뮬레이션을 실행할 수 있습니다. 새로운 통합 솔버를 사용하면 이러한 유형의 하드웨어를 사용하는 사용자는 일반적으로 HPC 클러스터 구성에서만 사용할 수 있었던 OpenMP/MPI 하이브리드 병렬화를 활용하여 모델을 실행할 수 있어 성능이 향상되는 것을 확인할 수 있습니다.
낮은 수준의 루틴으로 향상된 벡터화 및 메모리 액세스
대부분의 테스트 사례에서 10~20% 정도의 성능 향상이 관찰되었으며 일부 사례에서는 20%를 초과하는 런타임 이점이 나타났습니다.
정제된 체적 대류 안정성 한계
시간 단계 안정성 제한은 모델 런타임의 주요 동인이며, 2022R2에서는 새로운 시간 단계 안정성 제한인 3D 대류 안정성 제한을 숫자 위젯에서 사용할 수 있습니다. 실행 중이고 대류가 제한된(cx, cy 또는 cz 제한) 모델의 경우 새 옵션은 일반적인 속도 향상을 30% 정도 보여줍니다.
압력 솔버 프리컨디셔너
경우에 따라 까다로운 흐름 구성의 경우 과도한 압력 솔버 반복으로 인해 실행 시간이 길어질 수 있습니다. 이러한 어려운 경우 2022R2에서는 모델이 너무 많이 반복되면 FLOW-3D가 자동으로 새로운 사전 조절기를 활성화하여 압력 수렴을 돕습니다. 테스트의 런타임이 1.9~335배 더 빨라졌습니다!
점탄성 유체에 대한 로그 형태 텐서 방법
점탄성 유체에 대한 새로운 솔버 옵션을 사용자가 사용할 수 있으며 특히 높은 Weissenberg 수에 효과적입니다.
로그 구조 텐서 솔루션을 사용하여 점탄성 흐름에 대한 높은 Weissenberg 수의 개선된 솔루션의 예입니다. 제공: MF Tome 외, J. Non-Newton. 체액. 기계화. 175-176 (2012) 44–54
활성 시뮬레이션 제어 확장
능동 시뮬레이션 제어 기능이 확장되어 연속 주조 및 적층 제조 응용 분야에 일반적으로 사용되는 팬텀 개체는 물론 주조 및 기타 여러 열 관리 응용 분야에 사용되는 냉각 채널에도 사용됩니다.
연속 주조 응용 분야에 대한 가상 물체 속도 제어의 예융합 증착 모델링 애플리케이션을 위한 동적 열 제어의 예산업용 탱크 적용을 위한 동적 냉각 채널 제어의 예
향상된 공기 동반 기능
디퓨저 및 이와 유사한 산업용 기포 흐름 응용 분야의 경우 이제 질량 공급원을 사용하여 물기둥에 공기를 유입할 수 있습니다. 또한, 동반 공기 및 용존 산소의 난류 확산에 대한 기본값이 업데이트되었으며 매우 낮은 공기 농도에 대한 모델 정확도가 향상되었습니다.
금속 적층 제조 중 고체 상 변형 예측: Inconel-738의 전자빔 분말층 융합에 대한 사례 연구
Nana Kwabena Adomako a, Nima Haghdadi a, James F.L. Dingle bc, Ernst Kozeschnik d, Xiaozhou Liao bc, Simon P. Ringer bc, Sophie Primig a
Abstract
Metal additive manufacturing (AM) has now become the perhaps most desirable technique for producing complex shaped engineering parts. However, to truly take advantage of its capabilities, advanced control of AM microstructures and properties is required, and this is often enabled via modeling. The current work presents a computational modeling approach to studying the solid-state phase transformation kinetics and the microstructural evolution during AM. Our approach combines thermal and thermo-kinetic modelling. A semi-analytical heat transfer model is employed to simulate the thermal history throughout AM builds. Thermal profiles of individual layers are then used as input for the MatCalc thermo-kinetic software. The microstructural evolution (e.g., fractions, morphology, and composition of individual phases) for any region of interest throughout the build is predicted by MatCalc. The simulation is applied to an IN738 part produced by electron beam powder bed fusion to provide insights into how γ′ precipitates evolve during thermal cycling. Our simulations show qualitative agreement with our experimental results in predicting the size distribution of γ′ along the build height, its multimodal size character, as well as the volume fraction of MC carbides. Our findings indicate that our method is suitable for a range of AM processes and alloys, to predict and engineer their microstructures and properties.
Additive manufacturing (AM) is an advanced manufacturing method that enables engineering parts with intricate shapes to be fabricated with high efficiency and minimal materials waste. AM involves building up 3D components layer-by-layer from feedstocks such as powder [1]. Various alloys, including steel, Ti, Al, and Ni-based superalloys, have been produced using different AM techniques. These techniques include directed energy deposition (DED), electron- and laser powder bed fusion (E-PBF and L-PBF), and have found applications in a variety of industries such as aerospace and power generation[2], [3], [4]. Despite the growing interest, certain challenges limit broader applications of AM fabricated components in these industries and others. One of such limitations is obtaining a suitable and reproducible microstructure that offers the desired mechanical properties consistently. In fact, the AM as-built microstructure is highly complex and considerably distinctive from its conventionally processed counterparts owing to the complicated thermal cycles arising from the deposition of several layers upon each other [5], [6].
Several studies have reported that the solid-state phases and solidification microstructure of AM processed alloys such as CMSX-4, CoCr [7], [8], Ti-6Al-4V [9], [10], [11], IN738[6], 304L stainless steel[12], and IN718 [13], [14] exhibit considerable variations along the build direction. For instance, references [9], [10] have reported that there is a variation in the distribution of α and β phases along the build direction in Ti-alloys. Similarly, the microstructure of an L-PBF fabricated martensitic steel exhibits variations in the fraction of martensite [15]. Furthermore, some of the present authors and others [6], [16], [17], [18], [19], [20] have recently reviewed and reported that there is a difference in the morphology and fraction of nanoscale precipitates as a function of build height in Ni-based superalloys. These non-uniformities in the as-built microstructure result in an undesired heterogeneity in mechanical and other important properties such as corrosion and oxidation[19], [21], [22], [23]. To obtain the desired microstructure and properties, additional processing treatments are utilized, but this incurs extra costs and may lead to precipitation of detrimental phases and grain coarsening. Therefore, a through-process understanding of the microstructure evolution under repeated heating and cooling is now needed to further advance 3D printed microstructure and property control.
It is now commonly understood that the microstructure evolution during printing is complex, and most AM studies concentrate on the microstructure and mechanical properties of the final build only. Post-printing studies of microstructure characteristics at room temperature miss crucial information on how they evolve. In-situ measurements and modelling approaches are required to better understand the complex microstructural evolution under repeated heating and cooling. Most in-situ measurements in AM focus on monitoring the microstructural changes, such as phase transformations and melt pool dynamics during fabrication using X-ray scattering and high-speed X-ray imaging [24], [25], [26], [27]. For example, Zhao et al. [25] measured the rate of solidification and described the α/β phase transformation during L-PBF of Ti-6Al-4V in-situ. Also, Wahlmann et al. [21] recently used an L-PBF machine coupled with X-ray scattering to investigate the changes in CMSX-4 phase during successive melting processes. Although these techniques provide significant understanding of the basic principles of AM, they are not widely accessible. This is due to the great cost of the instrument, competitive application process, and complexities in terms of the experimental set-up, data collection, and analysis [26], [28].
Computational modeling techniques are promising and more widely accessible tools that enable advanced understanding, prediction, and engineering of microstructures and properties during AM. So far, the majority of computational studies have concentrated on physics based process models for metal AM, with the goal of predicting the temperature profile, heat transfer, powder dynamics, and defect formation (e.g., porosity) [29], [30]. In recent times, there have been efforts in modeling of the AM microstructure evolution using approaches such as phase-field [31], Monte Carlo (MC) [32], and cellular automata (CA) [33], coupled with finite element simulations for temperature profiles. However, these techniques are often restricted to simulating the evolution of solidification microstructures (e.g., grain and dendrite structure) and defects (e.g., porosity). For example, Zinovieva et al. [33] predicted the grain structure of L-PBF Ti-6Al-4V using finite difference and cellular automata methods. However, studies on the computational modelling of the solid-state phase transformations, which largely determine the resulting properties, remain limited. This can be attributed to the multi-component and multi-phase nature of most engineering alloys in AM, along with the complex transformation kinetics during thermal cycling. This kind of research involves predictions of the thermal cycle in AM builds, and connecting it to essential thermodynamic and kinetic data as inputs for the model. Based on the information provided, the thermokinetic model predicts the history of solid-state phase microstructure evolution during deposition as output. For example, a multi-phase, multi-component mean-field model has been developed to simulate the intermetallic precipitation kinetics in IN718 [34] and IN625 [35] during AM. Also, Basoalto et al. [36] employed a computational framework to examine the contrasting distributions of process-induced microvoids and precipitates in two Ni-based superalloys, namely IN718 and CM247LC. Furthermore, McNamara et al. [37] established a computational model based on the Johnson-Mehl-Avrami model for non-isothermal conditions to predict solid-state phase transformation kinetics in L-PBF IN718 and DED Ti-6Al-4V. These models successfully predicted the size and volume fraction of individual phases and captured the repeated nucleation and dissolution of precipitates that occur during AM.
In the current study, we propose a modeling approach with appreciably short computational time to investigate the detailed microstructural evolution during metal AM. This may include obtaining more detailed information on the morphologies of phases, such as size distribution, phase fraction, dissolution and nucleation kinetics, as well as chemistry during thermal cycling and final cooling to room temperature. We utilize the combination of the MatCalc thermo-kinetic simulator and a semi-analytical heat conduction model. MatCalc is a software suite for simulation of phase transformations, microstructure evolution and certain mechanical properties in engineering alloys. It has successfully been employed to simulate solid-state phase transformations in Ni-based superalloys [38], [39], steels [40], and Al alloys[41] during complex thermo-mechanical processes. MatCalc uses the classical nucleation theory as well as the so-called Svoboda-Fischer-Fratzl-Kozeschnik (SFFK) growth model as the basis for simulating precipitation kinetics [42]. Although MatCalc was originally developed for conventional thermo-mechanical processes, we will show that it is also applicable for AM if the detailed time-temperature profile of the AM build is known. The semi-analytical heat transfer code developed by Stump and Plotkowski [43] is used to simulate these profile throughout the AM build.
1.1. Application to IN738
Inconel-738 (IN738) is a precipitation hardening Ni-based superalloy mainly employed in high-temperature components, e.g. in gas turbines and aero-engines owing to its exceptional mechanical properties at temperatures up to 980 °C, coupled with high resistance to oxidation and corrosion [44]. Its superior high-temperature strength (∼1090 MPa tensile strength) is provided by the L12 ordered Ni3(Al,Ti) γ′ phase that precipitates in a face-centered cubic (FCC) γ matrix [45], [46]. Despite offering great properties, IN738, like most superalloys with high γ′ fractions, is challenging to process owing to its propensity to hot cracking [47], [48]. Further, machining of such alloys is challenging because of their high strength and work-hardening rates. It is therefore difficult to fabricate complex INC738 parts using traditional manufacturing techniques like casting, welding, and forging.
The emergence of AM has now made it possible to fabricate such parts from IN738 and other superalloys. Some of the current authors’ recent research successfully applied E-PBF to fabricate defect-free IN738 containing γ′ throughout the build [16], [17]. The precipitated γ′ were heterogeneously distributed. In particular, Haghdadi et al. [16] studied the origin of the multimodal size distribution of γ′, while Lim et al. [17] investigated the gradient in γ′ character with build height and its correlation to mechanical properties. Based on these results, the present study aims to extend the understanding of the complex and site-specific microstructural evolution in E-PBF IN738 by using a computational modelling approach. New experimental evidence (e.g., micrographs not published previously) is presented here to support the computational results.
2. Materials and Methods
2.1. Materials preparation
IN738 Ni-based superalloy (59.61Ni-8.48Co-7.00Al-17.47Cr-3.96Ti-1.01Mo-0.81W-0.56Ta-0.49Nb-0.47C-0.09Zr-0.05B, at%) gas-atomized powder was used as feedstock. The powders, with average size of 60 ± 7 µm, were manufactured by Praxair and distributed by Astro Alloys Inc. An Arcam Q10 machine by GE Additive with an acceleration voltage of 60 kV was used to fabricate a 15 × 15 × 25 mm3 block (XYZ, Z: build direction) on a 316 stainless steel substrate. The block was 3D-printed using a ‘random’ spot melt pattern. The random spot melt pattern involves randomly selecting points in any given layer, with an equal chance of each point being melted. Each spot melt experienced a dwell time of 0.3 ms, and the layer thickness was 50 µm. Some of the current authors have previously characterized the microstructure of the very same and similar builds in more detail [16], [17]. A preheat temperature of ∼1000 °C was set and kept during printing to reduce temperature gradients and, in turn, thermal stresses [49], [50], [51]. Following printing, the build was separated from the substrate through electrical discharge machining. It should be noted that this sample was simultaneously printed with the one used in [17] during the same build process and on the same build plate, under identical conditions.
2.2. Microstructural characterization
The printed sample was longitudinally cut in the direction of the build using a Struers Accutom-50, ground, and then polished to 0.25 µm suspension via standard techniques. The polished x-z surface was electropolished and etched using Struers A2 solution (perchloric acid in ethanol). Specimens for image analysis were polished using a 0.06 µm colloidal silica. Microstructure analyses were carried out across the height of the build using optical microscopy (OM) and scanning electron microscopy (SEM) with focus on the microstructure evolution (γ′ precipitates) in individual layers. The position of each layer being analyzed was determined by multiplying the layer number by the layer thickness (50 µm). It should be noted that the position of the first layer starts where the thermal profile is tracked (in this case, 2 mm from the bottom). SEM images were acquired using a JEOL 7001 field emission microscope. The brightness and contrast settings, acceleration voltage of 15 kV, working distance of 10 mm, and other SEM imaging parameters were all held constant for analysis of the entire build. The ImageJ software was used for automated image analysis to determine the phase fraction and size of γ′ precipitates and carbides. A 2-pixel radius Gaussian blur, following a greyscale thresholding and watershed segmentation was used [52]. Primary γ′ sizes (>50 nm), were measured using equivalent spherical diameters. The phase fractions were considered equal to the measured area fraction. Secondary γ′ particles (<50 nm) were not considered here. The γ′ size in the following refers to the diameter of a precipitate.
2.3. Hardness testing
A Struers DuraScan tester was utilized for Vickers hardness mapping on a polished x-z surface, from top to bottom under a maximum load of 100 mN and 10 s dwell time. 30 micro-indentations were performed per row. According to the ASTM standard [53], the indentations were sufficiently distant (∼500 µm) to assure that strain-hardened areas did not interfere with one another.
2.4. Computational simulation of E-PBF IN738 build
2.4.1. Thermal profile modeling
The thermal history was generated using the semi-analytical heat transfer code (also known as the 3DThesis code) developed by Stump and Plotkowski [43]. This code is an open-source C++ program which provides a way to quickly simulate the conductive heat transfer found in welding and AM. The key use case for the code is the simulation of larger domains than is practicable with Computational Fluid Dynamics/Finite Element Analysis programs like FLOW-3D AM. Although simulating conductive heat transfer will not be an appropriate simplification for some investigations (for example the modelling of keyholding or pore formation), the 3DThesis code does provide fast estimates of temperature, thermal gradient, and solidification rate which can be useful for elucidating microstructure formation across entire layers of an AM build. The mathematics involved in the code is as follows:
In transient thermal conduction during welding and AM, with uniform and constant thermophysical properties and without considering fluid convection and latent heat effects, energy conservation can be expressed as:(1)��∂�∂�=�∇2�+�̇where � is density, � specific heat, � temperature, � time, � thermal conductivity, and �̇ a volumetric heat source. By assuming a semi-infinite domain, Eq. 1 can be analytically solved. The solution for temperature at a given time (t) using a volumetric Gaussian heat source is presented as:(2)��,�,�,�−�0=33�����32∫0�1������exp−3�′�′2��+�′�′2��+�′�′2����′(3)and��=12��−�′+��2for�=�,�,�(4)and�′�′=�−���′Where � is the vector �,�,� and �� is the location of the heat source.
The numerical integration scheme used is an adaptive Gaussian quadrature method based on the following nondimensionalization:(5)�=��xy2�,�′=��xy2�′,�=��xy,�=��xy,�=��xy,�=���xy
A more detailed explanation of the mathematics can be found in reference [43].
The main source of the thermal cycling present within a powder-bed fusion process is the fusion of subsequent layers. Therefore, regions near the top of a build are expected to undergo fewer thermal cycles than those closer to the bottom. For this purpose, data from the single scan’s thermal influence on multiple layers was spliced to represent the thermal cycles experienced at a single location caused by multiple subsequent layers being fused.
The cross-sectional area simulated by this model was kept constant at 1 × 1 mm2, and the depth was dependent on the build location modelled with MatCalc. For a build location 2 mm from the bottom, the maximum number of layers to simulate is 460. Fig. 1a shows a stitched overview OM image of the entire build indicating the region where this thermal cycle is simulated and tracked. To increase similarity with the conditions of the physical build, each thermal history was constructed from the results of two simulations generated with different versions of a random scan path. The parameters used for these thermal simulations can be found in Table 1. It should be noted that the main purpose of the thermal profile modelling was to demonstrate how the conditions at different locations of the build change relative to each other. Accurately predicting the absolute temperature during the build would require validation via a temperature sensor measurement during the build process which is beyond the scope of the study. Nonetheless, to establish the viability of the heat source as a suitable approximation for this study, an additional sensitivity analysis was conducted. This analysis focused on the influence of energy input on γ′ precipitation behavior, the central aim of this paper. This was achieved by employing varying beam absorption energies (0.76, 0.82 – the values utilized in the simulation, and 0.9). The direct impact of beam absorption efficiency on energy input into the material was investigated. Specifically, the initial 20 layers of the build were simulated and subsequently compared to experimental data derived from SEM. While phase fractions were found to be consistent across all conditions, disparities emerged in the mean size of γ′ precipitates. An absorption efficiency of 0.76 yielded a mean size of approximately 70 nm. Conversely, absorption efficiencies of 0.82 and 0.9 exhibited remarkably similar mean sizes of around 130 nm, aligning closely with the outcomes of the experiments.
The numerical analyses of the evolution of precipitates was performed using MatCalc version 6.04 (rel 0.011). The thermodynamic (‘mc_ni.tdb’, version 2.034) and diffusion (‘mc_ni.ddb’, version 2.007) databases were used. MatCalc’s basic principles are elaborated as follows:
The nucleation kinetics of precipitates are computed using a computational technique based on a classical nucleation theory[54] that has been modified for systems with multiple components [42], [55]. Accordingly, the transient nucleation rate (�), which expresses the rate at which nuclei are formed per unit volume and time, is calculated as:(6)�=�0��*∙�xp−�*�∙�∙exp−��where �0 denotes the number of active nucleation sites, �* the rate of atomic attachment, � the Boltzmann constant, � the temperature, �* the critical energy for nucleus formation, τ the incubation time, and t the time. � (Zeldovich factor) takes into consideration that thermal excitation destabilizes the nucleus as opposed to its inactive state [54]. Z is defined as follows:(7)�=−12�kT∂2∆�∂�2�*12where ∆� is the overall change in free energy due to the formation of a nucleus and n is the nucleus’ number of atoms. ∆�’s derivative is evaluated at n* (critical nucleus size). �* accounts for the long-range diffusion of atoms required for nucleation, provided that the matrix’ and precipitates’ composition differ. Svoboda et al. [42] developed an appropriate multi-component equation for �*, which is given by:(8)�*=4��*2�4�∑�=1��ki−�0�2�0��0�−1where �* denotes the critical radius for nucleation, � represents atomic distance, and � is the molar volume. �ki and �0� represent the concentration of elements in the precipitate and matrix, respectively. The parameter �0� denotes the rate of diffusion of the ith element within the matrix. The expression for the incubation time � is expressed as [54]:(9)�=12�*�2
and �*, which represents the critical energy for nucleation:(10)�*=16�3�3∆�vol2where � is the interfacial energy, and ∆Gvol the change in the volume free energy. The critical nucleus’ composition is similar to the γ′ phase’s equilibrium composition at the same temperature. � is computed based on the precipitate and matrix compositions, using a generalized nearest neighbor broken bond model, with the assumption of interfaces being planar, sharp, and coherent [56], [57], [58].
In Eq. 7, it is worth noting that �* represents the fundamental variable in the nucleation theory. It contains �3/∆�vol2 and is in the exponent of the nucleation rate. Therefore, even small variations in γ and/or ∆�vol can result in notable changes in �, especially if �* is in the order of �∙�. This is demonstrated in [38] for UDIMET 720 Li during continuous cooling, where these quantities change steadily during precipitation due to their dependence on matrix’ and precipitate’s temperature and composition. In the current work, these changes will be even more significant as the system is exposed to multiple cycles of rapid cooling and heating.
Once nucleated, the growth of a precipitate is assessed using the radius and composition evolution equations developed by Svoboda et al. [42] with a mean-field method that employs the thermodynamic extremal principle. The expression for the total Gibbs free energy of a thermodynamic system G, which consists of n components and m precipitates, is given as follows:(11)�=∑���0��0�+∑�=1�4���33��+∑�=1��ki�ki+∑�=1�4���2��.
The chemical potential of component � in the matrix is denoted as �0�(�=1,…,�), while the chemical potential of component � in the precipitate is represented by �ki(�=1,…,�,�=1,…,�). These chemical potentials are defined as functions of the concentrations �ki(�=1,…,�,�=1,…,�). The interface energy density is denoted as �, and �� incorporates the effects of elastic energy and plastic work resulting from the volume change of each precipitate.
Eq. (12) establishes that the total free energy of the system in its current state relies on the independent state variables: the sizes (radii) of the precipitates �� and the concentrations of each component �ki. The remaining variables can be determined by applying the law of mass conservation to each component �. This can be represented by the equation:(12)��=�0�+∑�=1�4���33�ki,
Furthermore, the global mass conservation can be expressed by equation:(13)�=∑�=1���When a thermodynamic system transitions to a more stable state, the energy difference between the initial and final stages is dissipated. This model considers three distinct forms of dissipation effects [42]. These include dissipations caused by the movement of interfaces, diffusion within the precipitate and diffusion within the matrix.
Consequently, �̇� (growth rate) and �̇ki (chemical composition’s rate of change) of the precipitate with index � are derived from the linear system of equation system:(14)�ij��=��where �� symbolizes the rates �̇� and �̇ki [42]. Index i contains variables for precipitate radius, chemical composition, and stoichiometric boundary conditions suggested by the precipitate’s crystal structure. Eq. (10) is computed separately for every precipitate �. For a more detailed description of the formulae for the coefficients �ij and �� employed in this work please refer to [59].
The MatCalc software was used to perform the numerical time integration of �̇� and �̇ki of precipitates based on the classical numerical method by Kampmann and Wagner [60]. Detailed information on this method can be found in [61]. Using this computational method, calculations for E-PBF thermal cycles (cyclic heating and cooling) were computed and compared to experimental data. The simulation took approximately 2–4 hrs to complete on a standard laptop.
3. Results
3.1. Microstructure
Fig. 1 displays a stitched overview image and selected SEM micrographs of various γ′ morphologies and carbides after observations of the X-Z surface of the build from the top to 2 mm above the bottom. Fig. 2 depicts a graph that charts the average size and phase fraction of the primary γ′, as it changes with distance from the top to the bottom of the build. The SEM micrographs show widespread primary γ′ precipitation throughout the entire build, with the size increasing in the top to bottom direction. Particularly, at the topmost height, representing the 460th layer (Z = 22.95 mm), as seen in Fig. 1b, the average size of γ′ is 110 ± 4 nm, exhibiting spherical shapes. This is representative of the microstructure after it solidifies and cools to room temperature, without experiencing additional thermal cycles. The γ′ size slightly increases to 147 ± 6 nm below this layer and remains constant until 0.4 mm (∼453rd layer) from the top. At this position, the microstructure still closely resembles that of the 460th layer. After the 453rd layer, the γ′ size grows rapidly to ∼503 ± 19 nm until reaching the 437th layer (1.2 mm from top). The γ′ particles here have a cuboidal shape, and a small fraction is coarser than 600 nm. γ′ continue to grow steadily from this position to the bottom (23 mm from the top). A small fraction of γ′ is > 800 nm.
Besides primary γ′, secondary γ′ with sizes ranging from 5 to 50 nm were also found. These secondary γ′ precipitates, as seen in Fig. 1f, were present only in the bottom and middle regions. A detailed analysis of the multimodal size distribution of γ′ can be found in [16]. There is no significant variation in the phase fraction of the γ′ along the build. The phase fraction is ∼ 52%, as displayed in Fig. 2. It is worth mentioning that the total phase fraction of γ′ was estimated based on the primary γ′ phase fraction because of the small size of secondary γ′. Spherical MC carbides with sizes ranging from 50 to 400 nm and a phase fraction of 0.8% were also observed throughout the build. The carbides are the light grey precipitates in Fig. 1g. The light grey shade of carbides in the SEM images is due to their composition and crystal structure [52]. These carbides are not visible in Fig. 1b-e because they were dissolved during electro-etching carried out after electropolishing. In Fig. 1g, however, the sample was examined directly after electropolishing, without electro-etching.
Table 2 shows the nominal and measured composition of γ′ precipitates throughout the build by atom probe microscopy as determined in our previous study [17]. No build height-dependent composition difference was observed in either of the γ′ precipitate populations. However, there was a slight disparity between the composition of primary and secondary γ′. Among the main γ′ forming elements, the primary γ′ has a high Ti concentration while secondary γ′ has a high Al concentration. A detailed description of the atom distribution maps and the proxigrams of the constituent elements of γ′ throughout the build can be found in [17].
Table 2. Bulk IN738 composition determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES). Compositions of γ, primary γ′, and secondary γ′ at various locations in the build measured by APT. This information is reproduced from data in Ref. [17] with permission.
at%
Ni
Cr
Co
Al
Mo
W
Ti
Nb
C
B
Zr
Ta
Others
Bulk
59.12
17.47
8.48
7.00
1.01
0.81
3.96
0.49
0.47
0.05
0.09
0.56
0.46
γ matrix
Top
50.48
32.91
11.59
1.94
1.39
0.82
0.44
0.8
0.03
0.03
0.02
–
0.24
Mid
50.37
32.61
11.93
1.79
1.54
0.89
0.44
0.1
0.03
0.02
0.02
0.01
0.23
Bot
48.10
34.57
12.08
2.14
1.43
0.88
0.48
0.08
0.04
0.03
0.01
–
0.12
Primary γ′
Top
72.17
2.51
3.44
12.71
0.25
0.39
7.78
0.56
–
0.03
0.02
0.05
0.08
Mid
71.60
2.57
3.28
13.55
0.42
0.68
7.04
0.73
–
0.01
0.03
0.04
0.04
Bot
72.34
2.47
3.86
12.50
0.26
0.44
7.46
0.50
0.05
0.02
0.02
0.03
0.04
Secondary γ′
Mid
70.42
4.20
3.23
14.19
0.63
1.03
5.34
0.79
0.03
–
0.04
0.04
0.05
Bot
69.91
4.06
3.68
14.32
0.81
1.04
5.22
0.65
0.05
–
0.10
0.02
0.11
3.2. Hardness
Fig. 3a shows the Vickers hardness mapping performed along the entire X-Z surface, while Fig. 3b shows the plot of average hardness at different build heights. This hardness distribution is consistent with the γ′ precipitate size gradient across the build direction in Fig. 1, Fig. 2. The maximum hardness of ∼530 HV1 is found at ∼0.5 mm away from the top surface (Z = 22.5), where γ′ particles exhibit the smallest observed size in Fig. 2b. Further down the build (∼ 2 mm from the top), the hardness drops to the 440–490 HV1 range. This represents the region where γ′ begins to coarsen. The hardness drops further to 380–430 HV1 at the bottom of the build.
3.3. Modeling of the microstructural evolution during E-PBF
3.3.1. Thermal profile modeling
Fig. 4 shows the simulated thermal profile of the E-PBF build at a location of 23 mm from the top of the build, using a semi-analytical heat conduction model. This profile consists of the time taken to deposit 460 layers until final cooling, as shown in Fig. 4a. Fig. 4b-d show the magnified regions of Fig. 4a and reveal the first 20 layers from the top, a single layer (first layer from the top), and the time taken for the build to cool after the last layer deposition, respectively.
The peak temperatures experienced by previous layers decrease progressively as the number of layers increases but never fall below the build preheat temperature (1000 °C). Our simulated thermal cycle may not completely capture the complexity of the actual thermal cycle utilized in the E-PBF build. For instance, the top layer (Fig. 4c), also representing the first deposit’s thermal profile without additional cycles (from powder heating, melting, to solidification), recorded the highest peak temperature of 1390 °C. Although this temperature is above the melting range of the alloy (1230–1360 °C) [62], we believe a much higher temperature was produced by the electron beam to melt the powder. Nevertheless, the solidification temperature and dynamics are outside the scope of this study as our focus is on the solid-state phase transformations during deposition. It takes ∼25 s for each layer to be deposited and cooled to the build temperature. The interlayer dwell time is 125 s. The time taken for the build to cool to room temperature (RT) after final layer deposition is ∼4.7 hrs (17,000 s).
3.3.2. MatCalc simulation
During the MatCalc simulation, the matrix phase is defined as γ. γ′, and MC carbide are included as possible precipitates. The domain of these precipitates is set to be the matrix (γ), and nucleation is assumed to be homogenous. In homogeneous nucleation, all atoms of the unit volume are assumed to be potential nucleation sites. Table 3 shows the computational parameters used in the simulation. All other parameters were set at default values as recommended in the version 6.04.0011 of MatCalc. The values for the interfacial energies are automatically calculated according to the generalized nearest neighbor broken bond model and is one of the most outstanding features in MatCalc [56], [57], [58]. It should be noted that the elastic misfit strain was not included in the calculation. The output of MatCalc includes phase fraction, size, nucleation rate, and composition of the precipitates. The phase fraction in MatCalc is the volume fraction. Although the experimental phase fraction is the measured area fraction, it is relatively similar to the volume fraction. This is because of the generally larger precipitate size and similar morphology at the various locations along the build [63]. A reliable phase fraction comparison between experiment and simulation can therefore be made.
Table 3. Computational parameters used in the simulation.
γ′ = 0.080–0.140 J/m2 and MC carbide = 0.410–0.430 J/m2
3.3.2.1. Precipitate phase fraction
Fig. 5a shows the simulated phase fraction of γ′ and MC carbide during thermal cycling. Fig. 5b is a magnified view of 5a showing the simulated phase fraction at the center points of the top 70 layers, whereas Fig. 5c corresponds to the first two layers from the top. As mentioned earlier, the top layer (460th layer) represents the microstructure after solidification. The microstructure of the layers below is determined by the number of thermal cycles, which increases with distance to the top. For example, layers 459, 458, 457, up to layer 1 (region of interest) experience 1, 2, 3 and 459 thermal cycles, respectively. In the top layer in Fig. 5c, the volume fraction of γ′ and carbides increases with temperature. For γ′, it decreases to zero when the temperature is above the solvus temperature after a few seconds. Carbides, however, remain constant in their volume fraction reaching equilibrium (phase fraction ∼ 0.9%) in a short time. The topmost layer can be compared to the first deposit, and the peak in temperature symbolizes the stage where the electron beam heats the powder until melting. This means γ′ and carbide precipitation might have started in the powder particles during heating from the build temperature and electron beam until the onset of melting, where γ′ dissolves, but carbides remain stable [28].
During cooling after deposition, γ′ reprecipitates at a temperature of 1085 °C, which is below its solvus temperature. As cooling progresses, the phase fraction increases steadily to ∼27% and remains constant at 1000 °C (elevated build temperature). The calculated equilibrium fraction of phases by MatCalc is used to show the complex precipitation characteristics in this alloy. Fig. 6 shows that MC carbides form during solidification at 1320 °C, followed by γ′, which precipitate when the solidified layer cools to 1140 °C. This indicates that all deposited layers might contain a negligible amount of these precipitates before subsequent layer deposition, while being at the 1000 °C build temperature or during cooling to RT. The phase diagram also shows that the equilibrium fraction of the γ′ increases as temperature decreases. For instance, at 1000, 900, and 800 °C, the phase fractions are ∼30%, 38%, and 42%, respectively.
Deposition of subsequent layers causes previous layers to undergo phase transformations as they are exposed to several thermal cycles with different peak temperatures. In Fig. 5c, as the subsequent layer is being deposited, γ′ in the previous layer (459th layer) begins to dissolve as the temperature crosses the solvus temperature. This is witnessed by the reduction of the γ′ phase fraction. This graph also shows how this phase dissolves during heating. However, the phase fraction of MC carbide remains stable at high temperatures and no dissolution is seen during thermal cycling. Upon cooling, the γ′ that was dissolved during heating reprecipitates with a surge in the phase fraction until 1000 °C, after which it remains constant. This microstructure is similar to the solidification microstructure (layer 460), with a similar γ′ phase fraction (∼27%).
The complete dissolution and reprecipitation of γ′ continue for several cycles until the 50th layer from the top (layer 411), where the phase fraction does not reach zero during heating to the peak temperature (see Fig. 5d). This indicates the ‘partial’ dissolution of γ′, which continues progressively with additional layers. It should be noted that the peak temperatures for layers that underwent complete dissolution were much higher (1170–1300 °C) than the γ′ solvus.
The dissolution and reprecipitation of γ′ during thermal cycling are further confirmed in Fig. 7, which summarizes the nucleation rate, phase fraction, and concentration of major elements that form γ′ in the matrix. Fig. 7b magnifies a single layer (3rd layer from top) within the full dissolution region in Fig. 7a to help identify the nucleation and growth mechanisms. From Fig. 7b, γ′ nucleation begins during cooling whereby the nucleation rate increases to reach a maximum value of approximately 1 × 1020 m−3s−1. This fast kinetics implies that some rearrangement of atoms is required for γ′ precipitates to form in the matrix [65], [66]. The matrix at this stage is in a non-equilibrium condition. Its composition is similar to the nominal composition and remains unchanged. The phase fraction remains insignificant at this stage although nucleation has started. The nucleation rate starts declining upon reaching the peak value. Simultaneously, diffusion-controlled growth of existing nuclei occurs, depleting the matrix of γ′ forming elements (Al and Ti). Thus, from (7), (11), ∆�vol continuously decreases until nucleation ceases. The growth of nuclei is witnessed by the increase in phase fraction until a constant level is reached at 27% upon cooling to and holding at build temperature. This nucleation event is repeated several times.
At the onset of partial dissolution, the nucleation rate jumps to 1 × 1021 m−3s−1, and then reduces sharply at the middle stage of partial dissolution. The nucleation rate reaches 0 at a later stage. Supplementary Fig. S1 shows a magnified view of the nucleation rate, phase fraction, and thermal profile, underpinning this trend. The jump in nucleation rate at the onset is followed by a progressive reduction in the solute content of the matrix. The peak temperatures (∼1130–1160 °C) are lower than those in complete dissolution regions but still above or close to the γ′ solvus. The maximum phase fraction (∼27%) is similar to that of the complete dissolution regions. At the middle stage, the reduction in nucleation rate is accompanied by a sharp drop in the matrix composition. The γ′ fraction drops to ∼24%, where the peak temperatures of the layers are just below or at γ′ solvus. The phase fraction then increases progressively through the later stage of partial dissolution to ∼30% towards the end of thermal cycling. The matrix solute content continues to drop although no nucleation event is seen. The peak temperatures are then far below the γ′ solvus. It should be noted that the matrix concentration after complete dissolution remains constant. Upon cooling to RT after final layer deposition, the nucleation rate increases again, indicating new nucleation events. The phase fraction reaches ∼40%, with a further depletion of the matrix in major γ′ forming elements.
3.3.2.2. γ′ size distribution
Fig. 8 shows histograms of the γ′ precipitate size distributions (PSD) along the build height during deposition. These PSDs are predicted at the end of each layer of interest just before final cooling to room temperature, to separate the role of thermal cycles from final cooling on the evolution of γ′. The PSD for the top layer (layer 460) is shown in Fig. 8a (last solidified region with solidification microstructure). The γ′ size ranges from 120 to 230 nm and is similar to the 44 layers below (2.2 mm from the top).
Further down the build, γ′ begins to coarsen after layer 417 (44th layer from top). Fig. 8c shows the PSD after the 44th layer, where the γ′ size exhibits two peaks at ∼120–230 and ∼300 nm, with most of the population being in the former range. This is the onset of partial dissolution where simultaneously with the reprecipitation and growth of fresh γ′, the undissolved γ′ grows rapidly through diffusive transport of atoms to the precipitates. This is shown in Fig. 8c, where the precipitate class sizes between 250 and 350 represent the growth of undissolved γ′. Although this continues in the 416th layer, the phase fractions plot indicates that the onset of partial dissolution begins after the 411th layer. This implies that partial dissolution started early, but the fraction of undissolved γ′ was too low to impact the phase fraction. The reprecipitated γ′ are mostly in the 100–220 nm class range and similar to those observed during full dissolution.
As the number of layers increases, coarsening intensifies with continued growth of more undissolved γ′, and reprecipitation and growth of partially dissolved ones. Fig. 8d, e, and f show this sequence. Further down the build, coarsening progresses rapidly, as shown in Figs. 8d, 8e, and 8f. The γ′ size ranges from 120 to 1100 nm, with the peaks at 160, 180, and 220 nm in Figs. 8d, 8e, and 8f, respectively. Coarsening continues until nucleation ends during dissolution, where only the already formed γ′ precipitates continue to grow during further thermal cycling. The γ′ size at this point is much larger, as observed in layers 361 and 261, and continues to increase steadily towards the bottom (layer 1). Two populations in the ranges of ∼380–700 and ∼750–1100 nm, respectively, can be seen. The steady growth of γ′ towards the bottom is confirmed by the gradual decrease in the concentration of solute elements in the matrix (Fig. 7a). It should be noted that for each layer, the γ′ class with the largest size originates from continuous growth of the earliest set of the undissolved precipitates.
Fig. 9, Fig. 10 and supplementary Figs. S2 and S3 show the γ′ size evolution during heating and cooling of a single layer in the full dissolution region, and early, middle stages, and later stages of partial dissolution, respectively. In all, the size of γ′ reduces during layer heating. Depending on the peak temperature of the layer which varies with build height, γ′ are either fully or partially dissolved as mentioned earlier. Upon cooling, the dissolved γ′ reprecipitate.
In Fig. 9, those layers that underwent complete dissolution (top layers) were held above γ′ solvus temperature for longer. In Fig. 10, layers at the early stage of partial dissolution spend less time in the γ′ solvus temperature region during heating, leading to incomplete dissolution. In such conditions, smaller precipitates are fully dissolved while larger ones shrink [67]. Layers in the middle stages of partial dissolution have peak temperatures just below or at γ′ solvus, not sufficient to achieve significant γ′ dissolution. As seen in supplementary Fig. S2, only a few smaller γ′ are dissolved back into the matrix during heating, i.e., growth of precipitates is more significant than dissolution. This explains the sharp decrease in concentration of Al and Ti in the matrix in this layer.
The previous sections indicate various phenomena such as an increase in phase fraction, further depletion of matrix composition, and new nucleation bursts during cooling. Analysis of the PSD after the final cooling of the build to room temperature allows a direct comparison to post-printing microstructural characterization. Fig. 11 shows the γ′ size distribution of layer 1 (460th layer from the top) after final cooling to room temperature. Precipitation of secondary γ′ is observed, leading to the multimodal size distribution of secondary and primary γ′. The secondary γ′ size falls within the 10–80 nm range. As expected, a further growth of the existing primary γ′ is also observed during cooling.
3.3.2.3. γ′ chemistry after deposition
Fig. 12 shows the concentration of the major elements that form γ′ (Al, Ti, and Ni) in the primary and secondary γ′ at the bottom of the build, as calculated by MatCalc. The secondary γ′ has a higher Al content (13.5–14.5 at% Al), compared to 13 at% Al in the primary γ′. Additionally, within the secondary γ′, the smallest particles (∼10 nm) have higher Al contents than larger ones (∼70 nm). In contrast, for the primary γ′, there is no significant variation in the Al content as a function of their size. The Ni concentration in secondary γ′ (71.1–72 at%) is also higher in comparison to the primary γ′ (70 at%). The smallest secondary γ′ (∼10 nm) have higher Ni contents than larger ones (∼70 nm), whereas there is no substantial change in the Ni content of primary γ′, based on their size. As expected, Ti shows an opposite size-dependent variation. It ranges from ∼ 7.7–8.7 at% Ti in secondary γ′ to ∼9.2 at% in primary γ′. Similarly, within the secondary γ′, the smallest (∼10 nm) have lower Al contents than the larger ones (∼70 nm). No significant variation is observed for Ti content in primary γ′.
4. Discussion
A combined modelling method is utilized to study the microstructural evolution during E-PBF of IN738. The presented results are discussed by examining the precipitation and dissolution mechanism of γ′ during thermal cycling. This is followed by a discussion on the phase fraction and size evolution of γ′ during thermal cycling and after final cooling. A brief discussion on carbide morphology is also made. Finally, a comparison is made between the simulation and experimental results to assess their agreement.
4.1. γ′ morphology as a function of build height
4.1.1. Nucleation of γ′
The fast precipitation kinetics of the γ′ phase enables formation of γ′ upon quenching from higher temperatures (above solvus) during thermal cycling [66]. In Fig. 7b, for a single layer in the full dissolution region, during cooling, the initial increase in nucleation rate signifies the first formation of nuclei. The slight increase in nucleation rate during partial dissolution, despite a decrease in the concentration of γ′ forming elements, may be explained by the nucleation kinetics. During partial dissolution and as the precipitates shrink, it is assumed that the regions at the vicinity of partially dissolved precipitates are enriched in γ′ forming elements [68], [69]. This differs from the full dissolution region, in which case the chemical composition is evenly distributed in the matrix. Several authors have attributed the solute supersaturation of the matrix around primary γ′ to partial dissolution during isothermal ageing [69], [70], [71], [72]. The enhanced supersaturation in the regions close to the precipitates results in a much higher driving force for nucleation, leading to a higher nucleation rate upon cooling. This phenomenon can be closely related to the several nucleation bursts upon continuous cooling of Ni-based superalloys, where second nucleation bursts exhibit higher nucleation rates [38], [68], [73], [74].
At middle stages of partial dissolution, the reduction in the nucleation rate indicates that the existing composition and low supersaturation did not trigger nucleation as the matrix was closer to the equilibrium state. The end of a nucleation burst means that the supersaturation of Al and Ti has reached a low level, incapable of providing sufficient driving force during cooling to or holding at 1000 °C for further nucleation [73]. Earlier studies on Ni-based superalloys have reported the same phenomenon during ageing or continuous cooling from the solvus temperature to RT [38], [73], [74].
4.1.2. Dissolution of γ′ during thermal cycling
γ′ dissolution kinetics during heating are fast when compared to nucleation due to exponential increase in phase transformation and diffusion activities with temperature [65]. As shown in Fig. 9, Fig. 10, and supplementary Figs. S2 and S3, the reduction in γ′ phase fraction and size during heating indicates γ′ dissolution. This is also revealed in Fig. 5 where phase fraction decreases upon heating. The extent of γ′ dissolution mostly depends on the temperature, time spent above γ′ solvus, and precipitate size[75], [76], [77]. Smaller γ′ precipitates are first to be dissolved [67], [77], [78]. This is mainly because more solute elements need to be transported away from large γ′ precipitates than from smaller ones [79]. Also, a high temperature above γ′ solvus temperature leads to a faster dissolution rate[80]. The equilibrium solvus temperature of γ′ in IN738 in our MatCalc simulation (Fig. 6) and as reported by Ojo et al. [47] is 1140 °C and 1130–1180 °C, respectively. This means the peak temperature experienced by previous layers decreases progressively from γ′ supersolvus to subsolvus, near-solvus, and far from solvus as the number of subsequent layers increases. Based on the above, it can be inferred that the degree of dissolution of γ′ contributes to the gradient in precipitate distribution.
Although the peak temperatures during later stages of partial dissolution are much lower than the equilibrium γ′ solvus, γ′ dissolution still occurs but at a significantly lower rate (supplementary Fig. S3). Wahlmann et al. [28] also reported a similar case where they observed the rapid dissolution of γ′ in CMSX-4 during fast heating and cooling cycles at temperatures below the γ′ solvus. They attributed this to the γ′ phase transformation process taking place in conditions far from the equilibrium. While the same reasoning may be valid for our study, we further believe that the greater surface area to volume ratio of the small γ′ precipitates contributed to this. This ratio means a larger area is available for solute atoms to diffuse into the matrix even at temperatures much below the solvus [81].
4.2. γ′ phase fraction and size evolution
4.2.1. During thermal cycling
In the first layer, the steep increase in γ′ phase fraction during heating (Fig. 5), which also represents γ′ precipitation in the powder before melting, has qualitatively been validated in [28]. The maximum phase fraction of 27% during the first few layers of thermal cycling indicates that IN738 theoretically could reach the equilibrium state (∼30%), but the short interlayer time at the build temperature counteracts this. The drop in phase fraction at middle stages of partial dissolution is due to the low number of γ′ nucleation sites [73]. It has been reported that a reduction of γ′ nucleation sites leads to a delay in obtaining the final volume fraction as more time is required for γ′ precipitates to grow and reach equilibrium [82]. This explains why even upon holding for 150 s before subsequent layer deposition, the phase fraction does not increase to those values that were observed in the previous full γ′ dissolution regions. Towards the end of deposition, the increase in phase fraction to the equilibrium value of 30% is as a result of the longer holding at build temperature or close to it [83].
During thermal cycling, γ′ particles begin to grow immediately after they first precipitate upon cooling. This is reflected in the rapid increase in phase fraction and size during cooling in Fig. 5 and supplementary Fig. S2, respectively. The rapid growth is due to the fast diffusion of solute elements at high temperatures [84]. The similar size of γ′ for the first 44 layers from the top can be attributed to the fact that all layers underwent complete dissolution and hence, experienced the same nucleation event and growth during deposition. This corresponds with the findings by Balikci et al. [85], who reported that the degree of γ′ precipitation in IN738LC does not change when a solution heat treatment is conducted above a certain critical temperature.
The increase in coarsening rate (Fig. 8) during thermal cycling can first be ascribed to the high peak temperature of the layers [86]. The coarsening rate of γ′ is known to increase rapidly with temperature due to the exponential growth of diffusion activity. Also, the simultaneous dissolution with coarsening could be another reason for the high coarsening rate, as γ′ coarsening is a diffusion-driven process where large particles grow by consuming smaller ones [78], [84], [86], [87]. The steady growth of γ′ towards the bottom of the build is due to the much lower layer peak temperature, which is almost close to the build temperature, and reduced dissolution activity, as is seen in the much lower solute concentration in γ′ compared to those in the full and partial dissolution regions.
4.2.2. During cooling
The much higher phase fraction of ∼40% upon cooling signifies the tendency of γ′ to reach equilibrium at lower temperatures (Fig. 4). This is due to the precipitation of secondary γ′ and a further increase in the size of existing primary γ′, which leads to a multimodal size distribution of γ′ after cooling [38], [73], [88], [89], [90]. The reason for secondary γ′ formation during cooling is as follows: As cooling progresses, it becomes increasingly challenging to redistribute solute elements in the matrix owing to their lower mobility [38], [73]. A higher supersaturation level in regions away from or free of the existing γ′ precipitates is achieved, making them suitable sites for additional nucleation bursts. More cooling leads to the growth of these secondary γ′ precipitates, but as the temperature and in turn, the solute diffusivity is low, growth remains slow.
4.3. Carbides
MC carbides in IN738 are known to have a significant impact on the high-temperature strength. They can also act as effective hardening particles and improve the creep resistance [91]. Precipitation of MC carbides in IN738 and several other superalloys is known to occur during solidification or thermal treatments (e.g., hot isostatic pressing) [92]. In our case, this means that the MC carbides within the E-PBF build formed because of the thermal exposure from the E-PBF thermal cycle in addition to initial solidification. Our simulation confirms this as MC carbides appear during layer heating (Fig. 5). The constant and stable phase fraction of MC carbides during thermal cycling can be attributed to their high melting point (∼1360 °C) and the short holding time at peak temperatures [75], [93], [94]. The solvus temperature for most MC carbides exceeds most of the peak temperatures observed in our simulation, and carbide dissolution kinetics at temperatures above the solvus are known to be comparably slow [95]. The stable phase fraction and random distribution of MC carbides signifies the slight influence on the gradient in hardness.
4.4. Comparison of simulations and experiments
4.4.1. Precipitate phase fraction and morphology as a function of build height
A qualitative agreement is observed for the phase fraction of carbides, i.e. ∼0.8% in the experiment and ∼0.9% in the simulation. The phase fraction of γ′ differs, with the experiment reporting a value of ∼51% and the simulation, 40%. Despite this, the size distribution of primary γ′ along the build shows remarkable consistency between experimental and computational analyses. It is worth noting that the primary γ′ morphology in the experimental analysis is observed in the as-fabricated state, whereas the simulation (Fig. 8) captures it during deposition process. The primary γ′ size in the experiment is expected to experience additional growth during the cooling phase. Regardless, both show similar trends in primary γ′ size increments from the top to the bottom of the build. The larger primary γ’ size in the simulation versus the experiment can be attributed to the fact that experimental and simulation results are based on 2D and 3D data, respectively. The absence of stereological considerations [96] in our analysis could have led to an underestimation of the precipitate sizes from SEM measurements. The early starts of coarsening (8th layer) in the experiment compared to the simulation (45th layer) can be attributed to a higher actual γ′ solvus temperature than considered in our simulation [47]. The solvus temperature of γ′ in a Ni-based superalloy is mainly determined by the detailed composition. A high amount of Cr and Co are known to reduce the solvus temperature, whereas Ta and Mo will increase it [97], [98], [99]. The elemental composition from our experimental work was used for the simulation except for Ta. It should be noted that Ta is not included in the thermodynamic database in MatCalc used, and this may have reduced the solvus temperature. This could also explain the relatively higher γ′ phase fraction in the experiment than in simulation, as a higher γ′ solvus temperature will cause more γ′ to precipitate and grow early during cooling [99], [100].
Another possible cause of this deviation can be attributed to the extent of γ′ dissolution, which is mainly determined by the peak temperature. It can be speculated that individual peak temperatures at different layers in the simulation may have been over-predicted. However, one needs to consider that the true thermal profile is likely more complicated in the actual E-PBF process [101]. For example, the current model assumes that the thermophysical properties of the material are temperature-independent, which is not realistic. Many materials, including IN738, exhibit temperature-dependent properties such as thermal conductivity, specific heat capacity, and density [102]. This means that heat transfer simulations may underestimate or overestimate the temperature gradients and cooling rates within the powder bed and the solidified part. Additionally, the model does not account for the reduced thermal diffusivity through unmelted powder, where gas separating the powder acts as insulation, impeding the heat flow [1]. In E-PBF, the unmelted powder regions with trapped gas have lower thermal diffusivity compared to the fully melted regions, leading to localized temperature variations, and altered solidification behavior. These limitations can impact the predictions, particularly in relation to the carbide dissolution, as the peak temperatures may be underestimated.
While acknowledging these limitations, it is worth emphasizing that achieving a detailed and accurate representation of each layer’s heat source would impose tough computational challenges. Given the substantial layer count in E-PBF, our decision to employ a semi-analytical approximation strikes a balance between computational feasibility and the capture of essential trends in thermal profiles across diverse build layers. In future work, a dual-calibration strategy is proposed to further reduce simulation-experiment disparities. By refining temperature-independent thermophysical property approximations and absorptivity in the heat source model, and by optimizing interfacial energy descriptions in the kinetic model, the predictive precision could be enhanced. Further refining the simulation controls, such as adjusting the precipitate class size may enhance quantitative comparisons between modeling outcomes and experimental data in future work.
4.4.2. Multimodal size distribution of γ′ and concentration
Another interesting feature that sees qualitative agreement between the simulation and the experiment is the multimodal size distribution of γ′. The formation of secondary γ′ particles in the experiment and most E-PBF Ni-based superalloys is suggested to occur at low temperatures, during final cooling to RT [16], [73], [90]. However, so far, this conclusion has been based on findings from various continuous cooling experiments, as the study of the evolution during AM would require an in-situ approach. Our simulation unambiguously confirms this in an AM context by providing evidence for secondary γ′ precipitation during slow cooling to RT. Additionally, it is possible to speculate that the chemical segregation occurring during solidification, due to the preferential partitioning of certain elements between the solid and liquid phases, can contribute to the multimodal size distribution during deposition [51]. This is because chemical segregation can result in variations in the local composition of superalloys, which subsequently affects the nucleation and growth of γ′. Regions with higher concentrations of alloying elements will encourage the formation of larger γ′ particles, while regions with lower concentrations may favor the nucleation of smaller precipitates. However, it is important to acknowledge that the elevated temperature during the E-PBF process will largely homogenize these compositional differences [103], [104].
A good correlation is also shown in the composition of major γ′ forming elements (Al and Ti) in primary and secondary γ′. Both experiment and simulation show an increasing trend for Al content and a decreasing trend for Ti content from primary to secondary γ′. The slight composition differences between primary and secondary γ′ particles are due to the different diffusivity of γ′ stabilizers at different thermal conditions [105], [106]. As the formation of multimodal γ′ particles with different sizes occurs over a broad temperature range, the phase chemistry of γ′ will be highly size dependent. The changes in the chemistry of various γ′ (primary, secondary, and tertiary) have received significant attention since they have a direct influence on the performance [68], [105], [107], [108], [109]. Chen et al. [108], [109], reported a high Al content in the smallest γ′ precipitates compared to the largest, while Ti showed an opposite trend during continuous cooling in a RR1000 Ni-based superalloy. This was attributed to the temperature and cooling rate at which the γ′ precipitates were formed. The smallest precipitates formed last, at the lowest temperature and cooling rate. A comparable observation is evident in the present investigation, where the secondary γ′ forms at a low temperature and cooling rate in comparison to the primary. The temperature dependence of γ′ chemical composition is further evidenced in supplementary Fig. S4, which shows the equilibrium chemical composition of γ′ as a function of temperature.
5. Conclusions
A correlative modelling approach capable of predicting solid-state phase transformations kinetics in metal AM was developed. This approach involves computational simulations with a semi-analytical heat transfer model and the MatCalc thermo-kinetic software. The method was used to predict the phase transformation kinetics and detailed morphology and chemistry of γ′ and MC during E-PBF of IN738 Ni-based superalloy. The main conclusions are:
1.The computational simulations are in qualitative agreement with the experimental observations. This is particularly true for the γ′ size distribution along the build height, the multimodal size distribution of particles, and the phase fraction of MC carbides.
2.The deviations between simulation and experiment in terms of γ′ phase fraction and location in the build are most likely attributed to a higher γ′ solvus temperature during the experiment than in the simulation, which is argued to be related to the absence of Ta in the MatCalc database.
3.The dissolution and precipitation of γ′ occur fast and under non-equilibrium conditions. The level of γ′ dissolution determines the gradient in γ′ size distribution along the build. After thermal cycling, the final cooling to room temperature has further significant impacts on the final γ′ size, morphology, and distribution.
4.A negligible amount of γ′ forms in the first deposited layer before subsequent layer deposition, and a small amount of γ′ may also form in the powder induced by the 1000 °C elevated build temperature before melting.
Our findings confirm the suitability of MatCalc to predict the microstructural evolution at various positions throughout a build in a Ni-based superalloy during E-PBF. It also showcases the suitability of a tool which was originally developed for traditional thermo-mechanical processing of alloys to the new additive manufacturing context. Our simulation capabilities are likely extendable to other alloy systems that undergo solid-state phase transformations implemented in MatCalc (various steels, Ni-based superalloys, and Al-alloys amongst others) as well as other AM processes such as L-DED and L-PBF which have different thermal cycle characteristics. New tools to predict the microstructural evolution and properties during metal AM are important as they provide new insights into the complexities of AM. This will enable control and design of AM microstructures towards advanced materials properties and performances.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This research was sponsored by the Department of Industry, Innovation, and Science under the auspices of the AUSMURI program – which is a part of the Commonwealth’s Next Generation Technologies Fund. The authors acknowledge the facilities and the scientific and technical assistance at the Electron Microscope Unit (EMU) within the Mark Wainwright Analytical Centre (MWAC) at UNSW Sydney and Microscopy Australia. Nana Adomako is supported by a UNSW Scientia PhD scholarship. Michael Haines’ (UNSW Sydney) contribution to the revised version of the original manuscript is thankfully acknowledged.
[1]T. Debroy, H.L. Wei, J.S. Zuback, T. Mukherjee, J.W. Elmer, J.O. Milewski, A.M. Beese, A. Wilson-heid, A. De, W. ZhangAdditive manufacturing of metallic components – process, structure and propertiesProg. Mater. Sci., 92 (2018), pp. 112-224, 10.1016/j.pmatsci.2017.10.001View PDFView articleView in ScopusGoogle Scholar
[4]N.K. Adomako, J.J. Lewandowski, B.M. Arkhurst, H. Choi, H.J. Chang, J.H. KimMicrostructures and mechanical properties of multi-layered materials composed of Ti-6Al-4V, vanadium, and 17–4PH stainless steel produced by directed energy depositionAddit. Manuf., 59 (2022), Article 103174, 10.1016/j.addma.2022.103174View PDFView articleView in ScopusGoogle Scholar
[5]H. Wang, Z.G. Zhu, H. Chen, A.G. Wang, J.Q. Liu, H.W. Liu, R.K. Zheng, S.M.L. Nai, S. Primig, S.S. Babu, S.P. Ringer, X.Z. LiaoEffect of cyclic rapid thermal loadings on the microstructural evolution of a CrMnFeCoNi high-entropy alloy manufactured by selective laser meltingActa Mater., 196 (2020), pp. 609-625, 10.1016/J.ACTAMAT.2020.07.006View PDFView articleView in ScopusGoogle Scholar
[10]S.S. Al-Bermani, M.L. Blackmore, W. Zhang, I. ToddThe origin of microstructural diversity, texture, and mechanical properties in electron beam melted Ti-6Al-4VMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 41 (2010), pp. 3422-3434, 10.1007/s11661-010-0397-xView article View in ScopusGoogle Scholar
[13]H. Helmer, A. Bauereiß, R.F. Singer, C. KörnerErratum to: ‘Grain structure evolution in Inconel 718 during selective electron beam melting’ (Materials Science & Engineering A (2016) 668 (180–187 (S0921509316305536) (10.1016/j.msea.2016.05.046))Mater. Sci. Eng. A., 676 (2016), p. 546, 10.1016/j.msea.2016.09.016View PDFView articleView in ScopusGoogle Scholar
[16]N. Haghdadi, E. Whitelock, B. Lim, H. Chen, X. Liao, S.S. Babu, S.P. Ringer, S. PrimigMultimodal γ′ precipitation in Inconel-738 Ni-based superalloy during electron-beam powder bed fusion additive manufacturingJ. Mater. Sci., 55 (2020), pp. 13342-13350, 10.1007/s10853-020-04915-wView article View in ScopusGoogle Scholar
[17]B. Lim, H. Chen, Z. Chen, N. Haghdadi, X. Liao, S. Primig, S.S. Babu, A. Breen, S.P. RingerMicrostructure–property gradients in Ni-based superalloy (Inconel 738) additively manufactured via electron beam powder bed fusionAddit. Manuf. (2021), Article 102121, 10.1016/j.addma.2021.102121View PDFView articleView in ScopusGoogle Scholar
[18]P. Karimi, E. Sadeghi, P. Åkerfeldt, J. Ålgårdh, J. AnderssonInfluence of successive thermal cycling on microstructure evolution of EBM-manufactured alloy 718 in track-by-track and layer-by-layer designMater. Des., 160 (2018), pp. 427-441, 10.1016/j.matdes.2018.09.038View PDFView articleView in ScopusGoogle Scholar
[19]E. Chauvet, P. Kontis, E.A. Jägle, B. Gault, D. Raabe, C. Tassin, J.J. Blandin, R. Dendievel, B. Vayre, S. Abed, G. MartinHot cracking mechanism affecting a non-weldable Ni-based superalloy produced by selective electron Beam MeltingActa Mater., 142 (2018), pp. 82-94, 10.1016/j.actamat.2017.09.047View PDFView articleView in ScopusGoogle Scholar
[20]M. Ramsperger, R.F. Singer, C. KörnerMicrostructure of the nickel-base superalloy CMSX-4 fabricated by selective electron beam meltingMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 47 (2016), pp. 1469-1480, 10.1007/s11661-015-3300-y View PDF This article is free to access.View in ScopusGoogle Scholar
[21]B. Zhang, P. Wang, Y. Chew, Y. Wen, M. Zhang, P. Wang, G. Bi, J. WeiMechanical properties and microstructure evolution of selective laser melting Inconel 718 along building direction and sectional dimensionMater. Sci. Eng. A, 794 (2020), Article 139941, 10.1016/j.msea.2020.139941View PDFView articleView in ScopusGoogle Scholar
[22]C. Körner, M. Ramsperger, C. Meid, D. Bürger, P. Wollgramm, M. Bartsch, G. EggelerMicrostructure and mechanical properties of CMSX-4 single crystals prepared by additive manufacturingMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 49 (2018), pp. 3781-3792, 10.1007/s11661-018-4762-5 View PDF This article is free to access.View in ScopusGoogle Scholar
[23]B. Lim, H. Chen, K. Nomoto, Z. Chen, A.I. Saville, S. Vogel, A.J. Clarke, A. Paradowska, M. Reid, S. Primig, X. Liao, S.S. Babu, A.J. Breen, S.P. RingerAdditively manufactured Haynes-282 monoliths containing thin wall struts of varying thicknessesAddit. Manuf., 59 (2022), Article 103120, 10.1016/j.addma.2022.103120View PDFView articleView in ScopusGoogle Scholar
[24]C.L.A. Leung, S. Marussi, R.C. Atwood, M. Towrie, P.J. Withers, P.D. LeeIn situ X-ray imaging of defect and molten pool dynamics in laser additive manufacturingNat. Commun., 9 (2018), pp. 1-9, 10.1038/s41467-018-03734-7 View PDF This article is free to access.View in ScopusGoogle Scholar
[25]C. Zhao, K. Fezzaa, R.W. Cunningham, H. Wen, F. De Carlo, L. Chen, A.D. Rollett, T. SunReal-time monitoring of laser powder bed fusion process using high-speed X-ray imaging and diffractionSci. Rep., 7 (2017), pp. 1-11, 10.1038/s41598-017-03761-2 View PDF This article is free to access.View in ScopusGoogle Scholar
[26]C. Kenel, D. Grolimund, X. Li, E. Panepucci, V.A. Samson, D.F. Sanchez, F. Marone, C. LeinenbachIn situ investigation of phase transformations in Ti-6Al-4V under additive manufacturing conditions combining laser melting and high-speed micro-X-ray diffractionSci. Rep., 7 (2017), pp. 1-10, 10.1038/s41598-017-16760-0 View PDF This article is free to access.Google Scholar
[27]W.L. Bevilaqua, J. Epp, H. Meyer, J. Dong, H. Roelofs, A. da, S. Rocha, A. RegulyRevealing the dynamic transformation of austenite to bainite during uniaxial warm compression through in-situ synchrotron X-ray diffractionMetals, 11 (2021), pp. 1-14, 10.3390/met11030467View article View in ScopusGoogle Scholar
[28]B. Wahlmann, E. Krohmer, C. Breuning, N. Schell, P. Staron, E. Uhlmann, C. KörnerIn situ observation of γ′ phase transformation dynamics during selective laser melting of CMSX-4Adv. Eng. Mater., 23 (2021), 10.1002/adem.202100112 View PDF This article is free to access.Google Scholar
[35]M.J. Anderson, J. Benson, J.W. Brooks, B. Saunders, H.C. BasoaltoPredicting precipitation kinetics during the annealing of additive manufactured Inconel 625 componentsIntegr. Mater. Manuf. Innov., 8 (2019), pp. 154-166, 10.1007/S40192-019-00134-7/FIGURES/11 View PDFThis article is free to access.View in ScopusGoogle Scholar
[36]H.C. Basoalto, C. Panwisawas, Y. Sovani, M.J. Anderson, R.P. Turner, B. Saunders, J.W. BrooksA computational study on the three-dimensional printability of precipitate-strengthened nickel-based superalloysProc. R. Soc. A, 474 (2018), 10.1098/RSPA.2018.0295View article Google Scholar
[37]K. McNamara, Y. Ji, F. Lia, P. Promoppatum, S.C. Yao, H. Zhou, Y. Wang, L.Q. Chen, R.P. MartukanitzPredicting phase transformation kinetics during metal additive manufacturing using non-isothermal Johnson-Mehl-Avrami models: application to Inconel 718 and Ti-6Al-4VAddit. Manuf., 49 (2022), Article 102478, 10.1016/J.ADDMA.2021.102478View PDFView articleView in ScopusGoogle Scholar
[39]A. Drexler, B. Oberwinkler, S. Primig, C. Turk, E. Povoden-karadeniz, A. Heinemann, W. Ecker, M. StockingerMaterials Science & Engineering A Experimental and numerical investigations of the γ ″ and γ ′ precipitation kinetics in Alloy 718Mater. Sci. Eng. A., 723 (2018), pp. 314-323, 10.1016/j.msea.2018.03.013View PDFView articleView in ScopusGoogle Scholar
[44]A.V. Sotov, A.V. Agapovichev, V.G. Smelov, V.V. Kokareva, M.O. Dmitrieva, A.A. Melnikov, S.P. Golanov, Y.M. AnurovInvestigation of the IN-738 superalloy microstructure and mechanical properties for the manufacturing of gas turbine engine nozzle guide vane by selective laser meltingInt. J. Adv. Manuf. Technol., 107 (2020), pp. 2525-2535, 10.1007/s00170-020-05197-xView article View in ScopusGoogle Scholar
[49]S. Sanchez, P. Smith, Z. Xu, G. Gaspard, C.J. Hyde, W.W. Wits, I.A. Ashcroft, H. Chen, A.T. ClarePowder Bed Fusion of nickel-based superalloys: a reviewInt. J. Mach. Tools Manuf., 165 (2021), 10.1016/j.ijmachtools.2021.103729 View PDF This article is free to access.Google Scholar
[50]C.L.A. Leung, R. Tosi, E. Muzangaza, S. Nonni, P.J. Withers, P.D. LeeEffect of preheating on the thermal, microstructural and mechanical properties of selective electron beam melted Ti-6Al-4V componentsMater. Des., 174 (2019), Article 107792, 10.1016/j.matdes.2019.107792View PDFView articleView in ScopusGoogle Scholar
[51]S. Griffiths, H. Ghasemi Tabasi, T. Ivas, X. Maeder, A. De Luca, K. Zweiacker, R. Wróbel, J. Jhabvala, R.E. Logé, C. LeinenbachCombining alloy and process modification for micro-crack mitigation in an additively manufactured Ni-base superalloyAddit. Manuf., 36 (2020), 10.1016/j.addma.2020.101443View article Google Scholar
[52]P. Soille, L. Vincent Pierre Soille, L.M. Vincent, Determining watersheds in digital pictures via flooding simulations, Https://Doi.Org/10.1117/12.24211. 1360 (1990) 240–250. https://doi.org/10.1117/12.24211.Google Scholar
[53]ASTM Standard Test Method for Microindentation Hardness of MaterialsKnoop and Vickers Hardness of Materials 1, Annu. B ASTM Stand. i 2010 1 42.Google Scholar
[67]P. Strunz, M. Petrenec, J. Polák, U. Gasser, G. FarkasFormation and dissolution of’ precipitates in IN792 superalloy at elevated temperaturesMetals, 6 (2016), 10.3390/met6020037View article Google Scholar
[68]A.R.P. Singh, S. Nag, J.Y. Hwang, G.B. Viswanathan, J. Tiley, R. Srinivasan, H.L. Fraser, R. BanerjeeInfluence of cooling rate on the development of multiple generations of γ′ precipitates in a commercial nickel base superalloyMater. Charact., 62 (2011), pp. 878-886, 10.1016/j.matchar.2011.06.002View PDFView articleView in ScopusGoogle Scholar
[69]E. Balikci, A. Raman, R. MirshamsMicrostructure evolution in polycrystalline IN738LC in the range 1120 to 1250C, Zeitschrift FuerMet, 90 (1999), pp. 132-140View in ScopusGoogle Scholar
[71]Ł. Rakoczy, M. Grudzień-Rakoczy, F. Hanning, G. Cempura, R. Cygan, J. Andersson, A. Zielińska-LipiecInvestigation of the γ′ precipitates dissolution in a Ni-based superalloy during stress-free short-term annealing at high homologous temperaturesMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 52 (2021), pp. 4767-4784, 10.1007/s11661-021-06420-4 View PDF This article is free to access.View in ScopusGoogle Scholar
[73]F. Masoumi, D. Shahriari, M. Jahazi, J. Cormier, A. DevauxKinetics and Mechanisms of γ′ Reprecipitation in a Ni-based SuperalloySci. Rep., 6 (2016), pp. 1-16, 10.1038/srep28650View articleGoogle Scholar
[74]A.R.P. Singh, S. Nag, S. Chattopadhyay, Y. Ren, J. Tiley, G.B. Viswanathan, H.L. Fraser, R. BanerjeeMechanisms related to different generations of γ′ precipitation during continuous cooling of a nickel base superalloyActa Mater., 61 (2013), pp. 280-293, 10.1016/j.actamat.2012.09.058View PDFView articleView in ScopusGoogle Scholar
[76]N.D. Souza, M.C. Hardy, B. Roebuck, W.E.I. Li, G.D. West, D.M. Collins, On the Rate Dependence of Precipitate Formation and Dissolution in a Nickel-Base Superalloy, Metall. Mater. Trans. A. (n.d.). https://doi.org/10.1007/s11661–022-06680–8.Google Scholar
[79]H. Huang, G. Liu, H. Wang, A. Ullah, B. HuDissolution behavior and kinetics of γ′ phase during solution treatment in powder metallurgy nickel-based superalloyMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 51 (2020), pp. 1075-1084, 10.1007/s11661-019-05581-7View article View in ScopusGoogle Scholar
[80]A.J. Goodfellow, E.I. Galindo-Nava, K.A. Christofidou, N.G. Jones, T. Martin, P.A.J. Bagot, C.D. Boyer, M.C. Hardy, H.J. StoneGamma prime precipitate evolution during aging of a model nickel-based superalloyMetall. Mater. Trans. A Phys. Metall. Mater. Sci., 49 (2018), pp. 718-728, 10.1007/s11661-017-4336-y View PDF This article is free to access.View in ScopusGoogle Scholar
[81]T.P. Gabb, D.G. Backman, D.Y. Wei, D.P. Mourer, D. Furrer, A. Garg, D.L. Ellis, #947;’ Form. a Nickel-Base Disk Superalloy 2012 405 414 doi: 10.7449/2000/superalloys_2000_405_414.Google Scholar
[82]A. PlatiModelling of γ precipitation in superalloys University of CambridgeMater. Sci. (2003), p. 73Google Scholar
[91]F. Theska, W.F. Tse, B. Schulz, R. Buerstmayr, S.R. Street, M. Lison-Pick, S. PrimigReview of microstructure–mechanical property relationships in cast and wrought ni-based superalloys with boron, carbon, and zirconium microalloying additionsAdv. Eng. Mater. (2022), p. 2201514, 10.1002/ADEM.202201514 View PDF This article is free to access.Google Scholar
[93]L. Zhang, Y. Li, Q. Zhang, S. ZhangMicrostructure evolution, phase transformation and mechanical properties of IN738 superalloy fabricated by selective laser melting under different heat treatmentsMater. Sci. Eng. A, 844 (2022), Article 142947, 10.1016/j.msea.2022.142947View PDFView articleView in ScopusGoogle Scholar
[94]J.C. Franco-Correa, E. Martínez-Franco, J.M. Alvarado-Orozco, L.A. Cáceres-Díaz, D.G. Espinosa-Arbelaez, J.A. VilladaEffect of conventional heat treatments on the microstructure and microhardness of IN718 obtained by wrought and additive manufacturingJ. Mater. Eng. Perform., 30 (2021), pp. 7035-7045, 10.1007/s11665-021-06138-9View article View in ScopusGoogle Scholar
[96]N. Li, M.J. Anderson, H.C. BasoaltoAutomated stereology and uncertainty quantification considering spherical non-penetrating dispersionsPage 464.Cryst 2023, Vol. 13 (13) (2023), p. 464, 10.3390/CRYST13030464View article Google Scholar
[97]W.T. Loomis, J.W. Freeman, D.L. SponsellerInfluence of molybdenum on the γ′- phase in experimental nickelbase superalloysMet. Trans., 3 (1972), pp. 989-1000Google Scholar
[98]A.S. Shaikh Development of a γ’ Precipitation Hardening Ni-Base Superalloy for Additive Manufacturing Thesis 2018 102.〈https://odr.chalmers.se/handle/20.500.12380/255645%0Ahttps://www.researchgate.net/profile/Abdul_Shaafi_Shaikh2/publication/326226200%0Ahttps://drive.google.com/open?id=1BIez-aJyBTjnSgazzzTvv3jlrYpFC0N-〉.Google Scholar
Yupeng Ren abc, Huiguang Zhou cd, Houjie Wang ab, Xiao Wu ab, Guohui Xu cd, Qingsheng Meng cd
Abstract
해저 퇴적물 흐름은 퇴적물을 심해로 운반하는 주요 수단 중 하나이며, 종종 장거리를 이동하고 수십 또는 수백 킬로미터에 걸쳐 상당한 양의 퇴적물을 운반합니다. 그것의 강력한 파괴력은 종종 이동 과정에서 잠수함 유틸리티에 심각한 손상을 초래합니다.
퇴적물 흐름의 퇴적물 농도는 주변 해수와의 밀도차를 결정하며, 이 밀도 차이는 퇴적물 흐름의 흐름 능력을 결정하여 이송된 퇴적물의 최종 퇴적 위치에 영향을 미칩니다. 본 논문에서는 다양한 미사 및 점토 중량비(미사/점토 비율이라고 함)를 갖는 다양한 퇴적물 농도의 퇴적물 흐름을 수로 테스트를 통해 연구합니다.
우리의 테스트 결과는 특정 퇴적물 구성에 대해 퇴적물 흐름이 가장 빠르게 이동하는 임계 퇴적물 농도가 있음을 나타냅니다. 4가지 미사/점토 비율 각각에 대한 임계 퇴적물 농도와 이에 상응하는 최대 속도가 구해집니다. 결과는 점토 함량이 임계 퇴적물 농도와 선형적으로 음의 상관 관계가 있음을 나타냅니다.
퇴적물 농도가 증가함에 따라 퇴적물의 흐름 거동은 흐름 상태에서 붕괴된 상태로 변환되고 흐름 거동이 변화하는 두 탁한 현탁액의 유체 특성은 모두 Bingham 유체입니다.
또한 본 논문에서는 퇴적물 흐름 내 입자 배열을 분석하여 위에서 언급한 결과에 대한 미시적 설명도 제공합니다.
Submarine sediment flows is one of the main means for transporting sediment to the deep sea, often traveling long-distance and transporting significant volumes of sediment for tens or even hundreds of kilometers. Its strong destructive force often causes serious damage to submarine utilities on its course of movement. The sediment concentration of the sediment flow determines its density difference with the ambient seawater, and this density difference determines the flow ability of the sediment flow, and thus affects the final deposition locations of the transported sediment. In this paper, sediment flows of different sediment concentration with various silt and clay weight ratios (referred to as silt/clay ratio) are studied using flume tests. Our test results indicate that there is a critical sediment concentration at which sediment flows travel the fastest for a specific sediment composition. The critical sediment concentrations and their corresponding maximum velocities for each of the four silt/clay ratios are obtained. The results further indicate that the clay content is linearly negatively correlated with the critical sediment concentration. As the sediment concentration increases, the flow behaviors of sediment flows transform from the flow state to the collapsed state, and the fluid properties of the two turbid suspensions with changing flow behaviors are both Bingham fluids. Additionally, this paper also provides a microscopic explanation of the above-mentioned results by analyzing the arrangement of particles within the sediment flow.
Introduction
Submarine sediment flows are important carriers for sea floor sediment movement and may carry and transport significant volumes of sediment for tens or even hundreds of kilometers (Prior et al., 1987; Pirmez and Imran, 2003; Zhang et al., 2018). Earthquakes, storms, and floods may all trigger submarine sediment flow events (Hsu et al., 2008; Piper and Normark, 2009; Pope et al., 2017b; Gavey et al., 2017). Sediment flows have strong forces during the movement, which will cause great harm to submarine structures such as cables and pipelines (Pope et al., 2017a). It was first confirmed that the cable breaking event caused by the sediment flow occurred in 1929. The sediment flow triggered by the Grand Banks earthquake damaged 12 cables. According to the time sequence of the cable breaking, the maximum velocity of the sediment flow is as high as 28 m/s (Heezen and Ewing, 1952; Kuenen, 1952; Heezen et al., 1954). Subsequent research shows that the lowest turbidity velocity that can break the cable also needs to reach 19 m/s (Piper et al., 1988). Since then, there have been many damage events of submarine cables and oil and gas pipelines caused by sediment flows in the world (Hsu et al., 2008; Carter et al., 2012; Cattaneo et al., 2012; Carter et al., 2014). During its movement, the sediment flow will gradually deposit a large amount of sediment carried by it along the way, that is, the deposition process of the sediment flow. On the one hand, this process brings a large amount of terrestrial nutrients and other materials to the ocean, while on the other hand, it causes damage and burial to benthic organisms, thus forming the largest sedimentary accumulation on Earth – submarine fans, which are highly likely to become good reservoirs for oil and gas resources (Daly, 1936; Yuan et al., 2010; Wu et al., 2022). The study on sediment flows (such as, the study of flow velocity and the forces acting on seabed structures) can provide important references for the safe design of seabed structures, the protection of submarine ecosystems, and exploration of turbidity sediments related oil and gas deposits. Therefore, it is of great significance to study the movement of sediment flows.
The sediment flow, as a highly sediment-concentrated fluid flowing on the sea floor, has a dense bottom layer and a dilute turbulent cloud. Observations at the Monterey Canyon indicated that the sediment flow can maintain its movement over long distances if its bottom has a relatively high sediment concentration. This dense bottom layer can be very destructive along its movement path to any facilities on the sea floor (Paull et al., 2018; Heerema et al., 2020; Wang et al., 2020). The sediment flow mentioned in this research paper is the general term of sediment density flow.
The sediment flow, which occurs on the seafloor, has the potential to cause erosion along its path. In this process, the suspended sediment is replenished, allowing the sediment flow to maintain its continuous flow capacity (Zhao et al., 2018). The dynamic force of sediment flow movement stem from its own gravity and density difference with surrounding water. In cases that the gravity drive of the slope is absent (on a flat sea floor), the flow velocity and distance of sediment flows are essentially determined by the sediment composition and concentration of the sediment flows as previous studies have demonstrated. Ilstad et al. (2004) conducted underwater flow tests in a sloped tank and employed high speed video camera to perform particle tracking. The results indicated that the premixed sand-rich and clay-rich slurries demonstrated different flow velocity and flow behavior. Using mixed kaolinite(d50 = 6 μm) and silica flour(d50 = 9 μm) in three compositions with total volumetric concentration ranged 22% or 28%, Felix and Peakall (2006) carried out underwater flow tests in a 5° slope Perspex channel and found that the flow ability of sediment flows is different depending on sediment compositions and concentrations. Sumner et al. (2009) used annular flume experiments to investigate the depositional dynamics and deposits of waning sediment-laden flows, finding that decelerating fast flows with fixed sand content and variable mud content resulted in four different deposit types. Chowdhury and Testik (2011) used lock-exchange tank, and experimented the kaolin clay sediment flows in the concentration range of 25–350 g/L, and predicted the fluid mud sediment flows propagation characteristics, but this study focused on giving sediment flows propagate phase transition time parameters, and is limited to clay. Lv et al. (2017) found through experiments that the rheological properties and flow behavior of kaolin clay (d50 = 3.7 μm) sediment flows were correlated to clay concentrations. In the field monitoring conducted by Liu et al. (2023) at the Manila Trench in the South China Sea in 2021, significant differences in the velocity, movement distance, and flow morphology of turbidity currents were observed. These differences may be attributed to variations in the particle composition of the turbidity currents.
On low and gentle slopes, although sediment flow with sand as the main sediment composition moves faster, it is difficult to propagate over long distances because sand has greater settling velocity and subaqueous angle of repose. Whereas the sediment flows with silt and clay as main composition may maintain relatively stable currents. Although its movement speed is slow, it has the ability to propagate over long distances because of the low settling rate of the fine particles (Ilstad et al., 2004; Liu et al., 2023). In a field observation at the Gaoping submarine canyon, the sediments collected from the sediment flows exhibited grain size gradation and the sediment was mostly composed of silt and clay (Liu et al., 2012). At the largest deltas in the world, for instance, the Mississippi River Delta, the sediments are mainly composed of silt and clay, which generally distributed along the coast in a wide range and provided the sediment sources for further distribution. The sediment flows originated and transported sediment from the coast to the deep sea are therefore share the same sediment compositions as delta sediments. To study the sediment flows composed of silt and clay is of great importance.
The sediment concentration of the sediment flows determines the density difference between the sediment flows and the ambient water and plays a key role in its flow ability. For the sediment flow with sediment composed of silt and clay, low sediment concentration means low density and therefore leads to low flow ability; however, although high sediment concentration results in high density, since there is cohesion between fine particles, it changes fluid properties and leads to low flow ability as well. Therefore, there should be a critical sediment concentration with mixed composition of silt and clay, at which the sediment flow maintains its strongest flow capacity and have the highest movement speed. In other words, the two characteristics of particle diameter and concentration of the sediment flow determine its own motion ability, which, if occurs, may become the most destructive force to submarine structures.
The objectives of this work was to study how the sediment composition (measured in relative weight of silt and clay, and referred as silt/clay ratio) and sediment concentration affect flow ability and behavior of the sediment flows, and to quantify the critical sediment concentration at which the sediment flows reached the greatest flow velocity under the experiment setting. We used straight flume without slope and conducted a series of flume tests with varying sediment compositions (silt-rich or clay-rich) and concentrations (96 to 1212 g/L). Each sediment flow sample was tested and analyzed for rheological properties using a rheometer, in order to characterize the relationship between flow behavior and rheological properties. Combined with the particle diameter, density and viscosity characteristics of the sediment flows measured in the experiment, a numerical modeling study is conducted, which are mutually validated with the experimental results.
The sediment concentration determines the arrangements of the sediment particles in the turbid suspension, and the arrangement impacts the fluid properties of the turbid suspension. The microscopic mode of particle arrangement in the turbid suspension can be constructed to further analyze the relationship between the fluid properties of turbid suspension and the flow behaviors of the sediment flow, and then characterize the critical sediment concentration at which the sediment flow runs the fastest. A simplified microscopic model of particle arrangement in turbid suspension was constructed to analyze the microscopic arrangement characteristics of sediment particles in turbid suspension with the fastest velocity.
Section snippets
Equipment and materials
The sediment flows flow experiments were performed in a Perspex channel with smooth transparent walls. The layout and dimensions of the experimental set-up were shown in Fig. 1. The bottom of the channel was flat and straight, and a gate was arranged to separate the two tanks. In order to study the flow capacity of turbidity currents from the perspective of their own composition (particle size distribution and concentration), we used a straight channel instead of an inclined one, to avoid any
Relationship between sediment flow flow velocity and sediment concentration
After the sediment flow is generated, its movement in the first half (50 cm) of the channel is relatively stable, and there is obvious shock diffusion in the second half. The reason is that the excitation wave (similar to the surge) will be formed during the sediment flow movement, and its speed is much faster than the speed of the sediment flow head. When the excitation wave reaches the tail of the channel, it will be reflected, thus affecting the subsequent flow of the sediment flow.
Sediment flows motion simulation based on FLOW-3D
As a relatively mature 3D fluid simulation software, FLOW-3D can accurately predict the free surface flow, and has been used to simulate the movement process of sediment flows for many times (Heimsund, 2007). The model adopted in this paper is RNG turbulence model, which can better deal with the flow with high strain rate and is suitable for the simulation of sediment flows with variable shape during movement. The governing equations of the numerical model involved include continuity equation,
Conclusions
In this study, we conducted a series of sediment flow flume tests with mixed silt and clay sediment samples in four silt/clay ratios on a flat slope. Rheological measurements were carried out on turbid suspension samples and microstructure analysis of the sediment particle arrangements was conducted, we concluded that:
(1)The flow velocity of the sediment flow is controlled by the sediment concentration and its own particle diameter composition, the flow velocity increased with the increase of the
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the National Natural Science Foundation of China [Grant no. 42206055]; the National Natural Science Foundation of China [Grant no. 41976049]; and the National Natural Science Foundation of China [Grant no. 42272327].
R.A. BagnoldAuto-suspension of transported sediment; turbidity currentsProc. R. Soc. Lond.(1962)
L. Carter et al.Near-synchronous and delayed initiation of long run-out submarine sediment flows from a record-breaking river flood, offshore TaiwanGeophys. Res. Lett.(2012)
L. Carter et al.Insights into submarine geohazards from breaks in subsea telecommunication cablesOceanography(2014)
A. Cattaneo et al.Searching for the seafloor signature of the 21 May 2003 Boumerdès earthquake offshore Central AlgeriaNat. Hazard. Earth. Sys.(2012)
S. ChoiLayer-averaged modeling of two-dimensional turbidity currents with a dissipative; galerkin finite element method; part i, formulation and application exampleJ. Hydraul. Res.(1998)
S.U. Choi et al.k-ε Turbulence modeling of density currents developing two dimensionally on a slopeJ. Hydraul. Eng.(2002)
R.A. DalyOrigin of submarine canyonsAm. J. Sci.(1936)
M. Felix et al.Transformation of debris flows into turbidity currents : mechanisms inferred from laboratory experimentsSedimentology(2006)
F.H. Harlow et al.Turbulence transport equationsPhys. Fluids(1967)
G.J. Heerema et al.What determines the downstream evolution of turbidity currents?Earth Planet. Sci. Lett.(2020)
There are more references available in the full text version of this article.
전기화학 반응기에 대한 3D 수치 시뮬레이션 및 측정을 사용하여 동시 초음파 처리 유무에 관계없이 물에서 스트론튬 제거 효율을 분석했습니다. 초음파는 작동 주파수가 25kHz인 4개의 초음파 변환기를 사용하여 생성되었습니다. 반응기는 2개의 블록으로 배열된 8개의 알루미늄 전극을 사용했습니다.
LICHT K.1*, LONČAR G.1, POSAVČIĆ H.1, HALKIJEVIĆ I.1 1 Department of Hydroscience and Engineering, Faculty of Civil Engineering, University of Zagreb, Andrije Kačića-Miošića 26, 10000 Zagreb, Croatia *corresponding author: e-mail:katarina.licht@grad.unizg.hr
물 속의 스트론튬 이온은 3.2∙10-19C의 전하와 1.2∙10-8m의 직경을 특징으로 하는 입자로 모델링됩니다. 수치 모델은 기본 유체 역학 모듈, 정전기 모듈 및 일반 이동 객체 모듈을 사용하여 Flow-3D 소프트웨어에서 생성되었습니다.
수치 시뮬레이션을 통해 연구된 원자로 변형의 성능은 시뮬레이션 기간이 끝날 때 전극에 영구적으로 유지되는 모델 스트론튬 입자 수와 물 속의 초기 입자 수의 비율로 정의됩니다. 실험실 반응기의 경우 스트론튬 제거 효과는 실험 종료 시와 시작 시 물 내 균일한 스트론튬 농도의 비율로 정의됩니다.
결과는 초음파를 사용하면 수처리 180초 후에 스트론튬 제거 효과가 10.3%에서 11.2%로 증가한다는 것을 보여줍니다. 수치 시뮬레이션 결과는 동일한 기하학적 특성을 갖는 원자로에 대한 측정 결과와 일치합니다.
3D numerical simulations and measurements on an electrochemical reactor were used to analyze the efficiency of strontium removal from water, with and without simultaneous ultrasound treatment. Ultrasound was generated using 4 ultrasonic transducers with an operating frequency of 25 kHz. The reactor used 8 aluminum electrodes arranged in two blocks. Strontium ions in water are modeled as particles characterized by a charge of 3.2∙10-19 C and a diameter of 1.2∙10-8 m. The numerical model was created in Flow-3D software using the basic hydrodynamic module, electrostatic module, and general moving objects module. The performance of the studied reactor variants by numerical simulations is defined by the ratio of the number of model strontium particles permanently retained on the electrodes at the end of the simulation period to the initial number of particles in the water. For the laboratory reactor, the effect of strontium removal is defined by the ratio of the homogeneous strontium concentration in the water at the end and at the beginning of the experiments. The results show that the use of ultrasound increases the effect of strontium removal from 10.3% to 11.2% after 180 seconds of water treatment. The results of numerical simulations agree with the results of measurements on a reactor with the same geometrical characteristics.
Figure 1. US bath modified as an EC reactorFigure 2. Schematic view of the experimental set-up
References
Dong, B., Fishgold, A., Lee, P., Runge, K., Deymier, P. and Keswani, M. (2016), Sono-electrochemical recovery of metal ions from their aqueous solutions, Journal of Hazardous Materials, 318, 379–387. https://doi.org/10.1016/J.JHAZMAT.2016.07.007 EPA. (2014), Announcement of Final Regulatory Determinations for Contaminants on the Third Drinking Water Contaminant Candidate List. Retrieved from http://fdsys.gpo.gov/fdsys/search/home.action Fu, F., Lu, J., Cheng, Z. and Tang, B. (2016), Removal of selenite by zero-valent iron combined with ultrasound: Se(IV) concentration changes, Se(VI) generation, and reaction mechanism, Ultrasonics Sonochemistry, 29, 328–336. https://doi.org/10.1016/j.ultsonch.2015.10.007 Ince, N.H. (2018), Ultrasound-assisted advanced oxidation processes for water decontamination, Ultrasonics Sonochemistry, 40, 97–103. https://doi.org/10.1016/j.ultsonch.2017.04.009 Kamaraj, R. and Vasudevan, S. (2015), Evaluation of electrocoagulation process for the removal of strontium and cesium from aqueous solution, Chemical Engineering Research and Design, 93, 522–530. https://doi.org/10.1016/j.cherd.2014.03.021 Luczaj, J. and Masarik, K. (2015), Groundwater Quantity and Quality Issues in a Water-Rich Region: Examples from Wisconsin, USA, Resources, 4(2), 323–357. https://doi.org/10.3390/resources4020323 Mohapatra, D.P. and Kirpalani, D.M. (2019), Selenium in wastewater: fast analysis method development and advanced oxidation treatment applications, Water Science and Technology: A Journal of the International Association on Water Pollution Research, 79(5), 842–849. https://doi.org/10.2166/wst.2019.010
Mollah, M.Y.A., Schennach, R., Parga, J.R. and Cocke, D.L.(2001), Electrocoagulation (EC)- Science and applications, Journal of Hazardous Materials, 84(1), 29–41. https://doi.org/10.1016/S0304-3894(01)00176-5
Moradi, M., Vasseghian, Y., Arabzade, H. and Khaneghah, A.M. (2021), Various wastewaters treatment by sonoelectrocoagulation process: A comprehensive review of operational parameters and future outlook, Chemosphere, 263, 128314. https://doi.org/10.1016/J.CHEMOSPHERE.2020.12831 4 Peng, H., Yao, F., Xiong, S., Wu, Z., Niu, G. and Lu, T. (2021), Strontium in public drinking water and associated public health risks in Chinese cities, Environmental Science and Pollution Research International, 28(18), 23048. https://doi.org/10.1007/S11356-021-12378-Y Scott, V., Juran, L., Ling, E.J., Benham, B. and Spiller, A. (2020), Assessing strontium and vulnerability to strontium in private drinking water systems in Virginia, Water, 12(4). https://doi.org/10.3390/w12041053 Ziylan, A., Koltypin, Y., Gedanken, A. and Ince, N.H. (2013), More on sonolytic and sonocatalytic decomposition of Diclofenac using zero-valent iron, Ultrasonics Sonochemistry, 20(1), 580–586. https://doi.org/10.1016/j.ultsonch.2012.05.00
Analysis on inundation characteristics by compound external forces in coastal areas
연안 지역의 복합 외력에 의한 침수 특성 분석
Taeuk Kanga, Dongkyun Sunb, Sangho Leec* 강 태욱a, 선 동균b, 이 상호c*
aResearch Professor, Disaster Prevention Research Institute, Pukyong National University, Busan, Korea bResearcher, Disaster Prevention Research Institute, Pukyong National University, Busan, Korea cProfessor, Department of Civil Engineering, Pukyong National University, Busan, Korea a부경대학교 방재연구소 전임연구교수 b부경대학교 방재연구소 연구원 c부경대학교 공과대학 토목공학과 교수 *Corresponding Author
ABSTRACT
연안 지역은 강우, 조위, 월파 등 여러가지 외력에 의해 침수가 발생될 수 있다. 이에 이 연구에서는 연안 지역에서 발생될 수 있는 단일 및 복합 외력에 의한 지역별 침수 특성을 분석하였다. 연구에서 고려한 외력은 강우와 폭풍 해일에 의한 조위 및 월파이고, 분석 대상지역은 남해안 및 서해안의 4개 지역이다. 유역의 강우-유출 및 2차원 지표면 침수 분석에는 XP-SWMM이 사용되었고, 폭풍 해일에 의한 외력인 조위 및 월파량 산정에는 ADCSWAN (ADCIRC와 UnSWAN) 모형과 FLOW-3D 모형이 각각 활용되었다. 단일 외력을 이용한 분석 결과, 대부분의 연안 지역에서는 강우에 의한 침수 영향보다 폭풍 해일에 의한 침수 영향이 크게 나타났다. 복합 외력에 의한 침수 분석 결과는 대체로 단일 외력에 의한 침수 모의 결과를 중첩시켜 나타낸 결과와 유사하였다. 다만, 특정 지역에서는 복합 외력을 고려함에 따라 단일 외력만을 고려한 침수모의에서 나타나지 않았던 새로운 침수 영역이 발생하기도 하였다. 이러한 지역의 침수 피해 저감을 위해서는 복합 외력을 고려한 분석이 요구되는 것으로 판단되었다.
키워드
연안 지역
침수 분석
강우
폭풍 해일
복합 외력
The various external forces can cause inundation in coastal areas. This study is to analyze regional characteristics caused by single or compound external forces that can occur in coastal areas. Storm surge (tide level and wave overtopping) and rainfall were considered as the external forces in this study. The inundation analysis were applied to four coastal areas, located on the west and south coast in Republic of Korea. XP-SWMM was used to simulate rainfall-runoff phenomena and 2D ground surface inundation for watershed. A coupled model of ADCIRC and SWAN (ADCSWAN) was used to analyze tide level by storm surge and the FLOW-3D model was used to estimate wave overtopping. As a result of using a single external force, the inundation influence due to storm surge in most of the coastal areas was greater than rainfall. The results of using compound external forces were quite similar to those combined using one external force independently. However, a case of considering compound external forces sometimes created new inundation areas that didn’t appear when considering only a single external force. The analysis considering compound external forces was required to reduce inundation damage in these areas.
우리나라는 반도에 위치하여 삼면이 바다로 둘러싸여 있는 지리적 특성을 가지고 있다. 이에 따라 해양 산업을 중심으로 부산, 인천, 울산 등 대규모의 광역도시가 발달하였을 뿐만 아니라, 창원, 포항, 군산, 목포, 여수 등의 중․소규모 도시들도 발달되어 있다. 또한, 최근에는 연안 지역이 바다를 전망으로 하는 입지 조건을 가지고 있어 개발 선호도가 높고, 이에 따라 부산시 해운대의 마린시티, 엘시티와 같은 주거 및 상업시설의 개발이 지속되고 있다(Kang et al., 2019b).
한편, 최근 기후변화에 따른 지구 온난화 현상으로 평균 해수면이 상승하고, 해수면 온도도 상승하면서 태풍 및 강우의 강도가 커지고 있어 전 세계적으로 자연 재해로 인한 피해가 증가하고 있다(Kim et al., 2016). 실제로 2020년에는 최장기간의 장마가 발생하여 부산, 울산은 물론, 전국에서 50명의 인명 피해와 3,489세대의 이재민이 발생하였다1). 특히, 연안 지역은 강우, 만조 시 해수면 상승, 폭풍 해일(storm surge)에 의한 월파(wave overtopping) 등 복합적인 외력(compound external forces)에 의해 침수될 수 있다(Lee et al., 2020). 일례로, 2016년 태풍 차바 시 부산시 해운대구의 마린시티는 강우와 폭풍 해일에 의한 월파가 발생함에 따라 대규모 침수를 유발하였다(Kang et al., 2019b). 또한, 2020년 7월 23일에 부산에서는 시간당 81.6 mm의 집중호우와 약최고고조위를 상회하는 만조가 동시에 발생하였고, 이로 인해 감조 하천인 동천의 수위가 크게 상승하여 하천이 범람하였다(KSCE, 2021).
연안 지역의 복합 외력을 고려한 침수 분석에 관한 사례로서, 우선 강우와 조위를 고려한 연구 사례는 다음과 같다. Han et al. (2014)은 XP-SWMM을 이용하여 창원시 배수 구역을 대상으로 침수 모의를 수행하였는데, 연안 도시의 침수 모의에는 조위의 영향을 반드시 고려해야 함을 제시하였다. Choi et al. (2018a)은 경남 사천시 선구동 일대에 대하여 초과 강우 및 해수면 상승 시나리오를 조합하여 침수 분석을 수행하였다. Choi et al. (2018b)은 XP-SWMM을 이용하여 여수시 연등천 및 여수시청 지역에 대하여 강우 시나리오와 해수위 상승 시나리오를 고려한 복합 원인에 의한 침수 모의를 수행하여 홍수예경보 기준표를 작성하였다. 한편, 강우, 조위, 월파를 고려한 연구 사례로서, Song et al. (2017)은 부산시 해운대구 수영만 일원에 대하여 XP-SWMM으로 월파량의 적용 유무에 따른 침수 면적을 비교하였다. Suh and Kim (2018)은 부산시 마린시티 지역을 대상으로 태풍 차바 때 EurOtop의 경험식을 ADSWAN에 적용하여 월파량을 반영하였다. Chen et al. (2017)은 TELEMAC-2D 및 SWMM을 기반으로 한 극한 강우, 월파 및 조위를 고려하여 중국 해안 원자력 발전소의 침수를 예측하고 분석하기 위한 결합 모델을 개발한 바 있다. 한편, Lee et al. (2020)은 수리‧수문학 분야와 해양공학 분야에서 사용되는 물리 모형의 기술적 연계를 통해 연안 지역의 침수 모의의 재현성을 높였다.
상기의 연구들은 공통적으로 연안 지역에 대하여 복합 외력을 고려했을 때 발생되는 침수 현상의 재현 또는 예측을 목적으로 수행되었다. 이 연구는 이와 차별하여 복합 외력을 고려하는 경우 나타날 수 있는 연안 지역의 침수 특성 분석을 목적으로 수행되었다. 이를 위해 단일 외력을 독립적으로 고려했을 때 발생되는 침수 양상과 동시에 고려하는 경우의 침수 현상을 비교, 분석하였다. 복합 외력에 의한 지역적 침수 특성 분석은 우리나라 남해안과 서해안에 위치한 4개 지역에 대하여 적용되었다.
연안 지역의 침수는 크게 세 가지의 메카니즘으로 발생될 수 있다. 우선, 연안 지역은 바다와 인접하고 있기 때문에 그 영향을 직접적으로 받는다. Kim (2018)에 의하면, 연안 지역의 침수는 폭풍 해일에 의해 상승한 조위와 월파로 인해 발생될 수 있다(Table 1). 특히, 경상남도의 창원과 통영, 인천광역시의 소래포구 어시장 등 남해안 및 서해안 지역의 일부는 백중사리, 슈퍼문(super moon) 등 만조 시 조위의 상승으로 인한 침수가 발생하는 지역이 존재한다(Kang et al., 2019a). 두 번째는 강우에 의한 내수 침수 발생이다. ME (2011)에서는 도시 지역의 우수 관거를 10 ~ 30년 빈도로 계획하도록 지정하고 있고, 펌프 시설은 30 ~ 50년 빈도의 홍수를 배수시킬 수 있도록 정하고 있다. 하지만 최근에는 기후변화의 영향으로 도시 지역 배수시설의 설계 빈도를 초과하는 강우가 빈번하게 나타나고 있다. 실제로 2016년의 태풍 차바 시 울산 기상관측소에 관측된 시간 최대 강우량은 106.0 mm로서, 이는 300년 빈도 이상의 강우량에 해당하였다(Kang et al., 2019a). 따라서 배수시설의 설계 빈도 이상의 강우는 연안 도시 지역의 침수를 유발할 수 있다. 세 번째, 하천이 인접한 연안 도시에서는 하천의 범람으로 인해 침수가 발생할 수 있다. 하천의 경우, 기본계획이 수립되기는 하지만, 설계 빈도를 상회하는 강우의 발생, 제방, 수문 등 홍수 방어시설의 기능 저하, 예산 등의 문제로 하천기본계획 이행의 지연 등에 의해 범람할 가능성이 존재한다.
Table 1.
Type of natural hazard damage in coastal areas (Kim, 2018)
Item
Risk factor
Facilities damage
∙ Breaking of coastal facilities by wave – Breakwater, revetment, lighters wharf etc. ∙ Local scouring at the toe of the structures by wave ∙ Road collapse by wave overtopping
Inundation damage
∙ Inundation damage by wave overtopping ∙ Inundation of coastal lowlands by storm surge
Erosion damage
∙ Backshore erosion due to high swell waves ∙ Shoreline changes caused by construction of coastal erosion control structure ∙ Sediment transport due to the construction of artificial structures
상기의 내용을 종합하면, 연안 지역은 조위 및 월파에 의한 침수, 강우에 의한 내수 침수, 하천 범람에 의한 침수로 구분될 수 있다. 이 연구에서는 폭풍 해일에 의한 조위 상승 및 월파와 강우를 연안 지역의 침수 유발 외력으로 고려하였다. 하천 범람의 경우, 상대적으로 사례가 희소하여 제외하였다.
2.2 복합 외력을 고려한 침수 모의 방법
이 연구에서는 조위 및 월파와 강우를 연안 지역의 침수 발생에 관한 외력 조건으로 고려하였다. 따라서 해당 외력 조건을 고려하여 침수 분석을 수행할 수 있어야 한다. 이와 관련하여 Lee et al. (2020)은 Fig. 1과 같이 수리‧수문 및 해양공학 분야에서 사용되는 물리 기반 모형의 연계를 통해 조위, 월파, 강우를 고려한 침수 분석 방법을 제시하였고, 이 연구에서는 해당 방법을 이용하였다.
Fig. 1.
Connection among the models for inundation analysis in coastal areas (Lee et al., 2020)
우선, 태풍에 의해 발생되는 폭풍 해일의 영향을 분석하기 위해서는 태풍에 의해 발생되는 기압 강하, 해상풍, 진행 속도 등을 고려하여 해수면의 변화 양상 및 조석-해일-파랑을 충분히 재현 가능해야 한다. 이 연구에서는 국내․외에서 검증 및 공인된 폭풍 해일 모형인 ADCIRC 모형과 파랑 모형인 UnSWAN이 결합된 ADCSWAN (coupled model of ADCIRC and UnSWAN)을 이용하였다. 정수압 가정의 ADCSWAN은 월파량 산정에 단순 경험식을 적용하는 단점이 있지만 넓은 영역을 모의할 수 있고, FLOW-3D는 해안선의 경계를 고해상도로 재현이 가능하다. 이에 연구에서는 먼 바다 영역에 대해서는 ADCSWAN을 이용하여 분석하였고, 연안 주변의 바다 영역과 월파량 산정에 대해서는 FLOW-3D 모형을 이용하였다. 한편, 연안 지역의 침수 모의를 위해서는 유역에서 발생하는 강우-유출 현상과 우수 관거 등의 배수 체계에 대한 분석이 가능해야 한다. 또한, 배수 체계로부터 범람한 물이 지표면을 따라 흘러가는 현상을 해석할 수 있어야 하고, 바다의 조위 및 월파량을 경계조건으로 반영할 수 있어야 한다. 이 연구에서는 이러한 현상을 모의할 수 있고, 도시 침수 모의에 활용도가 높은 XP-SWMM을 이용하였다.
2.3 침수 분석 대상지역
연구의 대상지역은 조위 및 월파에 의한 침수와 강우에 의한 내수 침수의 영향이 복합적으로 발생할 수 있는 남해안과 서해안에 위치한 4개 지역이다. Table 2는 침수 분석 대상지역을 정리하여 나타낸 표이고, Fig. 2는 각 지역의 유역 경계를 나타낸 그림이다.
Table 2.
Target region for inundation analysis
Classification
Administrative district
Target region
Area (km2)
Main cause of inundation
Pump facility
Number of major outfall
The south coast
Haundae-gu, Busan
Marine City area
0.53
Wave overtopping
–
9
Haundae-gu, Busan
Centum City area
4.76
Poor interior drainage at high tide level
1
2
The west coast
Gunsan
Jungang-dong area
0.79
Poor interior drainage at high tide level
2
3
Boryeong
Ocheon Port area
0.41
High tide level
–
5
Fig. 2.
Watershed area
남해안의 분석 대상지역 중 부산시 해운대구의 마린시티는 바다 조망을 중심으로 조성된 주거지 및 상업시설 중심의 개발지역이다. 마린시티는 2016년 태풍 차바 및 2018년 태풍 콩레이 등 태풍 내습 시 월파에 의한 해수 월류로 인해 도로 및 상가 일부가 침수를 겪은 지역이다. 부산시 해운대구의 센텀시티는 과거 수영만 매립지였던 곳에 조성된 주거지 및 상업시설 중심의 신도시 지역이다. 센텀시티 유역의 북쪽은 해발고도 El. 634 m의 장산이 위치하는 등 산지 특성도 가지고 있어 상대적으로 유역 면적이 넓고, 배수시설의 규모도 크고 복잡하다. 하지만 수영강 하구의 저지대 지역에 위치함에 따라 강우 시 내수 배제가 불량하고, 특히 만조 시 침수가 잦은 지역이다.
서해안 분석 대상지역 중 전라북도 군산시의 중앙동 일원은 군산시 내항 내측에 조성된 구도시로서, 금강 및 경포천 하구에 위치하는 저지대이다. 이에 따라 군산시 풍수해저감종합계획에서는 해당 지역을 3개의 영역으로 구분하여 내수재해 위험지구(영동지구, 중동지구, 경암지구)로 지정하였고, 이 연구에서는 해당 지역을 모두 고려하였다. 한편, 군산시 중앙동 일원은 특히, 만조 시 내수 배제가 매우 불량하여 2개의 펌프시설이 운영되고 있다. 충청남도 보령시의 오천면에 위치한 오천항은 배후의 산지를 포함한 소규모 유역에 위치한다. 서해안의 특성에 따라 조석 간만의 차가 크고, 특히 태풍 내습 시 폭풍 해일에 의한 침수가 잦은 지역이다. 산지의 강우-유출수는 복개된 2개의 수로를 통해 바다로 배제되고, 상가들이 위치한 연안 주변 지역에는 강우-유출수 배제를 위한 3개의 배수 체계가 구성되어 있다.
3. 연구 결과
3.1 침수 모의 모형 구축
XP-SWMM을 이용하여 분석 대상지역별 침수 모의 모형을 구축하였다. 적절한 침수 분석 수행을 위해 지역별 수치지형도, 도시 공간 정보 시스템(urban information system, UIS), 하수 관망도 등의 수치 자료와 현장 조사를 통해 유역의 배수 체계를 구성하였다. 그리고 2차원 침수 분석을 위해 무인 드론 및 육상 라이다(LiDAR) 측량을 수행하여 평면해상도가 1 m 이하인 고해상도 수치지형모형(digital terrain model, DTM)을 구성하였고, 침수 모의 격자를 생성하였다.
Fig. 3은 XP-SWMM의 상세 구축 사례로서 부산시 마린시티 배수 유역에 대한 소유역 및 관거 분할 등을 통해 구성한 배수 체계와 고해상도 측량 결과를 이용하여 구성한 수치표면모형(digital surface model, DSM)을 나타낸다. Fig. 4는 각 대상지역에 대해 XP-SWMM을 이용하여 구축한 침수 모의 모형을 나타낸다. 침수 분석을 위해서는 침수 모의 영역에 대한 설정이 필요한데, 다수의 사전 모의를 통해 유역 내에서 침수가 발생되는 지역을 검토하여 결정하였다.
Fig. 3.
Analysis of watershed drainage system and high-resolution survey for Marine City
Fig. 4.
Simulation model for inundation analysis by target region using XP-SWMM
복합 외력에 의한 침수 영향을 검토하기 위해서는 외력 조건에 대한 빈도와 지속기간의 설정이 필요하다. 이 연구에서는 재해 현상이 충분히 나타날 수 있도록 강우와 조위 및 월파의 빈도를 모두 100년으로 설정하였다. 이때, 조위와 월파량의 산정에는 만조(약최고고조위) 시, 100년 빈도에 해당하는 태풍 내습에 따른 폭풍 해일의 발생 조건을 고려하였다.
지역별 강우 발생 특성과 유역 특성을 고려하기 위해 MOIS (2017)의 방재성능목표 기준에 따라 임계 지속기간을 결정하여 대상지역별 강우의 지속기간으로 설정하였다. 이때, 강우의 시간 분포는 MLTM (2011)의 Huff 3분위를 이용하였다. 그리고 조위와 월파의 경우, 일반적인 폭풍 해일의 지속기간을 고려하여 5시간으로 결정하였다. 한편, 침수 모의를 위한 계산 시간 간격, 2차원 모의 격자 등의 입력자료는 분석 대상지역의 유역 규모와 침수 분석 대상 영역을 고려하여 결정하였다. 참고로 침수 분석에 사용된 수치지형모형은 1 m 급의 고해상도로 구성되었지만, 2차원 침수 모의 격자의 크기는 지역별로 3 ~ 4 m이다. 이는 연구에서 사용된 XP-SWMM의 격자 수(100,000개) 제약에 따른 설정이나, Sun (2021)은 민감도 분석을 통해 2차원 침수 분석을 위한 적정 격자 크기를 3 ~ 4.5 m로 제시한 바 있다.
Table 3은 이 연구에서 설정한 침수 모의 조건과 분석 방법을 정리하여 나타낸 표이다.
Table 3.
Simulation condition and method
Classification
Target region
Simulation condition
Simulation method
Rainfall
Storm surge
Simulation time interval
2D grid size
Return period
Duration
Temporal distribution
Return period
Duration
Watershed routing
Channel routing
2D inundation
The south coast
Marine City area
100 yr
1 hr
3rd quartile of Huff’s method
100
5 hr
5 min
10 sec
1 sec
3 m
Centum City area
1 hr
100
5 min
10 sec
1 sec
4 m
The west coast
Jungang-dong area
2 hr
100
5 min
10 sec
1 sec
3.5 m
Ocheon Port area
1 hr
100
1 min
10 sec
1 sec
3 m
3.2.2 복합 재해의 동시 고려
이 연구의 대상지역들은 모두 소규모의 해안가 도시지역이고, 이러한 지역에 대한 강우의 임계지속기간은 1시간 ~ 2시간이나, 이 연구에서 분석한 폭풍 해일의 지속기간은 5시간으로 강우의 지속기간과 폭풍 해일의 지속기간이 상이하다. 이에 이 연구에서는 서로 다른 지속기간을 가진 강우와 폭풍 해일 또는 조위를 고려하기 위해 강우의 중심과 폭풍 해일의 중심이 동일한 시간에 위치하도록 설정하였다(Fig. 5).
XP-SWMM은 폭풍 해일이 지속되는 5시간 전체를 모의하도록 설정하였고, 폭풍 해일이 가장 큰 시점에 강우의 중심이 위치하도록 강우 발생 시기를 결정하였다. 다만, 부산 마린시티의 경우, 폭풍 해일에 의한 피해가 주로 월파에 의해 발생되므로 강우의 중심과 월파의 중심을 일치시켰고(Fig. 5(a)), 상대적으로 조위의 영향이 큰 3개 지역은 강우의 중심과 조위의 중심을 맞추었다. Fig. 5(b)는 군산시 중앙동 지역의 복합 외력에 의한 침수 분석에 사용된 강우와 조위의 조합이다.
한편, 100년 빈도의 확률강우량만을 고려한 침수 분석에서는 유역 유출부의 경계조건으로 우수 관거의 설계 조건을 고려하여 약최고고조위가 일정하게 유지되도록 설정하였다.
Fig. 5.
Consideration of external force conditions with different durations
3.2.3 XP-SWMM의 월파량 고려
XP-SWMM에 ADCSWAN 및 FLOW-3D 모형에 의해 산정된 월파량을 입력하기 위해 해안가 지역에 절점을 생성하여 월파 현상을 구현하였다. XP-SWMM에서 월파량을 입력하기 위한 절점의 위치는 FLOW-3D 모형에서 월파량을 산정한 격자의 중심 위치이다.
Fig. 6(a)는 마린시티 지역에 대한 월파량 입력 지점을 나타낸 것으로서, 유역 경계 주변에 동일 간격으로 원으로 표시한 지점들이 해당된다. Fig. 6(b)는 XP-SWMM에 월파량 입력 지점들을 반영하고, 하나의 절점에 월파량 시계열을 입력한 화면을 나타낸다.
Fig. 6.
Considering wave overtopping on XP-SWMM
3.3 침수 모의 결과
3.3.1 단일 외력에 의한 침수 모의 결과
Fig. 7은 단일 외력을 고려한 지역별 침수 모의 결과이다. 즉, Fig. 7의 왼쪽 그림들은 지역별로 100년 빈도 강우에 의한 침수 모의 결과를 나타내고, Fig. 7의 오른쪽 그림들은 만조 시 100년 빈도 폭풍 해일에 의한 침수 모의 결과이다. 대체로 강우에 의한 침수 영역은 유역 중․상류 지역의 유역 전반에 걸쳐 발생하였고, 폭풍 해일에 의한 침수 영역은 해안가 전면부에 위치하는 것을 볼 수 있다. 이는 폭풍 해일에 의한 조위 상승과 월파의 영향이 상류로 갈수록 감소하기 때문이다.
한편, 4개 지역 모두에서 공통적으로 강우에 비해 폭풍 해일에 의한 침수 영향이 상대적으로 크게 분석되었다. 이러한 결과는 연안 지역의 경우, 폭풍 해일에 대비한 침수 피해 저감 노력이 보다 중요함을 의미한다.
Fig. 7.
Simulation results by single external force (left: rainfall, right: storm surge)
3.3.2 복합 외력에 의한 침수 모의 결과
Fig. 8은 복합 외력을 고려한 지역별 침수 모의 결과이다. 즉, 강우 및 폭풍 해일을 동시에 고려함에 따라 발생된 침수 영역을 나타낸다. 복합 외력을 고려하는 경우, 단일 외력만을 고려한 분석 결과(Fig. 7)보다 침수 영역은 넓어졌고, 침수심은 깊어졌다.
복합 외력에 의한 침수 분석 결과는 대체로 단일 외력에 의한 침수 모의 결과를 중첩시켜 나타낸 결과와 유사하였고, 이는 일반적으로 예상할 수 있는 결과이다. 주목할만한 결과는 군산시 중앙동의 침수 분석에서 나타났다. 즉, 군산시 중앙동의 경우, 단일 외력만을 고려한 침수 모의 결과에서 나타나지 않았던 새로운 침수 영역이 발생하였다(Fig. 8(c)). 이와 관련된 상세 내용은 3.4절의 고찰에서 기술하였다.
Fig. 8.
Simulation results by compound external forces
3.4 결과 고찰
외력 조건별 침수의 영향을 정량적으로 비교하기 위해 침수 면적을 이용하였다. 이 연구에서는 강우만에 의해 유발된 침수 면적을 기준(기준값: 1)으로 하고, 폭풍 해일(조위+월파량)에 의한 침수 면적과 복합 외력에 의한 침수 면적의 상대적 비율로 분석하였다(Table 4).
Table 4.
Impact evaluation for inundation area by external force
Condition
Marine City, Busan
Centum City, Busan
Jungang-dong area, Gunsan
Ocheon Port area, Boryeong
Inundation area (km2)
Rate
Inundation area (km2)
Rate
Inundation area (km2)
Rate
Inundation area (km2)
Rate
Single external force
Rainfall (①)
0.0164
1.0
0.0759
1.0
0.0457
1.0
0.0175
1.0
Storm surge (②)
0.0363
2.21
0.0685
0.90
0.1463
3.20
0.0412
2.35
Compound external forces
Combination (①+②)
0.0524
3.19
0.1505
1.98
0.2632
5.76
0.0473
2.70
분석 결과, 부산 센텀시티를 제외한 3개 지역은 모두 폭풍 해일에 의한 침수 면적이 강우에 의한 침수 면적에 비해 2.2 ~ 3.2배 넓은 것으로 분석되었다. 한편, 복합 외력에 의한 침수 면적은 마린시티와 센텀시티의 경우, 각각의 외력에 의한 침수 면적의 합과 유사하게 나타났다. 이는 각각의 외력에 의한 침수 영역이 상이하여 거의 중복되지 않음을 의미한다. 반면에, 오천항에서는 각각의 외력에 의한 침수 면적의 합이 복합 외력에 의한 면적보다 크게 나타났다. 이는 오천항의 경우, 유역면적이 작고 배수 체계가 비교적 단순하여 강우와 폭풍 해일에 의한 침수 영역이 중복되기 때문인 것으로 분석되었다(Fig. 7(d)).
군산시 중앙동 일대의 경우, 복합 외력에 의한 침수 면적이 각각의 독립적인 외력 조건에 의한 침수 면적의 합에 비해 37.1% 크게 나타났다. 이러한 현상의 원인을 분석하기 위해 복합 외력 조건에서만 나타난 우수 관거(Fig. 8(c)의 A 구간)에 대하여 종단을 검토하였다(Fig. 9). Fig. 9(a)는 강우만에 의해 분석된 우수 관거 내 흐름 종단을 나타내고, Fig. 9(b)는 폭풍 해일만에 의한 우수 관거의 종단이다. 그림을 통해 각각의 독립적인 외력 조건 하에서는 해당 구간에서 침수가 발생되지 않은 것을 볼 수 있다. 다만, 강우만을 고려하더라도 우수 관거는 만관이 된 상태를 확인할 수 있다(Fig. 9(a)). 반면에, 만관 상태에서 폭풍 해일이 함께 고려됨에 따라 해수 범람과 조위 상승에 의해 우수 배제가 불량하게 되었고, 이로 인해 침수가 유발된 것으로 분석되었다(Fig. 9(c)). 따라서 이러한 지역은 복합 외력에 대한 취약지구로 판단할 수 있고, 단일 외력의 고려만으로는 침수를 예상하기 어려운 지역임을 알 수 있다.
Fig. 9.
A part of drainage profiles by external force in Jungang-dong area, Gunsan
4. 결 론
이 연구에서는 외력 조건에 따른 연안 지역의 침수 특성을 분석하였다. 연구에서 고려된 외력 조건은 두 가지로서 강우와 폭풍 해일(조위와 월파)이다. 분석 대상 연안 지역으로는 남해안에 위치하는 2개 지역(부산시 해운대구의 마린시티와 센텀시티)과 서해안의 2개 지역(군산시 중앙동 일원 및 보령시 오천항)이 선정되었다.
복합 외력을 고려한 연안 지역의 침수 모의를 위해서는 유역의 강우-유출 현상과 바다의 조위 및 월파량을 경계조건으로 반영할 수 있는 침수 모의 모형이 요구되는데, 이 연구에서는 XP-SWMM을 이용하였다. 한편, 조위 및 월파량 산정에는 ADCSWAN (ADCIRC와 UnSWAN) 및 FLOW-3D 모형이 이용되었다.
연안 지역별 침수 모의는 100년 빈도의 강우와 폭풍 해일을 독립적으로 고려한 경우와 복합적으로 고려한 경우를 구분하여 수행되었다. 우선, 외력을 독립적으로 고려한 결과, 대체로 폭풍 해일만 고려한 경우가 강우만 고려한 경우에 비해 침수 영향이 크게 나타났다. 따라서 연안 지역의 경우, 폭풍 해일에 의한 침수 피해 방지 계획이 상대적으로 중요한 것으로 분석되었다. 두 번째, 복합 외력에 의한 침수 분석 결과는 대체로 단일 외력에 의한 침수 모의 결과를 중첩시켜 나타낸 결과와 유사하였다. 다만, 특정 지역에서는 복합 외력을 고려함에 따라 단일 외력만을 고려한 침수 모의에서 나타나지 않았던 새로운 침수 영역이 발생하기도 하였다. 이러한 결과는 독립적인 외력 조건에서는 우수 관거가 만관 또는 그 이하의 상태가 되지만, 두 가지의 외력이 동시에 고려됨에 따라 우수 관거의 통수능 한계를 초과하여 나타났다. 이러한 지역은 복합 외력에 대한 취약지구로 판단되었고, 해당 지역의 적절한 침수 방지 대책 수립을 위해서는 복합적인 외력 조건이 고려되어야 함을 시사하였다.
현행, 자연재해저감종합계획에서는 침수와 관련된 재해 원인 지역을 내수재해, 해안재해, 하천재해 등으로 구분하고 있다. 하지만 이 연구에서 검토된 바와 같이, 연안 지역의 침수 원인은 복합적으로 나타날 뿐만 아니라, 복합 외력을 고려함에 따라 추가적으로 나타날 수 있는 침수 위험 지역도 존재한다. 따라서 기존의 획일적인 재해 원인의 구분보다는 지역의 특성에 맞는 복합적인 재해 원인을 검토할 필요가 있음을 제안한다.
Acknowledgements
본 논문은 행정안전부 극한 재난대응 기반기술 개발사업의 일환인 “해안가 복합재난 위험지역 피해저감 기술개발(연구과제번호: 2018-MOIS31-008)”의 지원으로 수행되었습니다.
References
1
Chen, X., Ji, P., Wu, Y., and Zhao, L. (2017). “Coupling simulation of overland flooding and underground network drainage in a coastal nuclear power plant.” Nuclear Engineering and Design, Vol. 325, pp. 129-134. 10.1016/j.nucengdes.2017.09.028
2
Choi, G., Song, Y., and Lee, J. (2018a). “Analysis of flood occurrence type according to complex characteristics of coastal cities.” 2018 Conference of the Korean Society of Hazard Mitigation, KOSHAM, p. 180.
3
Choi, J., Park, K., Choi, S., and Jun, H. (2018b). “A forecasting and alarm system for reducing damage from inland inundation in coastal urban areas: A case study of Yeosu City.” Journal of Korean Society of Hazard Mitigation, Vol. 18, No. 7, pp. 475-484. 10.9798/KOSHAM.2018.18.7.475
4
Han, H., Kim, Y., Kang, N., and, Kim, H.S. (2014). “Inundation analysis of a coastal urban area considering tide level.” 2014 Conference of Korean Society of Civil Engineers, KSCE, pp. 1507-1508.
5
Kang, T., Lee, S., and Sun, D. (2019a). “A technical review for reducing inundation damage to high-rise and underground-linked complex buildings in Coastal Areas (1): Proposal for analytical method.” Journal of Korean Society of Hazard Mitigation, Vol. 19, No. 5, pp. 35-43. 10.9798/KOSHAM.2019.19.5.35
6
Kang, T., Lee, S., Choi, H., and Yoon, S. (2019b). “A technical review for reducing inundation damage to high-rise and underground-linked complex buildings in coastal areas (2): Case analysis for application.” Journal of Korean Society of Hazard Mitigation, Vol. 19, No. 5, pp. 45-53. 10.9798/KOSHAM.2019.19.5.45
7
Kim, J.O., Kim, J.Y., and Lee, W.H. (2016). “Analysis on complex disaster information contents for building disaster map of coastal cities.” Journal of the Korean Association of Geographic Information Studies, Vol. 19, No. 3, pp. 43-60. 10.11108/kagis.2016.19.3.043
8
Kim, P.J. (2018). Improvement measures on the risk area designation of coastal disaster in consideration of natural hazards. Ph.D. dissertation, Chonnam National University.
9
Korean Society of Civil Engineers (KSCE) (2021). A report on the cause analysis and countermeasures establishment for Dongcheon flooding and lowland inundation. Busan/Ulsan, Gyungnam branch.
10
Lee, S., Kang, T., Sun, D., and Park, J.J. (2020). “Enhancing an analysis method of compound flooding in coastal areas by linking flow simulation models of coasts and watershed.” Sustainability, Vol. 12, No. 16, 6572. 10.3390/su12166572
11
Ministry of Environment (ME) (2011). Standard for sewerage facilities. Korea Water and Wastewater Works Association.
12
Ministry of Land, Transport and Maritime Affairs (MLTM) (2011). Improvement and complementary research for probability rainfall.
13
Ministry of the Interior and Safety (MOIS) (2017). Criteria for establishment and operation of disaster prevention performance target by region: Considering future climate change impacts.
14
Song, Y., Joo, J., Lee, J., and Park, M. (2017). “A study on estimation of inundation area in coastal urban area applying wave overtopping.” Journal of Korean Society of Hazard Mitigation, Vol. 17, No. 2, pp. 501-510. 10.9798/KOSHAM.2017.17.2.501
15
Suh, S.W., and Kim, H.J. (2018). “Simulation of wave overtopping and inundation over a dike caused by Typhoon Chaba at Marine City, Busan, Korea.” Journal of Coastal Research, Vol. 85, pp. 711-715.
16
Sun, D. (2021). Sensitivity analysis of XP-SWMM for inundation analysis in coastal area. M.Sc. Thesis, Pukyong National University.
Received: 10 June 2023 / Revised: 19 June 2023 / Accepted: 27 June 2023 / Published: 1 July 2023(This article belongs to the Section Ocean Engineering)
일부 수식이 손상되어 표시될 수 있습니다. 이 경우 원문을 참조하시기 바랍니다.
Abstract
Local scouring might result in the spanning of submarine cables, endangering their mechanical and electrical properties. In this contribution, a three-dimensional computational fluid dynamics simulation model is developed using FLOW-3D, and the scouring process of semi-exposed submarine cables is investigated. The effects of the sediment critical Shields number, sediment density, and ocean current velocity on local scouring are discussed, and variation rules for the submarine cables’ spanning time are provided. The results indicate that three scouring holes are formed around the submarine cables. The location of the bottom of the holes corresponds to that of the maximum shear velocity. The continuous development of scouring holes at the wake position leads to the spanning of the submarine cables. The increase in the sediment’s critical Shields number and sediment density, as well as the decrease in the ocean current velocity, will extend the time for maintaining the stability of the upstream scouring hole and retard the development velocity of the wake position and downstream scouring holes. The spanning time has a cubic relationship with the sediment’s critical Shields number, a linear relationship with the sediment density, and an exponential relationship with the ocean current velocity. In this paper, the local scouring process of semi-exposed submarine cables is studied, which provides a theoretical basis for the operation and maintenance of submarine cables.
As a key piece of equipment in cross-sea power grids, submarine cables are widely used to connect autonomous power grids, supply power to islands or offshore platforms, and transmit electric power generated by marine renewable energy installations to onshore substations [1]. Once submarine cables break down due to natural disasters or human-made damage, the normal operation of other marine electric power equipment connected to them may be affected. These chain reactions will cause great economic losses and serious social impacts [2].
To protect submarine cables, they are usually buried 1 to 3 m below the seabed [3]. However, submarine cables are still confronted with potential threats from the complex subsea environment. Under the influence of fishing, anchor damage, ocean current scouring, and other factors, the sediment above submarine cables will always inevitably migrate. When a submarine cable is partially exposed, the scouring at this position will be exacerbated; eventually, it will cause the submarine cable to span. According to a field investigation of the 500 kV oil-filled submarine cable that is part of the Hainan networking system, the total length of the span is 49 m [4]. Under strong ocean currents, spanning submarine cables may experience vortex-induced vibrations. Fatigue stress caused by vortex-induced vibrations may lead to metal sheath rupture [5], which endangers the mechanical and electrical properties of submarine cables. Therefore, understanding the local scouring processes of partially exposed submarine cables is crucial for predicting scouring patterns. This is the basis for developing effective operation and maintenance strategies for submarine cables.
The mechanism and influencing factors of sediment erosion have been examined by researchers around the world. In 1988, Sumer [6] conducted experiments to show that the shedding vortex in the wake of a pipeline would increase the Shields parameter by 3–4 times, which would result in severe scouring. In 1991, Chiew [7] performed experiments to prove that the maximum scouring depth could be obtained when the pipeline was located on a flat bed and was scoured by a unidirectional water flow. Based on the test results, they provided a prediction formula for the maximum scouring depth. In 2003, Mastbergen [8] proposed a one-dimensional, steady-state numerical model of turbidity currents, which considered the negative pore pressures in the seabed. The calculated results of this model were basically consistent with the actual scouring of a submarine canyon. In 2007, Dey [9] presented a semitheoretical model for the computation of the maximum clear-water scour depth below underwater pipelines in uniform sediments under a steady flow, and the predicted scour depth in clear water satisfactorily agreed with the observed values. In 2008, Dey [10] conducted experiments on clear-water scour below underwater pipelines under a steady flow and obtained a variation pattern of the depth of the scouring hole. In 2008, Liang [11] used a two-dimensional numerical simulation to study the scouring process of a tube bundle under the action of currents and waves. They discovered that, compared with the scouring of a single tube, the scouring depth of the tube bundle was deeper, and the scouring time was longer. In 2012, Yang [12] found that placing rubber sheets under pipes can greatly accelerate their self-burial. The rubber sheets had the best performance when their length was about 1.5 times the size of the pipe. In 2020, Li [13] investigated the two-dimensional local scour beneath two submarine pipelines in tandem under wave-plus-current conditions via numerical simulation. They found that for conditions involving waves plus a low-strength current, the scour pattern beneath the two pipelines behaved like that in the pure-wave condition. Conversely, when the current had equal strength to the wave-induced flow, the scour pattern beneath the two pipelines resembled that in the pure-current condition. In 2020, Guan [14] studied and discussed the interactive coupling effects among a vibrating pipeline, flow field, and scour process through experiments, and the experimental data showed that the evolution of the scour hole had significant influences on the pipeline vibrations. In 2021, Liu [15] developed a two-dimensional finite element numerical model and researched the local scour around a vibrating pipeline. The numerical results showed that the maximum vibration amplitude of the pipeline could reach about 1.2 times diameter, and the maximum scour depth occurred on the wake side of the vibrating pipeline. In 2021, Huang [16] carried out two-dimensional numerical simulations to investigate the scour beneath a single pipeline and piggyback pipelines subjected to an oscillatory flow condition at a KC number of 11 and captured typical steady-streaming structures around the pipelines due to the oscillatory flow condition. In 2021, Cui [17] investigated the characteristics of the riverbed scour profile for a pipeline buried at different depths under the condition of riverbed sediments with different particle sizes. The results indicated that, in general, the equilibrium scour depth changed in a spoon shape with the gradual increase in the embedment ratio. In 2022, Li [18] used numerical simulation to study the influence of the burial depth of partially buried pipelines on the surrounding flow field, but they did not investigate the scour depth. In 2022, Zhu [19] performed experiments to prove that the scour hole propagation rate under a pipeline decreases with an increasing pipeline embedment ratio and rises with the KC number. In 2022, Najafzadeh [20] proposed equations for the prediction of the scouring propagation rate around pipelines due to currents based on a machine learning model, and the prediction results were consistent with the experimental data. In 2023, Ma [21] used the computational fluid dynamics coarse-grained discrete element method to simulate the scour process around a pipeline. The results showed that this method can effectively reduce the considerable need for computing resources and excessive computation time. In 2023, through numerical simulations, Hu [22] discovered that the water velocity and the pipeline diameter had a significant effect on the depth of scouring.
In the preceding works, the researchers investigated the mechanism of sediment scouring and the effect of various factors on the local scouring of submarine pipelines. However, submarine cables are buried beneath the seabed, while submarine pipelines are erected above the seabed. The difference in laying methods leads to a large discrepancy between their local scouring processes. Therefore, the conclusions of the above investigations are not applicable to the local scouring of submarine cables. Currently, there is no report on the research of the local scouring of partially exposed submarine cables.
In this paper, a three-dimensional computational fluid dynamics (CFD) finite element model, based on two-phase flow, is established using FLOW-3D. The local scouring process of semi-exposed submarine cables under steady-state ocean currents is studied, and the variation rules of the depth and the shape of the scouring holes, as well as the shear velocity with time, are obtained. By setting different critical Shields numbers of the sediment, different sediment densities, and different ocean current velocities, the change rule of the scouring holes’ development rate and the time required for the spanning of submarine cables are explored.
2. Sediment Scouring Model
In the sediment scouring model, the sediment is set as the dispersed particle, which is regarded as a kind of quasifluid. In this context, sediment scouring is considered as a two-phase flow process between the liquid phase and solid particle phase. The sediment in this process is further divided into two categories: one is suspended in the fluid, and the other is deposited on the bottom.When the local Shields number of sediment is greater than the critical Shields number, the deposited sediment will be transformed into the suspended sediment under the action of ocean currents. The calculation formulae of the local Shields numbers θ and the critical Shields numbers
ρf is the fluid density, g is the acceleration of gravity, d
50 is the median size of sediment, and μ is the dynamic viscosity of sediment.And each sediment particle suspended in the fluid obeys the equations for mass conservation and energy conservation
𝑢�¯ is the mean velocity vector of the fluid and the sediment particle,
us is the velocity of the sediment particle,
fs is the volume fraction of the sediment particle, P is the pressure, F is the volumetric and viscous force, K is the drag force, and
ur is the relative velocity.
3. Numerical Setup and Modeling
In this paper, a three-dimensional submarine cable local scouring simulation model is established by FLOW-3D. Based on the numerical simulation, the process of the submarine cable, which gradually changes from semi-exposed to the spanning state under the steady-state ocean current, is studied. The geometric modeling, the mesh division, the physical field setup, and the grid independent test of CFD numerical model are as follows.
3.1. Geometric Modeling and Mesh Division
A three-dimensional (3D) numerical model of the local scouring of a semi-exposed submarine cable is established, which is shown in Figure 1. The dimensions of the model are marked in Figure 1. The inlet direction of the ocean current is defined as the upstream of the submarine cable (referred to as upstream), and the outlet direction of the ocean current is defined as the downstream of the submarine cable (referred to as downstream).
Figure 1. Three-dimensional finite element model of local scouring of semi-exposed submarine cable.
The submarine cable with a diameter of 0.2 m is positioned on sediment that is initially in a semi-exposed state. When the length of the span is short, the submarine cable will not show obvious deformation due to gravity or scouring from the ocean current. Therefore, the submarine cable surface is set as the fixed boundary. The model’s left boundary is set as the inlet, the right boundary is set as the outlet, the front and rear boundaries are set as symmetry, and the bottom boundary is set as the non-slip wall. Since the water depth above the submarine cable is more than 0.6 m in practice, the top boundary of the model is also set as symmetry. The sediment near the inlet and the outlet will be carried by ocean currents, which leads to the abnormal scouring terrain. At each end of the sediment, a baffle (thickness of 3 cm) is installed to ensure that the simulation results can reflect the real situation.
Due to the fact that the flow field around the semi-exposed submarine cable is not a simple two-dimensional symmetrical distribution, it should be solved by three-dimensional numerical simulation. Considering the accuracy and efficiency of the calculation, the size of mesh is set to 0.02 m. The total number of meshes after the dissection is 133,254.
3.2. Physical Field Setup
The CFD finite element model contains four physical field modules: sediment scouring module, gravity and non-inertial reference frame module, density evaluation module, and viscosity and turbulence module. In this paper, the renormalization group (RNG) k–ε turbulence model is used, which has high computational accuracy for turbulent vortices. Therefore, this turbulence model is suitable for calculating the sediment scouring process around the semi-exposed submarine cable [25]. The key parameters of the numerical simulation are referring to the survey results of submarine sediments in the Korean Peninsula [26], as listed in Table 1.Table 1. Key parameters of numerical simulation.
3.3. Mesh Independent Test
In order to eliminate errors caused by the quantity of grids in the calculation process, two sizes of mesh are set on the validation model, and the scour profiles under different mesh sizes are compared. The validation model is shown in Figure 2, and the scouring terrain under different mesh size is given in Figure 3.
Figure 2. Validation model.
Figure 3. Scouring terrain under different mesh sizes.
It can be seen from Figure 3 that with the increase in the number of meshes, the scouring terrain of the verification model changes slightly, and the scouring depth is basically unchanged. Considering the accuracy of the numerical simulation and the calculation’s time cost, it is reasonable to consider setting the mesh size to 0.02 m.
4. Results and Analysis
4.1. Analysis of Local Scouring Process
Based on the CFD finite element numerical simulation, the local scouring process of the submarine cable under the steady-state ocean current is analyzed. The end time of the simulation is 9 h, the initial time step is 0.01 s, and the fluid velocity is 0.40 m/s. Simulation results are saved every minute. Figure 4 illustrates the scouring terrain around the semi-exposed submarine cable, which has been scoured by the steady-state current for 5 h.
Figure 4. Scouring terrain around semi-exposed submarine cable (scour for 5 h).
As can be seen from Figure 4, three scouring holes were separately formed in the upstream wake position and downstream of the semi-exposed submarine cable. The scouring holes are labeled according to their locations. The variation of the scouring terrain around the semi-exposed submarine cable over time is given in Figure 5. The red circle in the picture corresponds to the position of the submarine cable, and the red box in the legend marks the time when the submarine cable is spanning.
Figure 5. Variation of scouring terrain around semi-exposed submarine cable adapted to time.
From Figure 5, in the first hour of scouring, the upstream (−0.5 m to −0.1 m) and downstream (0.43 m to 1.5 m) scouring holes appeared. The upstream scouring hole was relatively flat with depth of 0.04 m. The depth of the downstream scouring hole increased with the increase in distance, and the maximum depth was 0.13 m. The scouring hole that developed at the wake position was very shallow, and its depth was only 0.007 m.
In the second hour of scouring, the upstream scouring hole’s depth remained nearly constant. The depth of the downstream scouring hole only increased by 0.002 m. The scouring hole at the wake position developed steadily, and its depth increased from 0.007 m to 0.014 m.
The upstream and downstream scouring holes did not continue to develop during the third to the sixth hour. Compared to the first two hours, the development of scouring holes at the wake position accelerated significantly, with an average growth rate of 0.028 m/h. The growth rate in the fifth hour of the scouring hole at the wake position was slightly faster than the other times. After 6 h of scouring, the sediment on the right side of the submarine cable had been hollowed out.
In the seventh and the eighth hour of scouring, the upstream scouring hole’s depth increased slightly, the downstream scouring hole still remained stable, and the depth of the scouring hole at wake position increased by 0.019 m. The sediment under the submarine cable was gradually eroded as well. By the end of the eighth hour, the lower right part of the submarine cable had been exposed to water as well.
At 8 h 21 min of the scouring, the submarine cable was completely spanned, and the scouring holes were connected to each other. Within the next 10 min, the development of the scouring holes sped up significantly, and the maximum depth of scouring holes increased greatly to 0.27 m.
In reference [17], researchers have studied the local scouring process of semi-buried pipelines in sandy riverbeds through experiments. The test results show that the scouring process can be divided into a start-up stage, micropore formation stage, extension stage, and equilibrium stage. In this paper, the first three stages are simulated, and the results are in good agreement with the experiment, which proves the accuracy of the present numerical model.
In this research, the velocity of ocean currents at the sediment surface is defined as the shear velocity, which plays an important role in the process of local scouring. Figure 6 provides visual data on how the shear velocity varies over time.
Figure 6. Shear velocity changes in the scouring process.
The semi-exposed submarine cable protrudes from the seabed, which makes the shear velocity of its surface much higher than other locations. After the submarine cable is spanned, the shear velocity of the scouring hole surface below it is taken. This is the reason for the sudden change of shear velocity at the submarine cable’s location in Figure 6.The shear velocity in the initial state of the upstream scouring hole is obviously greater than in subsequent times. After 1 h of scouring, the shear velocity in the upstream scouring hole rapidly decreased from 1.1 × 10
−2 m/s to 3.98 × 10
−3 m/s and remained stable until the end of the sixth hour. This phenomenon explains why the upstream scouring hole developed rapidly in the first hour but remained stable for the following 5 h.The shear velocity in the downstream scouring hole reduced at first and then increased; its initial value was 1.41 × 10
−2 m/s. It took approximately 5 h for the shear velocity to stabilize, and the stable shear velocity was 2.26 × 10
−3 m/s. Therefore, compared with the upstream scouring hole, the downstream scouring hole was deeper and required more time to reach stability.The initial shear velocity in the scouring hole at the wake position was only 7.1 × 10
−3 m/s, which almost does not change in the first hour. This leads to a very slow development of the scouring hole at the wake position in the early stages. The maximum shear velocity in this scouring hole gradually increased to 1.05 × 10
−2 m/s from the second to the fifth hour, and then decreased to 6.61 × 10
−3 m/s by the end of the eighth hour. This is why the scouring hole at the wake position grows fastest around the fifth hour. Consistent with the pattern of change in the scouring hole’s terrain, the location of the maximal shear velocity also shifted to the right with time.
The shear velocity of all three scouring holes rose dramatically in the last hour. Combined with the terrain in Figure 5, this can be attributed to the complete spanning of the submarine cable.
From Equations (3)–(5), one can see the movement of the sediment is related directly with the sediment’s critical Shields number, sediment density, and ocean current velocity. Based on the parameters in Table 1, the influence of the above parameters on the local scouring process of semi-exposed submarine cables will be discussed.
4.2. Influence Factors
4.2.1. Sediment’s Critical Shields Number
The sediment’s critical Shields number
θcr is set as 0.02, 0.03, 0.04, 0.05, 0.06, and 0.07, and the variations of scouring terrain over time under each
Figure 7. Influence of sediment’s critical Shields number
θcr on local scouring around semi-exposed submarine cable: (a
) θcr = 0.02; (b
) θcr = 0.03; (c
) θcr = 0.04; (d
) θcr = 0.05; (e
) θcr = 0.06; and (f
) θcr = 0.07.From Figure 7, one can see that a change in
θcr will affect the depth of the upstream scouring hole and the development speed of the scouring hole at the wake position, but it will have no significant impact on the expansion of the downstream scouring hole.Under conditions of different
θcr, the upstream scouring hole will reach a temporary plateau within 1 h, at which time the stable depth will be about 0.04 m. When
θcr ≤ 0.05, the upstream scouring hole will continue to expand after a few hours. The stable time is obviously affected by
θcr, which will gradually increase from 1 h to 11 h with the increase in
θcr. The terrain of the upstream scouring hole will gradually convert to deep on the left and to shallow on the right. Since the scouring hole at the wake position has not been stable, its state at the time of submarine cable spanning is studied emphatically. In the whole process of scouring, the scouring hole at the wake position continues to develop and does not reach a stable state. With the increase in
θcr, the development velocity of the scouring hole at the wake position will decrease considerably. Its average evolution velocity decreases from 3.88 cm/h to 1.62 cm/h, and its depth decreases from 21.9 cm to 18.8 cm. Under the condition of each
θcr, the downstream scouring hole will stabilize within 1 h, and the stable depth will be basically unchanged (all about 13.5 cm).As
θcr increases, so does the sediment’s ability to withstand shearing forces, which will cause it to become increasingly difficult to be eroded or carried away by ocean currents. This effect has been directly reflected in the depth of scouring holes (upstream and wake position). Due to the blocking effect of semi-exposed submarine cables, the wake is elongated, which is why the downstream scouring hole develops before the scouring hole at the wake position and quickly reaches a stable state. However, due to the high wake intensity, this process is not significantly affected by the change of
θcr.
4.2.2. Sediment Density
The density of sediment
ρs is set as 1550 kg/m
3, 1600 kg/m
3, 1650 kg/m
3, 1700 kg/m
3, 1750 kg/m
3, and 1800 kg/m
3, and the variation of scouring terrain over time under each
ρs will also affect the depth of the upstream scouring hole and the development speed of the scouring hole at the wake position. In addition, it can even have an impact on the downstream scouring hole depth.Under different
ρs conditions, the upstream scouring hole will always reach a temporary stable state in 1 h, at which time the stable depth will be 0.04 m. When
ρs ≤ 1750 kg/m
3, the upstream scouring hole will continue to expand after a few hours. The stabilization time of upstream scouring hole is more clearly affected by
ρs, which will gradually increase from 3 h to 13 h with the increase in
ρs. The terrain of the upstream scouring hole will gradually change to deep on the left and to shallow on the right. Since the scouring hole at the wake position has not been stable, its state at the time of the submarine cable spanning is studied emphatically, too. In the whole process of scouring, the scouring hole at the wake position continues to develop and does not reach a stable state. When
ρs is large, the development rate of scouring hole obviously decreased with time. With the increase in
ρs, the development velocity of the scouring hole at the wake position reduces from 3.38 cm/h to 1.14 cm/h, and the depth of this scouring hole declines from 20 cm to 15 cm. As
ρs increases, the stabilization time of the downstream scouring hole increases from less than 1 h to about 2 h, but the stabilization depth of the downstream scouring hole remains essentially the same (all around 13.5 cm).As can be seen from Equation (1), the increase in
ρs will reduce the Shields number, thus weakening the shear action of the sediment by the ocean current, which explains the extension of the stability time of the upstream scouring hole. At the same time, with the increase in the depth of scouring hole at the wake position, its shear velocity will decreases. Therefore, under a larger
ρs value, the development speed of scouring hole at the wake position will decrease significantly with time. Possibly for the same reason,
ρs can affect the development rate of downstream scouring hole.
4.2.3. Ocean Current Velocity
The ocean current velocity v is set as 0.35 m/s, 0.40 m/s, 0.45 m/s, 0.50 m/s, 0.55 m/s, and 0.60 m/s. Figure 9 presents the variation in scouring terrain with time for each v.
Figure 9. Influence of ocean current velocity v on local scouring around semi-exposed submarine cable: (a) v = 0.35 m/s; (b) v = 0.40 m/s; (c) v = 0.45 m/s; (d) v = 0.50 m/s; (e) v = 0.55 m/s; and (f) v = 0.60 m/s.
Changes in v affect the depth of the upstream and downstream scouring holes, as well as the development velocity of the wake position and downstream scouring holes.
When v ≤ 0.45 m/s, the upstream scouring hole will reach a temporary stable state within 1 h, at which point the stable depth will be 0.04 m. The stabilization time of the upstream scouring hole is affected by v, which will gradually decrease from 15 h to 3 h with the increase in v. When v > 0.45 m/s, the upstream scouring hole is going to expand continuously. With the increase in v, its average development velocity increases from 6.68 cm/h to 8.66 cm/h, and its terrain changes to deep on the left and to shallow on the right. When the submarine cable is spanning, special attention should be paid to the depth of the scouring hole at the wake position. Throughout whole scouring process, the scouring hole at the wake position continues to develop and does not reach a stable state. With the increase in v, the depth of scouring hole at the wake position will increase from 14 cm to 20 cm, and the average development velocity will increase from 0.91 cm/h to 10.43 cm/h. As v increases, the time required to stabilize the downstream scouring hole is shortened from 1to 2 h to less than 1 h, but the stable depth is remains nearly constant at 13.5 cm.
An increase in v will increase the shear velocity. Therefore, when the depth of the scouring hole increases, the shear velocity in the hole will also increase, which can deepen both the upstream and downstream scouring hole. According to Equation (1), the Shields number is proportional to the square of the shear velocity. The increase in shear velocity significantly intensifies local scouring, which increases the development rate of scouring holes at the wake position and downstream.
4.3. Variation Rule of Spanning Time
In this paper, the spanning time is defined as the time taken for a semi-exposed submarine cable (initial state) to become a spanning submarine cable. Figure 10 illustrates the effect of the above parameters on the spanning time of the semi-exposed submarine cable.
Figure 10. Influence of different parameters on spanning time of the semi-exposed submarine cable: (a) Sediment critical Shields number; (b) Sediment density; and (c) Ocean current velocity.From Figure 10a, the spanning time monotonically increases with the increase in the critical Shields number of sediment. However, the slope of the curve decreases first and then increases, and the inflection point is at
θcr = 4.59 × 10
−2. The relationship between spanning time t and sediment’s critical Shields number
θcr can be formulated by a cubic function as shown in Equation (6):
𝑡=−2.98+6.76𝜃𝑐𝑟−1.45𝜃2𝑐𝑟+0.11𝜃3𝑐𝑟.�=−2.98+6.76���−1.45���2+0.11���3.(6)It can be seen from Figure 10b that with the increase in the sediment density, the spanning time increases monotonically and linearly. The relationship between the spanning time t and the sediment’s density
ρs can be formulated by the first order function as shown in Equation (7):
𝑡=−41.59+30.54𝜌𝑠.�=−41.59+30.54��.(7)Figure 10c shows that with the increase in the ocean current velocity, the spanning time decreases monotonically. The slope of the curve increases with the increase in the ocean current velocity, so it can be considered that there is saturation of the ocean current velocity effect. The relationship between the spanning time t and the ocean current velocity v can be formulated by the exponential function
𝑡=0.15𝑣−4.38.�=0.15�−4.38.(8)
5. Conclusions
In this paper, a three-dimensional CFD finite element numerical simulation model is established, which is used to research the local scouring process of the semi-exposed submarine cable under the steady-state ocean current. The relationship between shear velocity and scouring terrain is discussed, the influence of sediment critical Shields number, sediment density and ocean current velocity on the local scouring process is analyzed, and the variation rules of the spanning time of the semi-exposed submarine cable is given. The conclusions are as follows:
Under the steady-state ocean currents, scouring holes will be formed at the upstream, wake position and downstream of the semi-exposed submarine cable. The upstream and downstream scouring holes develop faster, which will reach a temporary stable state at about 1 h after the start of the scouring. The scouring hole at the wake position will continue to expand at a slower rate and eventually lead to the spanning of the submarine cable.
There is a close relationship between the distribution of shear velocity and the scouring terrain. As the local scouring process occurs, the location of the maximum shear velocity within the scouring hole shifts and causes the bottom of the hole to move as well.
When the sediment’s critical Shields number and density are significantly large and ocean current velocity is sufficiently low, the duration of the stable state of the upstream scouring hole will be prolonged, and the average development velocity of the scouring holes at the wake position and downstream will be reduced.
The relationship between the spanning time and the critical Shields number θcr can be formulated as a cubic function, in which the curve’s inflection point is θcr = 4.59 × 10−2. The relationship between spanning time and sediment density can be formulated as a linear function. The relationship between spanning time and ocean current velocity can be formulated by exponential function.
Based on the conclusions of this paper, even when it is too late to take measures or when the exposed position of the submarine cable cannot be located, the degree of burial depth development still can be predicted. This prediction is important for the operation and maintenance of the submarine cable. However, the study still leaves something to be desired. Only the local scouring process under the steady-state ocean current was studied, which is an extreme condition. In practice, exposed submarine cables are more likely to be scoured by reciprocating ocean currents. In the future, we will investigate the local scouring of submarine cables under the reciprocating ocean current.
Author Contributions
Conceptualization, Y.H. and Q.L.; methodology, Q.L., P.Z. and H.T.; software, Q.L.; validation, Q.L., L.C. and W.T.; writing—original draft preparation, Q.L.; writing—review and editing, Y.H. and Q.L.; supervision, Y.H. and L.Y. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the [Smart Grid Joint Fund Key Project between National Natural Science Foundation of China and State Grid Corporation] grant number [U1766220].
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data supporting the reported results cannot be shared at this time, as they have been used in producing more publications on this research.
Acknowledgments
This work is supported by the Smart Grid Joint Fund Key Project of the National Natural Science Foundation of China and State Grid Corporation (Grant No. U1766220).
Conflicts of Interest
The authors declare no conflict of interest.
References
Taormina, B.; Bald, J.; Want, A.; Thouzeau, G.; Lejart, M.; Desroy, N.; Carlier, A. A review of potential impacts of submarine power cables on the marine environment: Knowledge gaps, recommendations and future directions. Renew. Sust. Energ. Rev.2018, 96, 380–391. [Google Scholar] [CrossRef]
Gulski, E.; Anders, G.J.; Jogen, R.A.; Parciak, J.; Siemiński, J.; Piesowicz, E.; Paszkiewicz, S.; Irska, I. Discussion of electrical and thermal aspects of offshore wind farms’ power cables reliability. Renew. Sust. Energ. Rev.2021, 151, 111580. [Google Scholar] [CrossRef]
Wang, W.; Yan, X.; Li, S.; Zhang, L.; Ouyang, J.; Ni, X. Failure of submarine cables used in high-voltage power transmission: Characteristics, mechanisms, key issues and prospects. IET Gener. Transm. Distrib.2021, 15, 1387–1402. [Google Scholar] [CrossRef]
Chen, H.; Chen, Z.; Lu, H.; Wu, C.; Liang, J. Protection method for submarine cable detection and exposed suspension problem in Qiongzhou straits. Telecom Pow. Technol.2019, 36, 60–61+63. [Google Scholar]
Zhu, J.; Ren, B.; Dong, P.; Chen, W. Vortex-induced vibrations of a free spanning submarine power cable. Ocean Eng.2023, 272, 113792. [Google Scholar] [CrossRef]
Sumer, B.M.; Jensen, H.R.; Mao, Y.; Fredsøe, J. Effect of lee-wake on scour below pipelines in current. J. Waterw. Port Coast. Ocean. Eng.1988, 114, 599–614. [Google Scholar] [CrossRef]
Chiew, Y.M. Prediction of maximum scour depth at submarine pipelines. J. Hydraul. Eng.1991, 117, 452–466. [Google Scholar] [CrossRef]
Mastbergen, D.R.; Vandenberg, J.H. Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons. Sedimentology2003, 50, 625–637. [Google Scholar] [CrossRef]
Dey, S.; Singh, N.P. Clear-water scour below underwater pipelines under steady flow. J. Hydraul. Eng.2008, 134, 588–600. [Google Scholar] [CrossRef]
Liang, D.; Cheng, L. Numerical study of scour around a pipeline bundle. Proc. Inst. Civil Eng. Mar. Eng.2008, 161, 89–95. [Google Scholar] [CrossRef]
Yang, L.; Guo, Y.; Shi, B.; Kuang, C.; Xu, W.; Cao, S. Study of scour around submarine pipeline with a rubber plate or rigid spoiler in wave conditions. J. Waterw. Port Coast. Ocean Eng.2012, 138, 484–490. [Google Scholar] [CrossRef]
Li, Y.; Ong, M.C.; Fuhrman, D.R.; Larsen, B.E. Numerical investigation of wave-plus-current induced scour beneath two submarine pipelines in tandem. Coast. Eng.2020, 156, 103619. [Google Scholar] [CrossRef]
Guan, D.; Hsieh, S.C.; Chiew, Y.M.; Low, Y.M.; Wei, M. Local scour and flow characteristics around pipeline subjected to vortex-induced vibrations. J. Hydraul. Eng.2020, 146, 04019048. [Google Scholar] [CrossRef]
Liu, M.M.; Jin, X.; Wang, L.; Yang, F.; Tang, J. Numerical investigation of local scour around a vibrating pipeline under steady currents. Ocean Eng.2021, 221, 108546. [Google Scholar] [CrossRef]
Huang, J.; Yin, G.; Ong, M.C.; Myrhaug, D.; Jia, X. Numerical investigation of scour beneath pipelines subjected to an oscillatory flow condition. J. Mar. Sci. Eng.2021, 9, 1102. [Google Scholar] [CrossRef]
Cui, F.; Du, Y.; Hao, X.; Peng, S.; Bao, Z.; Peng, S. Experimental study on local scour and related mechanical effects at river-crossing underwater oil and gas pipelines. Adv. Civ. Eng.2021, 2021, 6689212. [Google Scholar] [CrossRef]
Li, B.; Ma, H. Scouring mechanism of suspended and partially-buried pipelines under steady flow. Coast. Eng.2022, 177, 104201. [Google Scholar] [CrossRef]
Najafzadeh, M.; Oliveto, G. Scour propagation rates around offshore pipelines exposed to currents by applying data-driven models. Water2022, 14, 493. [Google Scholar] [CrossRef]
Zhu, Y.; Xie, L.; Wong, T.; Su, T. Development of three-dimensional scour below pipelines in regular waves. J. Mar. Sci. Eng.2022, 10, 124. [Google Scholar] [CrossRef]
Ma, H.; Li, B. CFD-CGDEM coupling model for scour process simulation of submarine pipelines. Ocean Eng.2023, 271, 113789. [Google Scholar] [CrossRef]
Hu, K.; Bai, X.; Vaz, M.A. Numerical simulation on the local scour processing and influencing factors of submarine pipeline. J. Mar. Sci. Eng.2023, 11, 234. [Google Scholar] [CrossRef]
Yang, B.; Gao, F.; Wu, Y. Experimental study on local scour of sandy seabed under submarine pipeline in unidirectional currents. Eng. Mech.2008, 25, 206–210. [Google Scholar]
Cheng, Y.; Wang, X.; Luo, W.; Huang, X.; Lyu, X. Experimental study of local scour around a downstream inclined pile under combined waves and current. Adv. Eng. Sci.2021, 53, 64–71. [Google Scholar]
Lu, Y.; Zhou, L.; Shen, X. Different turbulence models for simulating a liquid-liquid hydro cyclone. J. Tsinghua Univ.2001, 41, 105–109. [Google Scholar]
Yun, D.H.; Kim, Y.T. Experimental study on settlement and scour characteristics of artificial reef with different reinforcement type and soil type. Geotext. Geomembr.2018, 46, 448–454. [Google Scholar] [CrossRef]
액체-증기 상 변화 모델은 밀폐된 용기의 자체 가압 프로세스 시뮬레이션에 매우 큰 영향을 미칩니다. Hertz-Knudsen 관계, 에너지 점프 모델 및 그 파생물과 같은 널리 사용되는 액체-증기 상 변화 모델은 실온 유체를 기반으로 개발되었습니다. 액체-증기 전이를 통한 극저온 시뮬레이션에 널리 적용되었지만 각 모델의 성능은 극저온 조건에서 명시적으로 조사 및 비교되지 않았습니다. 본 연구에서는 171가지 일반적인 액체-증기 상 변화 모델을 통합한 통합 다상 솔버가 제안되었으며, 이를 통해 이러한 모델을 실험 데이터와 직접 비교할 수 있습니다. 증발 및 응축 모델의 예측 정확도와 계산 속도를 평가하기 위해 총 <>개의 자체 가압 시뮬레이션이 수행되었습니다. 압력 예측은 최적화 전략이 서로 다른 모델 계수에 크게 의존하는 것으로 나타났습니다. 에너지 점프 모델은 극저온 자체 가압 시뮬레이션에 적합하지 않은 것으로 나타났습니다. 평균 편차와 CPU 소비량에 따르면 Lee 모델과 Tanasawa 모델은 다른 모델보다 안정적이고 효율적인 것으로 입증되었습니다.
Liquid-vapor phase change models vitally influence the simulation of self-pressurization processes in closed containers. Popular liquid-vapor phase change models, such as the Hertz-Knudsen relation, energy jump model, and their derivations were developed based on room-temperature fluids. Although they had widely been applied in cryogenic simulations with liquid-vapor transitions, the performance of each model was not explicitly investigated and compared yet under cryogenic conditions. A unified multi-phase solver incorporating four typical liquid-vapor phase change models has been proposed in the present study, which enables direct comparison among those models against experimental data. A total number of 171 self-pressurization simulations were conducted to evaluate the evaporation and condensation models’ prediction accuracy and calculation speed. It was found that the pressure prediction highly depended on the model coefficients, whose optimization strategies differed from each other. The energy jump model was found inadequate for cryogenic self-pressurization simulations. According to the average deviation and CPU consumption, the Lee model and the Tanasawa model were proven to be more stable and more efficient than the others.
Introduction
The liquid-vapor phase change of cryogenic fluids is widely involved in industrial applications, such as the hydrogen transport vehicles [1], shipborne liquid natural gas (LNG) containers [2] and on-orbit cryogenic propellant tanks [3]. These applications require cryogenic fluids to be stored for weeks to months. Although high-performance insulation measures are adopted, heat inevitably enters the tank via radiation and conduction. The self-pressurization in the tank induced by the heat leakage eventually causes the venting loss of the cryogenic fluids and threatens the safety of the craft in long-term missions. To reduce the boil-off loss and extend the cryogenic storage duration, a more comprehensive understanding of the self-pressurization mechanism is needed.
Due to the difficulties and limitations in implementing cryogenic experiments, numerical modeling is a convenient and powerful way to study the self-pressurization process of cryogenic fluids. However, how the phase change models influence the mass and heat transfer under cryogenic conditions is still unsettled [4]. As concluded by Persad and Ward [5], a seemingly slight variation in the liquid-vapor phase change models can lead to erroneous predictions.
Among the liquid-vapor phase change models, the kinetic theory gas (KTG) based models and the energy jump model are the most popular ones used in recent self-pressurization simulations [6]. The KTG based models, also known as the Hertz-Knudsen relation models, were developed on the concept of the Maxwell-Boltzmann distribution of the gas molecular [7]. The Hertz-Knudsen relation has evolved to several models, including the Schrage model [8], the Tanasawa model [9], the Lee model [10] and the statistical rate theory (SRT) [11], which will be described in Section 2.2. Since the Schrage model and the Lee model are embedded and configured as the default ones in the commercial CFD solvers Flow-3D® and Ansys Fluent® respectively, they have been widely used in self-pressurization simulations for liquid nitrogen [12], [13] and liquid hydrogen [14], [15]. The major drawback of the KTG models lies in the difficulty of selecting model coefficients, which were reported in a considerably wide range spanning three magnitudes even for the same working fluid [16], [17], [18], [19], [20], [21]. Studies showed that the liquid level, pressure and mass transfer rate are directly influenced by the model coefficients [16], [22], [23], [24], [25]. Wrong coefficients will lead to deviation or even divergence of the results. The energy jump model is also known as the thermal limitation model. It assumes that the evaporation and condensation at the liquid-vapor interface are induced only by heat conduction. The model is widely adopted in lumped node simulations due to its simplicity [6], [26], [27]. To improve the accuracy of mass flux prediction, the energy jump model was modified by including the convection heat transfer [28], [29]. However, the convection correlations are empirical and developed mainly for room-temperature fluids. Whether the correlation itself can be precisely applied in cryogenic simulations still needs further investigation.
Fig. 1 summarizes the cryogenic simulations involving the modeling of evaporation and condensation processes in recent years. The publication has been increasing rapidly. However, the characteristics of each evaporation and condensation model are not explicitly revealed when simulating self-pressurization. A comparative study of the phase change models is highly needed for cryogenic fluids for a better simulation of the self-pressurization processes.
In the present paper, a unified multi-phase solver incorporating four typical liquid-vapor phase change models, namely the Tanasawa model, the Lee model, the energy jump model, and the modified energy jump model has been proposed, which enables direct comparison among different models. The models are used to simulate the pressure and temperature evolutions in an experimental liquid nitrogen tank in normal gravity, which helps to evaluate themselves in the aspects of accuracy, calculation speed and robustness.
Section snippets
Governing equations for the self-pressurization tank
In the present study, both the fluid domain and the solid wall of the tank are modeled and discretized. The heat transportation at the solid boundaries is considered to be irrelevant with the nearby fluid velocity. Consequently, two sets of the solid and the fluid governing equations can be decoupled and solved separately. The pressures in the cryogenic container are usually from 100 kPa to 300 kPa. Under these conditions, the Knudsen number is far smaller than 0.01, and the fluids are
Self-pressurization results and phase change model comparison
This section compares the simulation results by different phase change models. Section 3.1 compares the pressure and temperature outputs from two KTG based models, namely the Lee model and the Tanasawa model. Section 3.2 presents the pressure predictions from the energy transport models, namely the energy jump model and the modified energy jump model, and compares pressure prediction performances between the KTG based models and the energy transport models. Section 3.3 evaluates the four models
Conclusion
A unified vapor-liquid-solid multi-phase numerical solver has been accomplished for the self pressurization simulation in cryogenic containers. Compared to the early fluid-only solver, the temperature prediction in the vicinity of the tank wall improves significantly. Four liquid-vapor phase change models were integrated into the solver, which enables fair and effective comparison for performances between each other. The pressure and temperature prediction accuracies, and the calculation speed
Authors declare that they have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, “Validity evaluation of popular liquid-vapor phase change models for cryogenic self-pressurization process”.
Acknowledgement
This project is supported by the National Natural Science Foundation of China (No. 51936006).
Progress in physical modelling and numerical simulation of phase transitions in cryogenic pool boiling and cavitation2023, Applied Mathematical ModellingCitation Excerpt :We will not delve into cryogenic evaporation phenomena, that predominantly drive the phase-change in well-insulated storage facilities, and thus are less relevant to spill scenarios. We instead refer the reader to the works of Zuo et al. [31–33]. If the static pressure at any location in a turbomachine drops below a fluid’s saturation pressure, localized evaporation events may occur, followed by rapid collapse of the vapour cavities in a process termed “cavitation” [34].Show abstract
Thermodynamic performance in a liquid oxygen tank during active-pressurization under different gas injection temperatures2023, International Communications in Heat and Mass TransferCitation Excerpt :The volume of fluid method is adopted to predict the tank pressurization performance. The associated governing equations could refer to previous published investigations [33–39,41,45,46]. Subjected to external heat input and gas injection, the phase change occurs at the interface and within the tank.Show abstract
Interfacial mass and energy transport during steady-state evaporation in liquid oxygen storage tanks2022, Applied EnergyCitation Excerpt :However, most of them simply used the Lee model for mass transport as did for regular fluids, and seldom focus on the evaporation itself related to the interfacial temperature distribution or were unable to validate their results against credible experimental data. A recent study proposed an optimized evaporation model for the cryogenic self-pressurization with a thorough comparison between popular phase change models [8], but still lacked of experimental data to validate the results. A series of experiments have been conducted on the heat and mass transport in a thin liquid layer in the vicinity of the liquid–vapor interface of room-temperature fluids [9–14].Show abstract
Thermal destratification of cryogenic liquid storage tanks by continuous bubbling of gases2022, International Journal of Hydrogen EnergyCitation Excerpt :It was concluded that a single injector with a larger diameter configuration showed a higher chance of developing a vertical temperature gradient. Zuo et al. [48] carried out a numerical analysis to investigate the temperature distribution within the LH2 storage tank with a self-pinning spraying bar. They used the SST turbulence model coupled with the 6-DOF model.Show abstract
Field observations provide valuable data regarding nearshore tsunami impact, yet only in inundation areas where tsunami waves have already flooded. Therefore, tsunami modeling is essential to understand tsunami behavior and prepare for tsunami inundation. It is necessary that all numerical models used in tsunami emergency planning be subject to benchmark tests for validation and verification. This study focuses on two numerical codes, NAMI DANCE and FLOW-3D®, for validation and performance comparison. NAMI DANCE is an in-house tsunami numerical model developed by the Ocean Engineering Research Center of Middle East Technical University, Turkey and Laboratory of Special Research Bureau for Automation of Marine Research, Russia. FLOW-3D® is a general purpose computational fluid dynamics software, which was developed by scientists who pioneered in the design of the Volume-of-Fluid technique. The codes are validated and their performances are compared via analytical, experimental and field benchmark problems, which are documented in the ‘‘Proceedings and Results of the 2011 National Tsunami Hazard Mitigation Program (NTHMP) Model Benchmarking Workshop’’ and the ‘‘Proceedings and Results of the NTHMP 2015 Tsunami Current Modeling Workshop”. The variations between the numerical solutions of these two models are evaluated through statistical error analysis.
현장 관찰은 연안 쓰나미 영향에 관한 귀중한 데이터를 제공하지만 쓰나미 파도가 이미 범람한 침수 지역에서만 가능합니다. 따라서 쓰나미 모델링은 쓰나미 행동을 이해하고 쓰나미 범람에 대비하는 데 필수적입니다.
쓰나미 비상 계획에 사용되는 모든 수치 모델은 검증 및 검증을 위한 벤치마크 테스트를 받아야 합니다. 이 연구는 검증 및 성능 비교를 위해 NAMI DANCE 및 FLOW-3D®의 두 가지 숫자 코드에 중점을 둡니다.
NAMI DANCE는 터키 중동 기술 대학의 해양 공학 연구 센터와 러시아 해양 연구 자동화를 위한 특별 조사국 연구소에서 개발한 사내 쓰나미 수치 모델입니다. FLOW-3D®는 Volume-of-Fluid 기술의 설계를 개척한 과학자들이 개발한 범용 전산 유체 역학 소프트웨어입니다.
코드의 유효성이 검증되고 분석, 실험 및 현장 벤치마크 문제를 통해 코드의 성능이 비교되며, 이는 ‘2011년 NTHMP(National Tsunami Hazard Mitigation Program) 모델 벤치마킹 워크숍의 절차 및 결과’와 ”절차 및 NTHMP 2015 쓰나미 현재 모델링 워크숍 결과”. 이 두 모델의 수치 해 사이의 변동은 통계적 오류 분석을 통해 평가됩니다.
The distribution of the computed maximum current speed during the entire duration of the NAMI DANCE and FLOW-3D simulations. The
resolution of computational domain is 10 m
Apotsos, A., Buckley, M., Gelfenbaum, G., Jafe, B., & Vatvani, D. (2011). Nearshore tsunami inundation and sediment transport modeling: towards model validation and application. Pure and Applied Geophysics,168(11), 2097–2119. https://doi.org/10.1007/s00024-011-0291-5.ArticleGoogle Scholar
Barberopoulou, A., Legg, M. R., & Gica, E. (2015). Time evolution of man-made harbor modifications in San Diego: effects on Tsunamis. Journal of Marine Science and Engineering,3, 1382–1403.ArticleGoogle Scholar
Basu, D., Green, S., Das, K., Janetzke, R. and Stamatakos, J. (2009). Numerical Simulation of Surface Waves Generated by a Subaerial Landslide at Lituya Bay, Alaska. Proceedings of 28th International Conference on Ocean, Offshore and Arctic Engineering. Honolulu, Hawaii, USA.
Briggs, M. J., Synolakis, C. E., Harkins, G. S., & Green, D. R. (1995). Laboratory experiments of tsunami run-up on a circular island. Pure and Applied Geophysics,144(3/4), 569–593.ArticleGoogle Scholar
Cheung, K. F., Bai, Y., & Yamazaki, Y. (2013). Surges around the Hawaiian Islands from the 2011 Tohoku Tsunami. Journal of Geophysical Research: Oceans,118, 5703–5719. https://doi.org/10.1002/jgrc.20413.Google Scholar
Choi, B. H., Dong, C. K., Pelinovsky, E., & Woo, S. B. (2007). Three-dimensional Simulation of Tsunami Run-up Around Conical Island. Coastal Engineering,54, 618–629.ArticleGoogle Scholar
Cox, D., T. Tomita, P. Lynett, R.A., Holman. (2008). Tsunami Inundation with Macroroughness in the Constructed Environment. Proceedings of 31st International Conference on Coastal Engineering, ASCE, pp. 1421–1432.
Flow Science. (2002). FLOW-3D User’s Manual.
Hirt, C. W., & Nichols, B. D. (1981). Volume of fluid (VOF) method for the dynamics of free boundaries. Journal of Computational Physics,39, 201–225.ArticleGoogle Scholar
Horrillo, J., Grilli, S. T., Nicolsky, D., Roeber, V., & Zang, J. (2015). Performance benchmarking Tsunami models for NTHMP’s inundation mapping activities. Pure and Applied Geophysics,172, 869–884.ArticleGoogle Scholar
Kim, K. O., Kim, D. C., Choi, B.-H., Jung, T. K., Yuk, J. H., & Pelinovsky, E. (2015). The role of diffraction effects in extreme run-up inundation at Okushiri Island due to 1993 Tsunami. Natural Hazards and Earth Systems Sciences,15, 747–755.ArticleGoogle Scholar
Liu, P. L.-F. (1994). Model equations for wave propagations from deep to shallow water. (P.-F. Liu, Ed.) Advances in Coastal and Ocean Engineering, 1, 125–158.
Liu, P. L.-F., Yeh, H., & Synolakis, C. E. (2008). Advanced numerical models for simulating Tsunami waves and run-up. Advances in Coastal and Ocean Engineering,10, 344.Google Scholar
Lynett, P. J., Gately, K., Wilson, R., Montoya, L., Arcas, D., Aytore, B., et al. (2017). Inter-model analysis of Tsunami-induced coastal currents. Ocean Modelling,114, 14–32.ArticleGoogle Scholar
Lynett, P. J., Wu, T.-R., & Liu, P. L.-F. (2002). Modeling wave run-up with depth-integrated equations. Coastal Engineering,46(2), 89–107.ArticleGoogle Scholar
Macias, J., Castro, M. J., Ortega, S., Escalante, C., & Gonzalez-Vida, J. M. (2017). Performance benchmarking of Tsunami-HySEA model for nthmp’s inundation mapping activities. Pure and Applied Geophysics,174, 3147–3183.ArticleGoogle Scholar
Matsuyama, M., & Tanaka, H. (2001). An experimental study of the highest run-up height in the 1993 Hokkaidō Nansei-Oki Earthquake Tsunami. Proceedings of ITS,2001, 879–889.Google Scholar
National Tsunami Hazard Mitigation Program. 2012. Proceedings and Results of the 2011 NTHMP Model Benchmarking Workshop. Boulder: U.S. Department of Commerce/NOAA/NTHMP; (NOAA Special Report). p. 436.
National Tsunami Hazard Mitigation Program. (2017). Proceedings and Results of the National Tsunami Hazard Mitigation Program 2015 Tsunami Current Modeling Workshop, February 9-10, 2015, Portland, Oregon: compiled by Patrick Lynett and Rick Wilson, p 194.
Necmioglu, O., & Ozel, N. M. (2014). An earthquake source sensitivity analysis for Tsunami propagation in the Eastern Mediterranean. Oceanography,27(2), 76–85.ArticleGoogle Scholar
Nichols, B.D. and Hirt, C.W. (1975). Methods for Calculating Multi-Dimensional, Transient Free Surface Flows Past Bodies. Proceedings of 1st International Conference Num. Ship Hydrodynamics. Gaithersburg.
Nicolsky, D. J., Suleimani, E. N., & Hansen, R. A. (2011). Validation and verification of a numerical model for Tsunami propagation and run-up. Pure and Applied Geophysics,168(6), 1199–1222.ArticleGoogle Scholar
Park, H., Cox, D. T., Lynett, P. J., Wiebe, D. M., & Shin, S. (2013). Tsunami inundation modeling in constructed environments: a physical and numerical comparison of free-surface elevation, velocity, and momentum flux. Coastal Engineering,79, 9–21.ArticleGoogle Scholar
Patel, V. M., Dholakia, M. B., & Singh, A. P. (2016). Emergency preparedness in the case of Makran Tsunami: a case study on Tsunami risk visualization for the Western Parts of Gujarat, India. Geomatics Natural Hazard and Risk,7(2), 826–842.ArticleGoogle Scholar
Pelinovsky, E., Kim, D.-C., Kim, K.-O., & Choi, B.-H. (2013). Three-dimensional simulation of extreme run-up heights during the 2004 Indonesian and 2011 Japanese Tsunamis. Vienna: EGU General Assembly.Google Scholar
Rueben, M., Holman, R., Cox, D., Shin, S., Killian, J., & Stanley, J. (2011). Optical measurements of Tsunami inundation through an urban waterfront modeled in a large-scale laboratory basin. Coastal Engineering,58, 229–238.ArticleGoogle Scholar
Shuto, N. (1991). Numerical simulation of Tsunamis—its present and near future. Natural Hazards,4, 171–191.ArticleGoogle Scholar
Synolakis, C. E. (1986). The run-up of long waves. Ph.D. Thesis. California Institute of Technology, Pasadena, California.
Synolakis, C. E., Bernard, E. N., Titov, V. V., Kanoglu, U. & Gonzalez, F. (2007). Standards, criteria, and procedures for NOAA evaluation of Tsunami Numerical Models. 55 p. Seattle, Washington: NOAA OAR Special Report, Contribution No 3053, NOAA/OAR/PMEL.
Synolakis, C. E., Bernard, E. N., Titov, V. V., Kanoglu, U., & Gonzalez, F. I. (2008). Validation and verification of Tsunami numerical models. Pure and Applied Geophysics,165, 2197–2228.ArticleGoogle Scholar
Tolkova, E. (2014). Land-water boundary treatment for a tsunami model with dimensional splitting. Pure and Applied Geophysics,171(9), 2289–2314.ArticleGoogle Scholar
Velioglu, D. (2017). Advanced two- and three-dimensional Tsunami models: benchmarking and validation. Ph.D. Thesis. Middle East Technical University, Ankara.
Velioglu, D., Kian, R., Yalciner, A.C. and Zaytsev, A. (2016). Performance assessment of NAMI DANCE in Tsunami evolution and currents using a benchmark problem. (R. Signell, Ed.) J. Mar. Sci. Eng., 4(3), 49.
Wu, T. (2001). A unified theory for modeling water waves. Advances in Applied Mechanics,37, 1–88.ArticleGoogle Scholar
Wu, N.-J., Hsiao, S.-C., Chen, H.-H., & Yang, R.-Y. (2016). The study on solitary waves generated by a piston-type wave maker. Ocean Engineering,117, 114–129.ArticleGoogle Scholar
Yalciner, A. C., Dogan, P. and Sukru. E. (2005). December 26 2004, Indian Ocean Tsunami Field Survey, North of Sumatra Island. UNESCO.
Yalciner, A. C., Gülkan, P., Dilmen, I., Aytore, B., Ayca, A., Insel, I., et al. (2014). Evaluation of Tsunami scenarios For Western Peloponnese, Greece. Bollettino di Geofisica Teorica ed Applicata,55, 485–500.Google Scholar
Yen, B. C. (1991). Hydraulic resistance in open channels. In B. C. Yen (Ed.), Channel flow resistance: centennial of manning’s formula (pp. 1–135). Highlands Ranch: Water Resource Publications.Google Scholar
Zaitsev, A. I., Kovalev, D. P., Kurkin, A. A., Levin, B. V., Pelinovskii, E. N., Chernov, A. G., et al. (2009). The Tsunami on Sakhalin on August 2, 2007: mareograph evidence and numerical simulation. Tikhookeanskaya Geologiya,28, 30–35.Google Scholar
The authors wish to thank Dr. Andrey Zaytsev due to his undeniable contributions to the development of in-house numerical model, NAMI DANCE. The Turkish branch of Flow Science, Inc. is also acknowledged. Finally, the National Tsunami Hazard Mitigation Program (NTHMP), who provided most of the benchmark data, is appreciated. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author information
Author notes
Deniz Velioglu SogutPresent address: 1212 Computer Science, Department of Civil Engineering, Stony Brook University, Stony Brook, NY, 11794, USA
Authors and Affiliations
Middle East Technical University, 06800, Ankara, TurkeyDeniz Velioglu Sogut & Ahmet Cevdet Yalciner
Velioglu Sogut, D., Yalciner, A.C. Performance Comparison of NAMI DANCE and FLOW-3D® Models in Tsunami Propagation, Inundation and Currents using NTHMP Benchmark Problems. Pure Appl. Geophys.176, 3115–3153 (2019). https://doi.org/10.1007/s00024-018-1907-9
A numerical study was performed on the embankment weir overflows with various surface roughness and tailwater submergence, to better understand the effects of weir roughness on discharge performances under the free and submerged conditions. The variation of flow regime is captured, from the free overflow, submerged hydraulic jump, to surface flow with increasing tailwater depth. A roughness factor is introduced to reflect the reduction in discharge caused by weir roughness. The roughness factor decreases with the roughness height, and it also depends on the tailwater depth, highlighting various relations of the roughness factor with the roughness height between different flow regimes, which is linear for the free overflow and submerged hydraulic jump while exponential for the surface flow. Accordingly, the effects of weir roughness on overflow discharge appear nonnegligible for the significant roughness height and the surface flow regime occurring under considerable tailwater submergence. The established empirical expressions of discharge coefficient and submergence and roughness factors make it possible to predict the discharge over embankment weirs considering both tailwater submergence and surface roughness.
자유 및 침수 조건에서 방류 성능에 대한 둑 거칠기의 영향을 더 잘 이해하기 위해 다양한 표면 거칠기와 테일워터 침수를 갖는 제방 둑 범람에 대한 수치 연구가 수행되었습니다.
자유 범람, 수중 수압 점프, 테일워터 깊이가 증가하는 표면 유동에 이르기까지 유동 체제의 변화가 캡처됩니다. 위어 거칠기로 인한 배출 감소를 반영하기 위해 거칠기 계수가 도입되었습니다.
조도 계수는 조도 높이와 함께 감소하고, 또한 테일워터 깊이에 따라 달라지며, 서로 다른 흐름 영역 사이의 조도 높이와 조도 계수의 다양한 관계를 강조합니다.
이는 자유 범람 및 수중 수압 점프에 대해 선형인 반면 표면에 대해 지수적입니다. 흐름. 따라서 월류 방류에 대한 웨어 조도의 영향은 상당한 조도 높이와 상당한 방수 침수 하에서 발생하는 표면 흐름 체제에 대해 무시할 수 없는 것으로 보입니다.
방류계수와 침수 및 조도계수의 확립된 실증식은 방류수 침수와 지표조도를 모두 고려한 제방보 위의 방류량을 예측할 수 있게 합니다.
References
Kindsvater C. E. Discharge characteristics of embankment -shaped weirs (No. 1617) [R]. Washington DC, USA: US Government Printing Office, 1964.Google Scholar
Fritz H. M., Hager W. H. Hydraulics of embankment weirs [J]. Journal of Hydraulic Engineering, ASCE, 1998, 124(9): 963–971.ArticleGoogle Scholar
Azimi A. H., Rajaratnam N., Zhu D. Z. Water surface characteristics of submerged rectangular sharp-crested weirs [J]. Journal of Hydraulic Engineering, ASCE, 2016, 142(5): 06016001.ArticleGoogle Scholar
Felder S., Islam N. Hydraulic performance of an embankment weir with rough crest [J]. Journal of Hydraulic Engineering, ASCE, 2017, 143(3): 04016086.ArticleGoogle Scholar
Hakim S. S., Azimi A. H. Hydraulics of submerged traingular weirs and weirs of finite-crest length with upstream and downstream ramps [J]. Journal of Irrigation and Drainage Engineering, 2017, 143(8): 06017008.ArticleGoogle Scholar
Safarzadeh A., Mohajeri S. H. Hydrodynamics of rectangular broad-crested porous weirs [J]. Journal of Hydraulic Engineering, ASCE, 2018, 144(10): 04018028.Google Scholar
Sargison J. E., Percy A. Hydraulics of broad-crested weirs with varying side slopes [J]. Journal of Irrigation and Drainage Engineering, 2009, 35(1): 115–118.ArticleGoogle Scholar
Yang Z., Bai F., Huai W. et al. Lattice Boltzmann method for simulating flows in the open-channel with partial emergent rigid vegetation cover [J]. Journal of Hydrodynamics, 2019, 31(4): 717–724.ArticleGoogle Scholar
Fathi-moghaddam M., Sadrabadi M. T., Rahmanshahi M. Numerical simulation of the hydraulic performance of triangular and trapezoidal gabion weirs in free flow condtion [J]. Flow Measurement on Instrumentation, 2018, 62: 93–104.ArticleGoogle Scholar
Zerihun Y. T. A one-dimensional Boussinesq-type momentum model for steady rapidly varied open channel flows [D]. Doctoral Thesis, Melbourne, Australia: The University of Melbourne, 2004.Google Scholar
Pařílková J., Říha J., Zachoval Z. The influence of roughness on the discharge coefficient of a broad-crested weir [J]. Journal of Hydrology and Hydromechanics, 2012, 60(2): 101–114.ArticleGoogle Scholar
Říha J., Duchan D., Zachoval Z. et al. Performance of a shallow-water model for simulating flow over trapezoidal broad-crested weirs [J]. Journal of Hydrology and Hydromechanics, 2019, 67(4): 322–328.ArticleGoogle Scholar
Yan X., Ghodoosipour B., Mohammadian A. Three-dimensional numerical study of multiple vertical buoyant jets in stationary ambient water [J]. Journal of Hydraulic Engineering, ASCE, 2020, 146(7): 04020049.ArticleGoogle Scholar
Qian S., Xu H., Feng J. Flume experiments on baffle-posts for retarding open channel flow: By C. UBING, R. ETTEMA and CI THORNTON, J. Hydraulic Res. 55 (3), 2017, 430–437 [J]. Journal of Hydraulic Research, 2019, 57(2): 280–282.ArticleGoogle Scholar
Sun J., Qian S., Xu H. et al. Three-dimensional numerical simulation of stepped dropshaft with different step shape [J]. Water Science and Technology Water Supply, 2020, 21(1): 581–592.Google Scholar
Qian S., Wu J., Zhou Y. et al. Discussion of “Hydraulic performance of an embankment weir with rough crest” by Stefan Felder and Nushan Islam [J]. Journal of Hydraulic Engineering, ASCE, 2018, 144(4): 07018003.ArticleGoogle Scholar
Mohammadpour R., Ghani A. A., Azamathulla H. M. Numerical modeling of 3-D flow on porous broad crested weirs [J]. Applied Mathematical Modelling, 2013, 37(22): 9324–9337.ArticleGoogle Scholar
Savage B. M., Brian M. C., Greg S. P. Physical and numerical modeling of large headwater ratios for a 15° labyrinth spillway [J]. Journal of Hydraulic Engineering, ASCE, 2016, 142(11): 04016046.ArticleGoogle Scholar
Al-Husseini T. R., Al-Madhhachi A. S. T., Naser Z. A. Laboratory experiments and numerical model of local scour around submerged sharp crested weirs [J]. Journal of King Saud University Science, 2020, 32(3): 167–176.ArticleGoogle Scholar
Zerihun Y. T., Fenton J. D. A Boussinesq-type model for flow over trapezoidal profile weirs [J]. Journal of Hydraulic Research, 2007, 45(4): 519–528.ArticleGoogle Scholar
Flow Science, Inc. FLOW-3D ® Version 12.0 Users Manual (2018) [EB/OL]. Santa Fe, NM, USA: Flow Science, Inc., 2019.Google Scholar
Bazin H. Expériences nouvelles sur l’ecoulement par déversoir [R]. Paris, France: Annales des Ponts et Chaussées, 1898.MATHGoogle Scholar
Hager W. H., Schwalt M. Broad-crested weir [J]. Journal of Irrigation and Drainage Engineering, 1994, 120(1): 13–26.ArticleGoogle Scholar
A series of numerical simulation were conducted to study the local scour around umbrella suction anchor foundation (USAF) under random waves. In this study, the validation was carried out firstly to verify the accuracy of the present model. Furthermore, the scour evolution and scour mechanism were analyzed respectively. In addition, two revised models were proposed to predict the equilibrium scour depth Seq around USAF. At last, a parametric study was carried out to study the effects of the Froude number Fr and Euler number Eu for the Seq. The results indicate that the present numerical model is accurate and reasonable for depicting the scour morphology under random waves. The revised Raaijmakers’s model shows good agreement with the simulating results of the present study when KCs,p < 8. The predicting results of the revised stochastic model are the most favorable for n = 10 when KCrms,a < 4. The higher Fr and Eu both lead to the more intensive horseshoe vortex and larger Seq.
The rapid expansion of cities tends to cause social and economic problems, such as environmental pollution and traffic jam. As a kind of clean energy, offshore wind power has developed rapidly in recent years. The foundation of offshore wind turbine (OWT) supports the upper tower, and suffers the cyclic loading induced by waves, tides and winds, which exerts a vital influence on the OWT system. The types of OWT foundation include the fixed and floating foundation, and the fixed foundation was used usually for nearshore wind turbine. After the construction of fixed foundation, the hydrodynamic field changes in the vicinity of the foundation, leading to the horseshoe vortex formation and streamline compression at the upside and sides of foundation respectively [1,2,3,4]. As a result, the neighboring soil would be carried away by the shear stress induced by vortex, and the scour hole would emerge in the vicinity of foundation. The scour holes increase the cantilever length, and weaken the lateral bearing capacity of foundation [5,6,7,8,9]. Moreover, the natural frequency of OWT system increases with the increase of cantilever length, causing the resonance occurs when the system natural frequency equals the wave or wind frequency [10,11,12]. Given that, an innovative foundation called umbrella suction anchor foundation (USAF) has been designed for nearshore wind power. The previous studies indicated the USAF was characterized by the favorable lateral bearing capacity with the low cost [6,13,14]. The close-up of USAF is shown in Figure 1, and it includes six parts: 1-interal buckets, 2-external skirt, 3-anchor ring, 4-anchor branch, 5-supporting rod, 6-telescopic hook. The detailed description and application method of USAF can be found in reference [13].
Figure 1. The close-up of umbrella suction anchor foundation (USAF).
Numerical and experimental investigations of scour around OWT foundation under steady currents and waves have been extensively studied by many researchers [1,2,15,16,17,18,19,20,21,22,23,24]. The seabed scour can be classified as two types according to Shields parameter θ, i.e., clear bed scour (θ < θcr) or live bed scour (θ > θcr). Due to the set of foundation, the adverse hydraulic pressure gradient exists at upstream foundation edges, resulting in the streamline separation between boundary layer flow and seabed. The separating boundary layer ascended at upstream anchor edges and developed into the horseshoe vortex. Then, the horseshoe vortex moved downstream gradually along the periphery of the anchor, and the vortex shed off continually at the lee-side of the anchor, i.e., wake vortex. The core of wake vortex is a negative pressure center, liking a vacuum cleaner. Hence, the soil particles were swirled into the negative pressure core and carried away by wake vortexes. At the same time, the onset of scour at rear side occurred. Finally, the wake vortex became downflow when the turbulence energy could not support the survival of wake vortex. According to Tavouktsoglou et al. [25], the scale of pile wall boundary layer is proportional to 1/ln(Rd) (Rd is pile Reynolds), which means the turbulence intensity induced by the flow-structure interaction would decrease with Rd increases, but the effects of Rd can be neglected only if the flow around the foundation is fully turbulent [26]. According to previous studies [1,15,27,28,29,30,31,32], the scour development around pile foundation under waves was significantly influenced by Shields parameter θ and KC number simultaneously (calculated by Equation (1)). Sand ripples widely existed around pile under waves in the case of live bed scour, and the scour morphology is related with θ and KC. Compared with θ, KC has a greater influence on the scour morphology [21,27,28]. The influence mechanism of KC on the scour around the pile is reflected in two aspects: the horseshoe vortex at upstream and wake vortex shedding at downstream.
KC=UwmTD��=�wm��(1)
where, Uwm is the maximum velocity of the undisturbed wave-induced oscillatory flow at the sea bottom above the wave boundary layer, T is wave period, and D is pile diameter.
There are two prerequisites to satisfy the formation of horseshoe vortex at upstream pile edges: (1) the incoming flow boundary layer with sufficient thickness and (2) the magnitude of upstream adverse pressure gradient making the boundary layer separating [1,15,16,18,20]. The smaller KC results the lower adverse pressure gradient, and the boundary layer cannot separate, herein, there is almost no horseshoe vortex emerging at upside of pile. Sumer et al. [1,15] carried out several sets of wave flume experiments under regular and irregular waves respectively, and the experiment results show that there is no horseshoe vortex when KC is less than 6. While the scale and lifespan of horseshoe vortex increase evidently with the increase of KC when KC is larger than 6. Moreover, the wake vortex contributes to the scour at lee-side of pile. Similar with the case of horseshoe vortex, there is no wake vortex when KC is less than 6. The wake vortex is mainly responsible for scour around pile when KC is greater than 6 and less than O(100), while horseshoe vortex controls scour nearly when KC is greater than O(100).
Sumer et al. [1] found that the equilibrium scour depth was nil around pile when KC was less than 6 under regular waves for live bed scour, while the equilibrium scour depth increased with the increase of KC. Based on that, Sumer proposed an equilibrium scour depth predicting equation (Equation (2)). Carreiras et al. [33] revised Sumer’s equation with m = 0.06 for nonlinear waves. Different with the findings of Sumer et al. [1] and Carreiras et al. [33], Corvaro et al. [21] found the scour still occurred for KC ≈ 4, and proposed the revised equilibrium scour depth predicting equation (Equation (3)) for KC > 4.
Rudolph and Bos [2] conducted a series of wave flume experiments to investigate the scour depth around monopile under waves only, waves and currents combined respectively, indicting KC was one of key parameters in influencing equilibrium scour depth, and proposed the equilibrium scour depth predicting equation (Equation (4)) for low KC (1 < KC < 10). Through analyzing the extensive data from published literatures, Raaijmakers and Rudolph [34] developed the equilibrium scour depth predicting equation (Equation (5)) for low KC, which was suitable for waves only, waves and currents combined. Khalfin [35] carried out several sets of wave flume experiments to study scour development around monopile, and proposed the equilibrium scour depth predicting equation (Equation (6)) for low KC (0.1 < KC < 3.5). Different with above equations, the Khalfin’s equation considers the Shields parameter θ and KC number simultaneously in predicting equilibrium scour depth. The flow reversal occurred under through in one wave period, so sand particles would be carried away from lee-side of pile to upside, resulting in sand particles backfilled into the upstream scour hole [20,29]. Considering the backfilling effects, Zanke et al. [36] proposed the equilibrium scour depth predicting equation (Equation (7)) around pile by theoretical analysis, and the equation is suitable for the whole range of KC number under regular waves and currents combined.
where, γ is safety factor, depending on design process, typically γ = 1.5, Kwave is correction factor considering wave action, Khw is correction factor considering water depth.
where, n is the 1/n’th highest wave for random waves
For predicting equilibrium scour depth under irregular waves, i.e., random waves, Sumer and Fredsøe [16] found it’s suitable to take Equation (2) to predict equilibrium scour depth around pile under random waves with the root-mean-square (RMS) value of near-bed orbital velocity amplitude Um and peak wave period TP to calculate KC. Khalfin [35] recommended the RMS wave height Hrms and peak wave period TP were used to calculate KC for Equation (6). References [37,38,39,40] developed a series of stochastic theoretical models to predict equilibrium scour depth around pile under random waves, nonlinear random waves plus currents respectively. The stochastic approach thought the 1/n’th highest wave were responsible for scour in vicinity of pile under random waves, and the KC was calculated in Equation (8) with Um and mean zero-crossing wave period Tz. The results calculated by Equation (8) agree well with experimental values of Sumer and Fredsøe [16] if the 1/10′th highest wave was used. To author’s knowledge, the stochastic approach proposed by Myrhaug and Rue [37] is the only theoretical model to predict equilibrium scour depth around pile under random waves for the whole range of KC number in published documents. Other methods of predicting scour depth under random waves are mainly originated from the equation for regular waves-only, waves and currents combined, which are limited to the large KC number, such as KC > 6 for Equation (2) and KC > 4 for Equation (3) respectively. However, situations with relatively low KC number (KC < 4) often occur in reality, for example, monopile or suction anchor for OWT foundations in ocean environment. Moreover, local scour around OWT foundations under random waves has not yet been investigated fully. Therefore, further study are still needed in the aspect of scour around OWT foundations with low KC number under random waves. Given that, this study presents the scour sediment model around umbrella suction anchor foundation (USAF) under random waves. In this study, a comparison of equilibrium scour depth around USAF between this present numerical models and the previous theoretical models and experimental results was presented firstly. Then, this study gave a comprehensive analysis for the scour mechanisms around USAF. After that, two revised models were proposed according to the model of Raaijmakers and Rudolph [34] and the stochastic model developed by Myrhaug and Rue [37] respectively to predict the equilibrium scour depth. Finally, a parametric study was conducted to study the effects of the Froude number (Fr) and Euler number (Eu) to equilibrium scour depth respectively.
2. Numerical Method
2.1. Governing Equations of Flow
The following equations adopted in present model are already available in Flow 3D software. The authors used these theoretical equations to simulate scour in random waves without modification. The incompressible viscous fluid motion satisfies the Reynolds-averaged Navier-Stokes (RANS) equation, so the present numerical model solves RANS equations:
where, VF is the volume fraction; u, v, and w are the velocity components in x, y, z direction respectively with Cartesian coordinates; Ai is the area fraction; ρf is the fluid density, fi is the viscous fluid acceleration, Gi is the fluid body acceleration (i = x, y, z).
2.2. Turbulent Model
The turbulence closure is available by the turbulent model, such as one-equation, the one-equation k-ε model, the standard k-ε model, RNG k-ε turbulent model and large eddy simulation (LES) model. The LES model requires very fine mesh grid, so the computational time is large, which hinders the LES model application in engineering. The RNG k-ε model can reduce computational time greatly with high accuracy in the near-wall region. Furthermore, the RNG k-ε model computes the maximum turbulent mixing length dynamically in simulating sediment scour model. Therefore, the RNG k-ε model was adopted to study the scour around anchor under random waves [41,42].
where, kT is specific kinetic energy involved with turbulent velocity, GT is the turbulent energy generated by buoyancy; εT is the turbulent energy dissipating rate, PT is the turbulent energy, Diffε and DiffkT are diffusion terms associated with VF, Ai; CDIS1, CDIS2 and CDIS3 are dimensionless parameters, and CDIS1, CDIS3 have default values of 1.42, 0.2 respectively. CDIS2 can be obtained from PT and kT.
2.3. Sediment Scour Model
The sand particles may suffer four processes under waves, i.e., entrainment, bed load transport, suspended load transport, and deposition, so the sediment scour model should depict the above processes efficiently. In present numerical simulation, the sediment scour model includes the following aspects:
2.3.1. Entrainment and Deposition
The combination of entrainment and deposition determines the net scour rate of seabed in present sediment scour model. The entrainment lift velocity of sand particles was calculated as [43]:
where, αi is the entrainment parameter, ns is the outward point perpendicular to the seabed, d* is the dimensionless diameter of sand particles, which was calculated by Equation (15), θcr is the critical Shields parameter, g is the gravity acceleration, di is the diameter of sand particles, ρi is the density of seabed species.
In Equation (14), the entrainment parameter αi confirms the rate at which sediment erodes when the given shear stress is larger than the critical shear stress, and the recommended value 0.018 was adopted according to the experimental data of Mastbergen and Von den Berg [43]. ns is the outward pointing normal to the seabed interface, and ns = (0,0,1) according to the Cartesian coordinates used in present numerical model.
The shields parameter was obtained from the following equation:
θ=U2f,m(ρi/ρf−1)gd50�=�f,m2(��/�f−1)��50(16)
where, Uf,m is the maximum value of the near-bed friction velocity; d50 is the median diameter of sand particles. The detailed calculation procedure of θ was available in Soulsby [44].
The critical shields parameter θcr was obtained from the Equation (17) [44]
The sand particles begin to deposit on seabed when the turbulence energy weaken and cann’t support the particles suspending. The setting velocity of the particles was calculated from the following equation [44]:
This is called bed load transport when the sand particles roll or bounce over the seabed and always have contact with seabed. The bed load transport velocity was computed by [45]:
where, qb,i is the bed load transport rate, which was obtained from Equation (20), δi is the bed load thickness, which was calculated by Equation (21), cb,i is the volume fraction of sand i in the multiple species, fb is the critical packing fraction of the seabed.
where, Cs,i is the suspended sand particles mass concentration of sand i in the multiple species, us,i is the sand particles velocity of sand i, Df is the diffusivity.
The velocity of sand i in the multiple species could be obtained from the following equation:
where, u¯�¯ is the velocity of mixed fluid-particles, which can be calculated by the RANS equation with turbulence model, cs,i is the suspended sand particles volume concentration, which was computed from Equation (24).
cs,i=Cs,iρi�s,�=�s,���(24)
3. Model Setup
The seabed-USAF-wave three-dimensional scour numerical model was built using Flow-3D software. As shown in Figure 2, the model includes sandy seabed, USAF model, sea water, two baffles and porous media. The dimensions of USAF are shown in Table 1. The sandy bed (210 m in length, 30 m in width and 11 m in height) is made up of uniform fine sand with median diameter d50 = 0.041 cm. The USAF model includes upper steel tube with the length of 20 m, which was installed in the middle of seabed. The location of USAF is positioned at 140 m from the upstream inflow boundary and 70 m from the downstream outflow boundary. Two baffles were installed at two ends of seabed. In order to eliminate the wave reflection basically, the porous media was set at the outflow side on the seabed.
Figure 2. (a) The sketch of seabed-USAF-wave three-dimensional model; (b) boundary condation:Wv-wave boundary, S-symmetric boundary, O-outflow boundary; (c) USAF model.
Table 1. Numerical simulating cases.
3.1. Mesh Geometric Dimensions
In the simulation of the scour under the random waves, the model includes the umbrella suction anchor foundation, seabed and fluid. As shown in Figure 3, the model mesh includes global mesh grid and nested mesh grid, and the total number of grids is 1,812,000. The basic procedure for building mesh grid consists of two steps. Step 1: Divide the global mesh using regular hexahedron with size of 0.6 × 0.6. The global mesh area is cubic box, embracing the seabed and whole fluid volume, and the dimensions are 210 m in length, 30 m in width and 32 m in height. The details of determining the grid size can see the following mesh sensitivity section. Step 2: Set nested fine mesh grid in vicinity of the USAF with size of 0.3 × 0.3 so as to shorten the computation cost and improve the calculation accuracy. The encryption range is −15 m to 15 m in x direction, −15 m to 15 m in y direction and 0 m to 32 m in z direction, respectively. In order to accurately capture the free-surface dynamics, such as the fluid-air interface, the volume of fluid (VOF) method was adopted for tracking the free water surface. One specific algorithm called FAVORTM (Fractional Area/Volume Obstacle Representation) was used to define the fractional face areas and fractional volumes of the cells which are open to fluid flow.
Figure 3. The sketch of mesh grid.
3.2. Boundary Conditions
As shown in Figure 2, the initial fluid length is 210 m as long as seabed. A wave boundary was specified at the upstream offshore end. The details of determining the random wave spectrum can see the following wave parameters section. The outflow boundary was set at the downstream onshore end. The symmetry boundary was used at the top and two sides of the model. The symmetric boundaries were the better strategy to improve the computation efficiency and save the calculation cost [46]. At the seabed bottom, the wall boundary was adopted, which means the u = v = w= 0. Besides, the upper steel tube of USAF was set as no-slip condition.
3.3. Wave Parameters
The random waves with JONSWAP wave spectrum were used for all simulations as realistic representation of offshore conditions. The unidirectional JONSWAP frequency spectrum was described as [47]:
where, α is wave energy scale parameter, which is calculated by Equation (26), ω is frequency, ωp is wave spectrum peak frequency, which can be obtained from Equation (27). γ is wave spectrum peak enhancement factor, in this study γ = 3.3. σ is spectral width factor, σ equals 0.07 for ω ≤ ωp and 0.09 for ω > ωp respectively.
α=0.0076(gXU2)−0.22�=0.0076(���2)−0.22(26)
ωp=22(gU)(gXU2)−0.33�p=22(��)(���2)−0.33(27)
where, X is fetch length, U is average wind velocity at 10 m height from mean sea level.
In present numerical model, the input key parameters include X and U for wave boundary with JONSWAP wave spectrum. The objective wave height and period are available by different combinations of X and U. In this study, we designed 9 cases with different wave heights, periods and water depths for simulating scour around USAF under random waves (see Table 2). For random waves, the wave steepness ε and Ursell number Ur were acquired form Equations (28) and (29) respectively
ε=2πgHsT2a�=2���s�a2(28)
Ur=Hsk2h3w�r=�s�2ℎw3(29)
where, Hs is significant wave height, Ta is average wave period, k is wave number, hw is water depth. The Shield parameter θ satisfies θ>θcr for all simulations in current study, indicating the live bed scour prevails.
Table 2. Numerical simulating cases.
3.4. Mesh Sensitivity
In this section, a mesh sensitivity analysis was conducted to investigate the influence of mesh grid size to results and make sure the calculation is mesh size independent and converged. Three mesh grid size were chosen: Mesh 1—global mesh grid size of 0.75 × 0.75, nested fine mesh grid size of 0.4 × 0.4, and total number of grids 1,724,000, Mesh 2—global mesh grid size of 0.6 × 0.6, nested fine mesh grid size of 0.3 × 0.3, and total number of grids 1,812,000, Mesh 3—global mesh grid size of 0.4 × 0.4, nested fine mesh grid size of 0.2 × 0.2, and total number of grids 1,932,000. The near-bed shear velocity U* is an important factor for influencing scour process [1,15], so U* at the position of (4,0,11.12) was evaluated under three mesh sizes. As the Figure 4 shown, the maximum error of shear velocity ∆U*1,2 is about 39.8% between the mesh 1 and mesh 2, and 4.8% between the mesh 2 and mesh 3. According to the mesh sensitivity criterion adopted by Pang et al. [48], it’s reasonable to think the results are mesh size independent and converged with mesh 2. Additionally, the present model was built according to prototype size, and the mesh size used in present model is larger than the mesh size adopted by Higueira et al. [49] and Corvaro et al. [50]. If we choose the smallest cell size, it will take too much time. For example, the simulation with Mesh3 required about 260 h by using a computer with Intel Xeon Scalable Gold 4214 CPU @24 Cores, 2.2 GHz and 64.00 GB RAM. Therefore, in this case, considering calculation accuracy and computation efficiency, the mesh 2 was chosen for all the simulation in this study.
Figure 4. Comparison of near-bed shear velocity U* with different mesh grid size.
The nested mesh block was adopted for seabed in vicinity of the USAF, which was overlapped with the global mesh block. When two mesh blocks overlap each other, the governing equations are by default solved on the mesh block with smaller average cell size (i.e., higher grid resolution). It is should be noted that the Flow 3D software used the moving mesh captures the scour evolution and automatically adjusts the time step size to be as large as possible without exceeding any of the stability limits, affecting accuracy, or unduly increasing the effort required to enforce the continuity condition [51].
3.5. Model Validation
In order to verify the reliability of the present model, the results of present study were compared with the experimental data of Khosronejad et al. [52]. The experiment was conducted in an open channel with a slender vertical pile under unidirectional currents. The comparison of scour development between the present results and the experimental results is shown in Figure 5. The Figure 5 reveals that the present results agree well with the experimental data of Khosronejad et al. [52]. In the first stage, the scour depth increases rapidly. After that, the scour depth achieves a maximum value gradually. The equilibrium scour depth calculated by the present model is basically corresponding with the experimental results of Khosronejad et al. [52], although scour depth in the present model is slightly larger than the experimental results at initial stage.
Figure 5. Comparison of time evolution of scour between the present study and Khosronejad et al. [52], Petersen et al. [17].
Secondly, another comparison was further conducted between the results of present study and the experimental data of Petersen et al. [17]. The experiment was carried out in a flume with a circular vertical pile in combined waves and current. Figure 4 shows a comparison of time evolution of scour depth between the simulating and the experimental results. As Figure 5 indicates, the scour depth in this study has good overall agreement with the experimental results proposed in Petersen et al. [17]. The equilibrium scour depth calculated by the present model is 0.399 m, which equals to the experimental value basically. Overall, the above verifications prove the present model is accurate and capable in dealing with sediment scour under waves.
In addition, in order to calibrate and validate the present model for hydrodynamic parameters, the comparison of water surface elevation was carried out with laboratory experiments conducted by Stahlmann [53] for wave gauge No. 3. The Figure 6 depicts the surface wave profiles between experiments and numerical model results. The comparison indicates that there is a good agreement between the model results and experimental values, especially the locations of wave crest and trough. Comparison of the surface elevation instructs the present model has an acceptable relative error, and the model is a calibrated in terms of the hydrodynamic parameters.
Figure 6. Comparison of surface elevation between the present study and Stahlmann [53].
Finally, another comparison was conducted for equilibrium scour depth or maximum scour depth under random waves with the experimental data of Sumer and Fredsøe [16] and Schendel et al. [22]. The Figure 7 shows the comparison between the numerical results and experimental data of Run01, Run05, Run21 and Run22 in Sumer and Fredsøe [16] and test A05 and A09 in Schendel et al. [22]. As shown in Figure 7, the equilibrium scour depth or maximum scour depth distributed within the ±30 error lines basically, meaning the reliability and accuracy of present model for predicting equilibrium scour depth around foundation in random waves. However, compared with the experimental values, the present model overestimated the equilibrium scour depth generally. Given that, a calibration for scour depth was carried out by multiplying the mean reduced coefficient 0.85 in following section.
Figure 7. Comparison of equilibrium (or maximum) scour depth between the present study and Sumer and Fredsøe [16], Schendel et al. [22].
Through the various examination for hydrodynamic and morphology parameters, it can be concluded that the present model is a validated and calibrated model for scour under random waves. Thus, the present numerical model would be utilized for scour simulation around foundation under random waves.
4. Numerical Results and Discussions
4.1. Scour Evolution
Figure 8 displays the scour evolution for case 1–9. As shown in Figure 8a, the scour depth increased rapidly at the initial stage, and then slowed down at the transition stage, which attributes to the backfilling occurred in scour holes under live bed scour condition, resulting in the net scour decreasing. Finally, the scour reached the equilibrium state when the amount of sediment backfilling equaled to that of scouring in the scour holes, i.e., the net scour transport rate was nil. Sumer and Fredsøe [16] proposed the following formula for the scour development under waves
St=Seq(1−exp(−t/Tc))�t=�eq(1−exp(−�/�c))(30)
where Tc is time scale of scour process.
Figure 8. Time evolution of scour for case 1–9: (a) Case 1–5; (b) Case 6–9.
The computing time is 3600 s and the scour development curves in Figure 8 kept fluctuating, meaning it’s still not in equilibrium scour stage in these cases. According to Sumer and Fredsøe [16], the equilibrium scour depth can be acquired by fitting the data with Equation (30). From Figure 8, it can be seen that the scour evolution obtained from Equation (30) is consistent with the present study basically at initial stage, but the scour depth predicted by Equation (30) developed slightly faster than the simulating results and the Equation (30) overestimated the scour depth to some extent. Overall, the whole tendency of the results calculated by Equation (30) agrees well with the simulating results of the present study, which means the Equation (30) is applicable to depict the scour evolution around USAF under random waves.
4.2. Scour Mechanism under Random Waves
The scour morphology and scour evolution around USAF are similar under random waves in case 1~9. Taking case 7 as an example, the scour morphology is shown in Figure 9.
Figure 9. Scour morphology under different times for case 7.
From Figure 9, at the initial stage (t < 1200 s), the scour occurred at upstream foundation edges between neighboring anchor branches. The maximum scour depth appeared at the lee-side of the USAF. Correspondingly, the sediments deposited at the periphery of the USAF, and the location of the maximum accretion depth was positioned at an angle of about 45° symmetrically with respect to the wave propagating direction in the lee-side of the USAF. After that, when t > 2400 s, the location of the maximum scour depth shifted to the upside of the USAF at an angle of about 45° with respect to the wave propagating direction.
According to previous studies [1,15,16,19,30,31], the horseshoe vortex, streamline compression and wake vortex shedding were responsible for scour around foundation. The Figure 10 displays the distribution of flow velocity in vicinity of foundation, which reflects the evolving processes of horseshoe vertex.
Figure 10. Velocity profile around USAF: (a) Flow runup and down stream at upstream anchor edges; (b) Horseshoe vortex at upstream anchor edges; (c) Flow reversal during wave through stage at lee side.
As shown in Figure 10, the inflow tripped to the upstream edges of the USAF and it was blocked by the upper tube of USAF. Then, the downflow formed the horizontal axis clockwise vortex and rolled on the seabed bypassing the tube, that is, the horseshoe vortex (Figure 11). The Figure 12 displays the turbulence intensity around the tube on the seabed. From Figure 12, it can be seen that the turbulence intensity was high-intensity with respect to the region of horseshoe vortex. This phenomenon occurred because of drastic water flow momentum exchanging in the horseshoe vortex. As a result, it created the prominent shear stress on the seabed, causing the local scour at the upstream edges of USAF. Besides, the horseshoe vortex moved downstream gradually along the periphery of the tube and the wake vortex shed off continually at the lee-side of the USAF, i.e., wake vortex.
Figure 11. Sketch of scour mechanism around USAF under random waves.
Figure 12. Turbulence intensity: (a) Turbulence intensity of horseshoe vortex; (b) Turbulence intensity of wake vortex; (c) Turbulence intensity of accretion area.
The core of wake vortex is a negative pressure center, liking a vacuum cleaner [11,42]. Hence, the soil particles were swirled into the negative pressure core and carried away by wake vortex. At the same time, the onset of scour at rear side occurred. Finally, the wake vortex became downflow at the downside of USAF. As is shown in Figure 12, the turbulence intensity was low where the downflow occurred at lee-side, which means the turbulence energy may not be able to support the survival of wake vortex, leading to accretion happening. As mentioned in previous section, the formation of horseshoe vortex was dependent with adverse pressure gradient at upside of foundation. As shown in Figure 13, the evaluated range of pressure distribution is −15 m to 15 m in x direction. The t = 450 s and t = 1800 s indicate that the wave crest and trough arrived at the upside and lee-side of the foundation respectively, and the t = 350 s was neither the wave crest nor trough. The adverse gradient pressure reached the maximum value at t = 450 s corresponding to the wave crest phase. In this case, it’s helpful for the wave boundary separating fully from seabed, which leads to the formation of horseshoe vortex with high turbulence intensity. Therefore, the horseshoe vortex is responsible for the local scour between neighboring anchor branches at upside of USAF. What’s more, due to the combination of the horseshoe vortex and streamline compression, the maximum scour depth occurred at the upside of the USAF with an angle of about 45° corresponding to the wave propagating direction. This is consistent with the findings of Pang et al. [48] and Sumer et al. [1,15] in case of regular waves. At the wave trough phase (t = 1800 s), the pressure gradient became positive at upstream USAF edges, which hindered the separating of wave boundary from seabed. In the meantime, the flow reversal occurred (Figure 10) and the adverse gradient pressure appeared at downstream USAF edges, but the magnitude of adverse gradient pressure at lee-side was lower than the upstream gradient pressure under wave crest. In this way, the intensity of horseshoe vortex behind the USAF under wave trough was low, which explains the difference of scour depth at upstream and downstream, i.e., the scour asymmetry. In other words, the scour asymmetry at upside and downside of USAF was attributed to wave asymmetry for random waves, and the phenomenon became more evident for nonlinear waves [21]. Briefly speaking, the vortex system at wave crest phase was mainly related to the scour process around USAF under random waves.
Figure 13. Pressure distribution around USAF.
4.3. Equilibrium Scour Depth
The KC number is a key parameter for horseshoe vortex emerging and evolving under waves. According to Equation (1), when pile diameter D is fixed, the KC depends on the maximum near-bed velocity Uwm and wave period T. For random waves, the Uwm can be denoted by the root-mean-square (RMS) value of near-bed velocity amplitude Uwm,rms or the significant value of near-bed velocity amplitude Uwm,s. The Uwm,rms and Uwm,s for all simulating cases of the present study are listed in Table 3 and Table 4. The T can be denoted by the mean up zero-crossing wave period Ta, peak wave period Tp, significant wave period Ts, the maximum wave period Tm, 1/10′th highest wave period Tn = 1/10 and 1/5′th highest wave period Tn = 1/5 for random waves, so the different combinations of Uwm and T will acquire different KC. The Table 3 and Table 4 list 12 types of KC, for example, the KCrms,s was calculated by Uwm,rms and Ts. Sumer and Fredsøe [16] conducted a series of wave flume experiments to investigate the scour depth around monopile under random waves, and found the equilibrium scour depth predicting equation (Equation (2)) for regular waves was applicable for random waves with KCrms,p. It should be noted that the Equation (2) is only suitable for KC > 6 under regular waves or KCrms,p > 6 under random waves.
Table 3.Uwm,rms and KC for case 1~9.
Table 4.Uwm,s and KC for case 1~9.
Raaijmakers and Rudolph [34] proposed the equilibrium scour depth predicting model (Equation (5)) around pile under waves, which is suitable for low KC. The format of Equation (5) is similar with the formula proposed by Breusers [54], which can predict the equilibrium scour depth around pile at different scour stages. In order to verify the applicability of Raaijmakers’s model for predicting the equilibrium scour depth around USAF under random waves, a validation of the equilibrium scour depth Seq between the present study and Raaijmakers’s equation was conducted. The position where the scour depth Seq was evaluated is the location of the maximum scour depth, and it was depicted in Figure 14. The Figure 15 displays the comparison of Seq with different KC between the present study and Raaijmakers’s model.
Figure 14. Sketch of the position where the Seq was evaluated.
Figure 15. Comparison of the equilibrium scour depth between the present model and the model of Raaijmakers and Rudolph [34]: (a) KCrms,s, KCrms,a; (b) KCrms,p, KCrms,m; (c) KCrms,n = 1/10, KCrms,n = 1/5; (d) KCs,s, KCs,a; (e) KCs,p, KCs,m; (f) KCs,n = 1/10, KCs,n = 1/5.
As shown in Figure 15, there is an error in predicting Seq between the present study and Raaijmakers’s model, and Raaijmakers’s model underestimates the results generally. Although the error exists, the varying trend of Seq with KC obtained from Raaijmakers’s model is consistent with the present study basically. What’s more, the error is minimum and the Raaijmakers’s model is of relatively high accuracy for predicting scour around USAF under random waves by using KCs,p. Based on this, a further revision was made to eliminate the error as much as possible, i.e., add the deviation value ∆S/D in the Raaijmakers’s model. The revised equilibrium scour depth predicting equation based on Raaijmakers’s model can be written as
As the Figure 16 shown, through trial-calculation, when ∆S/D = 0.05, the results calculated by Equation (31) show good agreement with the simulating results of the present study. The maximum error is about 18.2% and the engineering requirements have been met basically. In order to further verify the accuracy of the revised model for large KC (KCs,p > 4) under random waves, a validation between the revised model and the previous experimental results [21]. The experiment was conducted in a flume (50 m in length, 1.0 m in width and 1.3 m in height) with a slender vertical pile (D = 0.1 m) under random waves. The seabed is composed of 0.13 m deep layer of sand with d50 = 0.6 mm and the water depth is 0.5 m for all tests. The significant wave height is 0.12~0.21 m and the KCs,p is 5.52~11.38. The comparison between the predicting results by Equation (31) and the experimental results of Corvaro et al. [21] is shown in Figure 17. From Figure 17, the experimental data evenly distributes around the predicted results and the prediction accuracy is favorable when KCs,p < 8. However, the gap between the predicting results and experimental data becomes large and the Equation (31) overestimates the equilibrium scour depth to some extent when KCs,p > 8.
Figure 16. Comparison of Seq between the simulating results and the predicting values by Equation (31).
Figure 17. Comparison of Seq/D between the Experimental results of Corvaro et al. [21] and the predicting values by Equation (31).
In ocean environment, the waves are composed of a train of sinusoidal waves with different frequencies and amplitudes. The energy of constituent waves with very large and very small frequencies is relatively low, and the energy of waves is mainly concentrated in a certain range of moderate frequencies. Myrhaug and Rue [37] thought the 1/n’th highest wave was responsible for scour and proposed the stochastic model to predict the equilibrium scour depth around pile under random waves for full range of KC. Noteworthy is that the KC was denoted by KCrms,a in the stochastic model. To verify the application of the stochastic model for predicting scour depth around USAF, a validation between the simulating results of present study and predicting results by the stochastic model with n = 2,3,5,10,20,500 was carried out respectively.
As shown in Figure 18, compared with the simulating results, the stochastic model underestimates the equilibrium scour depth around USAF generally. Although the error exists, the varying trend of Seq with KCrms,a obtained from the stochastic model is consistent with the present study basically. What’s more, the gap between the predicting values by stochastic model and the simulating results decreases with the increase of n, but for large n, for example n = 500, the varying trend diverges between the predicting values and simulating results, meaning it’s not feasible only by increasing n in stochastic model to predict the equilibrium scour depth around USAF.
Figure 18. Comparison of Seq between the simulating results and the predicting values by Equation (8).
The Figure 19 lists the deviation value ∆Seq/D′ between the predicting values and simulating results with different KCrms,a and n. Then, fitted the relationship between the ∆S′and n under different KCrms,a, and the fitting curve can be written by Equation (32). The revised stochastic model (Equation (33)) can be acquired by adding ∆Seq/D′ to Equation (8).
The comparison between the predicting results by Equation (33) and the simulating results of present study is shown in Figure 20. According to the Figure 20, the varying trend of Seq with KCrms,a obtained from the stochastic model is consistent with the present study basically. Compared with predicting results by the stochastic model, the results calculated by Equation (33) is favorable. Moreover, comparison with simulating results indicates that the predicting results are the most favorable for n = 10, which is consistent with the findings of Myrhaug and Rue [37] for equilibrium scour depth predicting around slender pile in case of random waves.
Figure 20. Comparison of Seq between the simulating results and the predicting values by Equation (33).
In order to further verify the accuracy of the Equation (33) for large KC (KCrms,a > 4) under random waves, a validation was conducted between the Equation (33) and the previous experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21]. The details of experiments conducted by Corvaro et al. [21] were described in above section. Sumer and Fredsøe [16] investigated the local scour around pile under random waves. The experiments were conducted in a wave basin with a slender vertical pile (D = 0.032, 0.055 m). The seabed is composed of 0.14 m deep layer of sand with d50 = 0.2 mm and the water depth was maintained at 0.5 m. The JONSWAP wave spectrum was used and the KCrms,a was 5.29~16.95. The comparison between the predicting results by Equation (33) and the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21] are shown in Figure 21. From Figure 21, contrary to the case of low KCrms,a (KCrms,a < 4), the error between the predicting values and experimental results increases with decreasing of n for KCrms,a > 4. Therefore, the predicting results are the most favorable for n = 2 when KCrms,a > 4.
Figure 21. Comparison of Seq between the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21] and the predicting values by Equation (33).
Noteworthy is that the present model was built according to prototype size, so the errors between the numerical results and experimental data of References [16,21] may be attribute to the scale effects. In laboratory experiments on scouring process, it is typically impossible to ensure a rigorous similarity of all physical parameters between the model and prototype structure, leading to the scale effects in the laboratory experiments. To avoid a cohesive behaviour, the bed material was not scaled geometrically according to model scale. As a consequence, the relatively large-scaled sediments sizes may result in the overestimation of bed load transport and underestimation of suspended load transport compared with field conditions. What’s more, the disproportional scaled sediment presumably lead to the difference of bed roughness between the model and prototype, and thus large influences for wave boundary layer on the seabed and scour process. Besides, according to Corvaro et al. [21] and Schendel et al. [55], the pile Reynolds numbers and Froude numbers both affect the scour depth for the condition of non fully developed turbulent flow in laboratory experiments.
4.4. Parametric Study
4.4.1. Influence of Froude Number
As described above, the set of foundation leads to the adverse pressure gradient appearing at upstream, leading to the wave boundary layer separating from seabed, then horseshoe vortex formatting and the horseshoe vortex are mainly responsible for scour around foundation (see Figure 22). The Froude number Fr is the key parameter to influence the scale and intensity of horseshoe vortex. The Fr under waves can be calculated by the following formula [42]
Fr=UwgD−−−√�r=�w��(34)
where Uw is the mean water particle velocity during 1/4 cycle of wave oscillation, obtained from the following formula. Noteworthy is that the root-mean-square (RMS) value of near-bed velocity amplitude Uwm,rms is used for calculating Uwm.
Figure 22. Sketch of flow field at upstream USAF edges.
Tavouktsoglou et al. [25] proposed the following formula between Fr and the vertical location of the stagnation y
yh∝Fer�ℎ∝�r�(36)
where e is constant.
The Figure 23 displays the relationship between Seq/D and Fr of the present study. In order to compare with the simulating results, the experimental data of Corvaro et al. [21] was also depicted in Figure 23. As shown in Figure 23, the equilibrium scour depth appears a logarithmic increase as Fr increases and approaches the mathematical asymptotic value, which is also consistent with the experimental results of Corvaro et al. [21]. According to Figure 24, the adverse pressure gradient pressure at upstream USAF edges increases with the increase of Fr, which is benefit for the wave boundary layer separating from seabed, resulting in the high-intensity horseshoe vortex, hence, causing intensive scour around USAF. Based on the previous study of Tavouktsoglou et al. [25] for scour around pile under currents, the high Fr leads to the stagnation point is closer to the mean sea level for shallow water, causing the stronger downflow kinetic energy. As mentioned in previous section, the energy of downflow at upstream makes up the energy of the subsequent horseshoe vortex, so the stronger downflow kinetic energy results in the more intensive horseshoe vortex. Therefore, the higher Fr leads to the more intensive horseshoe vortex by influencing the position of stagnation point y presumably. Qi and Gao [19] carried out a series of flume tests to investigate the scour around pile under regular waves, and proposed the fitting formula between Seq/D and Fr as following
lg(Seq/D)=Aexp(B/Fr)+Clg(�eq/�)=�exp(�/�r)+�(37)
where A, B and C are constant.
Figure 23. The fitting curve between Seq/D and Fr.
Figure 24. Sketch of adverse pressure gradient at upstream USAF edges.
Took the Equation (37) to fit the simulating results with A = −0.002, B = 0.686 and C = −0.808, and the results are shown in Figure 23. From Figure 23, the simulating results evenly distribute around the Equation (37) and the varying trend of Seq/D and Fr in present study is consistent with Equation (37) basically, meaning the Equation (37) is applicable to express the relationship of Seq/D with Fr around USAF under random waves.
4.4.2. Influence of Euler Number
The Euler number Eu is the influencing factor for the hydrodynamic field around foundation. The Eu under waves can be calculated by the following formula. The Eu can be represented by the Equation (38) for uniform cylinders [25]. The root-mean-square (RMS) value of near-bed velocity amplitude Um,rms is used for calculating Um.
Eu=U2mgD�u=�m2��(38)
where Um is depth-averaged flow velocity.
The Figure 25 displays the relationship between Seq/D and Eu of the present study. In order to compare with the simulating results, the experimental data of Sumer and Fredsøe [16] and Corvaro et al. [21] were also plotted in Figure 25. As shown in Figure 25, similar with the varying trend of Seq/D and Fr, the equilibrium scour depth appears a logarithmic increase as Eu increases and approaches the mathematical asymptotic value, which is also consistent with the experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21]. According to Figure 24, the adverse pressure gradient pressure at upstream USAF edges increases with the increasing of Eu, which is benefit for the wave boundary layer separating from seabed, inducing the high-intensity horseshoe vortex, hence, causing intensive scour around USAF.
Figure 25. The fitting curve between Seq/D and Eu.
Therefore, the variation of Fr and Eu reflect the magnitude of adverse pressure gradient pressure at upstream. Given that, the Equation (37) also was used to fit the simulating results with A = 8.875, B = 0.078 and C = −9.601, and the results are shown in Figure 25. From Figure 25, the simulating results evenly distribute around the Equation (37) and the varying trend of Seq/D and Eu in present study is consistent with Equation (37) basically, meaning the Equation (37) is also applicable to express the relationship of Seq/D with Eu around USAF under random waves. Additionally, according to the above description of Fr, it can be inferred that the higher Fr and Eu both lead to the more intensive horseshoe vortex by influencing the position of stagnation point y presumably.
5. Conclusions
A series of numerical models were established to investigate the local scour around umbrella suction anchor foundation (USAF) under random waves. The numerical model was validated for hydrodynamic and morphology parameters by comparing with the experimental data of Khosronejad et al. [52], Petersen et al. [17], Sumer and Fredsøe [16] and Schendel et al. [22]. Based on the simulating results, the scour evolution and scour mechanisms around USAF under random waves were analyzed respectively. Two revised models were proposed according to the model of Raaijmakers and Rudolph [34] and the stochastic model developed by Myrhaug and Rue [37] to predict the equilibrium scour depth around USAF under random waves. Finally, a parametric study was carried out with the present model to study the effects of the Froude number Fr and Euler number Eu to the equilibrium scour depth around USAF under random waves. The main conclusions can be described as follows.(1)
The packed sediment scour model and the RNG k−ε turbulence model were used to simulate the sand particles transport processes and the flow field around UASF respectively. The scour evolution obtained by the present model agrees well with the experimental results of Khosronejad et al. [52], Petersen et al. [17], Sumer and Fredsøe [16] and Schendel et al. [22], which indicates that the present model is accurate and reasonable for depicting the scour morphology around UASF under random waves.(2)
The vortex system at wave crest phase is mainly related to the scour process around USAF under random waves. The maximum scour depth appeared at the lee-side of the USAF at the initial stage (t < 1200 s). Subsequently, when t > 2400 s, the location of the maximum scour depth shifted to the upside of the USAF at an angle of about 45° with respect to the wave propagating direction.(3)
The error is negligible and the Raaijmakers’s model is of relatively high accuracy for predicting scour around USAF under random waves when KC is calculated by KCs,p. Given that, a further revision model (Equation (31)) was proposed according to Raaijmakers’s model to predict the equilibrium scour depth around USAF under random waves and it shows good agreement with the simulating results of the present study when KCs,p < 8.(4)
Another further revision model (Equation (33)) was proposed according to the stochastic model established by Myrhaug and Rue [37] to predict the equilibrium scour depth around USAF under random waves, and the predicting results are the most favorable for n = 10 when KCrms,a < 4. However, contrary to the case of low KCrms,a, the predicting results are the most favorable for n = 2 when KCrms,a > 4 by the comparison with experimental results of Sumer and Fredsøe [16] and Corvaro et al. [21].(5)
The same formula (Equation (37)) is applicable to express the relationship of Seq/D with Eu or Fr, and it can be inferred that the higher Fr and Eu both lead to the more intensive horseshoe vortex and larger Seq.
Author Contributions
Conceptualization, H.L. (Hongjun Liu); Data curation, R.H. and P.Y.; Formal analysis, X.W. and H.L. (Hao Leng); Funding acquisition, X.W.; Writing—original draft, R.H. and P.Y.; Writing—review & editing, X.W. and H.L. (Hao Leng); The final manuscript has been approved by all the authors. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the Fundamental Research Funds for the Central Universities (grant number 202061027) and the National Natural Science Foundation of China (grant number 41572247).
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data presented in this study are available on request from the corresponding author.
Conflicts of Interest
The authors declare no conflict of interest.
References
Sumer, B.M.; Fredsøe, J.; Christiansen, N. Scour Around Vertical Pile in Waves. J. Waterw. Port. Coast. Ocean Eng.1992, 118, 15–31. [Google Scholar] [CrossRef]
Rudolph, D.; Bos, K. Scour around a monopile under combined wave-current conditions and low KC-numbers. In Proceedings of the 6th International Conference on Scour and Erosion, Amsterdam, The Netherlands, 1–3 November 2006; pp. 582–588. [Google Scholar]
Nielsen, A.W.; Liu, X.; Sumer, B.M.; Fredsøe, J. Flow and bed shear stresses in scour protections around a pile in a current. Coast. Eng.2013, 72, 20–38. [Google Scholar] [CrossRef]
Ahmad, N.; Bihs, H.; Myrhaug, D.; Kamath, A.; Arntsen, Ø.A. Three-dimensional numerical modelling of wave-induced scour around piles in a side-by-side arrangement. Coast. Eng.2018, 138, 132–151. [Google Scholar] [CrossRef]
Li, H.; Ong, M.C.; Leira, B.J.; Myrhaug, D. Effects of Soil Profile Variation and Scour on Structural Response of an Offshore Monopile Wind Turbine. J. Offshore Mech. Arct. Eng.2018, 140, 042001. [Google Scholar] [CrossRef]
Li, H.; Liu, H.; Liu, S. Dynamic analysis of umbrella suction anchor foundation embedded in seabed for offshore wind turbines. Géoméch. Energy Environ.2017, 10, 12–20. [Google Scholar] [CrossRef]
Fazeres-Ferradosa, T.; Rosa-Santos, P.; Taveira-Pinto, F.; Vanem, E.; Carvalho, H.; Correia, J.A.F.D.O. Editorial: Advanced research on offshore structures and foundation design: Part 1. Proc. Inst. Civ. Eng. Marit. Eng.2019, 172, 118–123. [Google Scholar] [CrossRef]
Chavez, C.E.A.; Stratigaki, V.; Wu, M.; Troch, P.; Schendel, A.; Welzel, M.; Villanueva, R.; Schlurmann, T.; De Vos, L.; Kisacik, D.; et al. Large-Scale Experiments to Improve Monopile Scour Protection Design Adapted to Climate Change—The PROTEUS Project. Energies2019, 12, 1709. [Google Scholar] [CrossRef][Green Version]
Wu, M.; De Vos, L.; Chavez, C.E.A.; Stratigaki, V.; Fazeres-Ferradosa, T.; Rosa-Santos, P.; Taveira-Pinto, F.; Troch, P. Large Scale Experimental Study of the Scour Protection Damage Around a Monopile Foundation Under Combined Wave and Current Conditions. J. Mar. Sci. Eng.2020, 8, 417. [Google Scholar] [CrossRef]
Sørensen, S.P.H.; Ibsen, L.B. Assessment of foundation design for offshore monopiles unprotected against scour. Ocean Eng.2013, 63, 17–25. [Google Scholar] [CrossRef]
Prendergast, L.; Gavin, K.; Doherty, P. An investigation into the effect of scour on the natural frequency of an offshore wind turbine. Ocean Eng.2015, 101, 1–11. [Google Scholar] [CrossRef][Green Version]
Fazeres-Ferradosa, T.; Chambel, J.; Taveira-Pinto, F.; Rosa-Santos, P.; Taveira-Pinto, F.; Giannini, G.; Haerens, P. Scour Protections for Offshore Foundations of Marine Energy Harvesting Technologies: A Review. J. Mar. Sci. Eng.2021, 9, 297. [Google Scholar] [CrossRef]
Yang, Q.; Yu, P.; Liu, Y.; Liu, H.; Zhang, P.; Wang, Q. Scour characteristics of an offshore umbrella suction anchor foundation under the combined actions of waves and currents. Ocean Eng.2020, 202, 106701. [Google Scholar] [CrossRef]
Yu, P.; Hu, R.; Yang, J.; Liu, H. Numerical investigation of local scour around USAF with different hydraulic conditions under currents and waves. Ocean Eng.2020, 213, 107696. [Google Scholar] [CrossRef]
Sumer, B.M.; Christiansen, N.; Fredsøe, J. The horseshoe vortex and vortex shedding around a vertical wall-mounted cylinder exposed to waves. J. Fluid Mech.1997, 332, 41–70. [Google Scholar] [CrossRef]
Sumer, B.M.; Fredsøe, J. Scour around Pile in Combined Waves and Current. J. Hydraul. Eng.2001, 127, 403–411. [Google Scholar] [CrossRef]
Petersen, T.U.; Sumer, B.M.; Fredsøe, J. Time scale of scour around a pile in combined waves and current. In Proceedings of the 6th International Conference on Scour and Erosion, Paris, France, 27–31 August 2012. [Google Scholar]
Petersen, T.U.; Sumer, B.M.; Fredsøe, J.; Raaijmakers, T.C.; Schouten, J.-J. Edge scour at scour protections around piles in the marine environment—Laboratory and field investigation. Coast. Eng.2015, 106, 42–72. [Google Scholar] [CrossRef]
Qi, W.; Gao, F. Equilibrium scour depth at offshore monopile foundation in combined waves and current. Sci. China Ser. E Technol. Sci.2014, 57, 1030–1039. [Google Scholar] [CrossRef][Green Version]
Corvaro, S.; Marini, F.; Mancinelli, A.; Lorenzoni, C.; Brocchini, M. Hydro- and Morpho-dynamics Induced by a Vertical Slender Pile under Regular and Random Waves. J. Waterw. Port. Coast. Ocean Eng.2018, 144, 04018018. [Google Scholar] [CrossRef]
Schendel, A.; Welzel, M.; Schlurmann, T.; Hsu, T.-W. Scour around a monopile induced by directionally spread irregular waves in combination with oblique currents. Coast. Eng.2020, 161, 103751. [Google Scholar] [CrossRef]
Fazeres-Ferradosa, T.; Taveira-Pinto, F.; Romão, X.; Reis, M.; das Neves, L. Reliability assessment of offshore dynamic scour protections using copulas. Wind. Eng.2018, 43, 506–538. [Google Scholar] [CrossRef]
Fazeres-Ferradosa, T.; Welzel, M.; Schendel, A.; Baelus, L.; Santos, P.R.; Pinto, F.T. Extended characterization of damage in rubble mound scour protections. Coast. Eng.2020, 158, 103671. [Google Scholar] [CrossRef]
Ettema, R.; Melville, B.; Barkdoll, B. Scale Effect in Pier-Scour Experiments. J. Hydraul. Eng.1998, 124, 639–642. [Google Scholar] [CrossRef]
Umeda, S. Scour Regime and Scour Depth around a Pile in Waves. J. Coast. Res. Spec. Issue2011, 64, 845–849. [Google Scholar]
Umeda, S. Scour process around monopiles during various phases of sea storms. J. Coast. Res.2013, 165, 1599–1604. [Google Scholar] [CrossRef]
Baykal, C.; Sumer, B.; Fuhrman, D.R.; Jacobsen, N.; Fredsøe, J. Numerical simulation of scour and backfilling processes around a circular pile in waves. Coast. Eng.2017, 122, 87–107. [Google Scholar] [CrossRef][Green Version]
Miles, J.; Martin, T.; Goddard, L. Current and wave effects around windfarm monopile foundations. Coast. Eng.2017, 121, 167–178. [Google Scholar] [CrossRef][Green Version]
Miozzi, M.; Corvaro, S.; Pereira, F.A.; Brocchini, M. Wave-induced morphodynamics and sediment transport around a slender vertical cylinder. Adv. Water Resour.2019, 129, 263–280. [Google Scholar] [CrossRef]
Yu, T.; Zhang, Y.; Zhang, S.; Shi, Z.; Chen, X.; Xu, Y.; Tang, Y. Experimental study on scour around a composite bucket foundation due to waves and current. Ocean Eng.2019, 189, 106302. [Google Scholar] [CrossRef]
Carreiras, J.; Larroudé, P.; Seabra-Santos, F.; Mory, M. Wave Scour Around Piles. In Proceedings of the Coastal Engineering 2000, American Society of Civil Engineers (ASCE), Sydney, Australia, 16–21 July 2000; pp. 1860–1870. [Google Scholar]
Raaijmakers, T.; Rudolph, D. Time-dependent scour development under combined current and waves conditions—Laboratory experiments with online monitoring technique. In Proceedings of the 4th International Conference on Scour and Erosion, Tokyo, Japan, 5–7 November 2008; pp. 152–161. [Google Scholar]
Khalfin, I.S. Modeling and calculation of bed score around large-diameter vertical cylinder under wave action. Water Resour.2007, 34, 357. [Google Scholar] [CrossRef][Green Version]
Zanke, U.C.; Hsu, T.-W.; Roland, A.; Link, O.; Diab, R. Equilibrium scour depths around piles in noncohesive sediments under currents and waves. Coast. Eng.2011, 58, 986–991. [Google Scholar] [CrossRef]
Myrhaug, D.; Rue, H. Scour below pipelines and around vertical piles in random waves. Coast. Eng.2003, 48, 227–242. [Google Scholar] [CrossRef]
Myrhaug, D.; Ong, M.C.; Føien, H.; Gjengedal, C.; Leira, B.J. Scour below pipelines and around vertical piles due to second-order random waves plus a current. Ocean Eng.2009, 36, 605–616. [Google Scholar] [CrossRef]
Myrhaug, D.; Ong, M.C. Random wave-induced onshore scour characteristics around submerged breakwaters using a stochastic method. Ocean Eng.2010, 37, 1233–1238. [Google Scholar] [CrossRef]
Ong, M.C.; Myrhaug, D.; Hesten, P. Scour around vertical piles due to long-crested and short-crested nonlinear random waves plus a current. Coast. Eng.2013, 73, 106–114. [Google Scholar] [CrossRef]
Yakhot, V.; Orszag, S.A. Renormalization group analysis of turbulence. I. Basic theory. J. Sci. Comput.1986, 1, 3–51. [Google Scholar] [CrossRef]
Yakhot, V.; Smith, L.M. The renormalization group, the e-expansion and derivation of turbulence models. J. Sci. Comput.1992, 7, 35–61. [Google Scholar] [CrossRef]
Mastbergen, D.R.; Berg, J.V.D. Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons. Sedimentology2003, 50, 625–637. [Google Scholar] [CrossRef]
Soulsby, R. Dynamics of Marine Sands; Thomas Telford Ltd.: London, UK, 1998. [Google Scholar] [CrossRef]
Van Rijn, L.C. Sediment Transport, Part I: Bed Load Transport. J. Hydraul. Eng.1984, 110, 1431–1456. [Google Scholar] [CrossRef][Green Version]
Zhang, Q.; Zhou, X.-L.; Wang, J.-H. Numerical investigation of local scour around three adjacent piles with different arrangements under current. Ocean Eng.2017, 142, 625–638. [Google Scholar] [CrossRef]
Yu, Y.X.; Liu, S.X. Random Wave and Its Applications to Engineering, 4th ed.; Dalian University of Technology Press: Dalian, China, 2011. [Google Scholar]
Pang, A.; Skote, M.; Lim, S.; Gullman-Strand, J.; Morgan, N. A numerical approach for determining equilibrium scour depth around a mono-pile due to steady currents. Appl. Ocean Res.2016, 57, 114–124. [Google Scholar] [CrossRef]
Higuera, P.; Lara, J.L.; Losada, I.J. Three-dimensional interaction of waves and porous coastal structures using Open-FOAM®. Part I: Formulation and validation. Coast. Eng.2014, 83, 243–258. [Google Scholar] [CrossRef]
Corvaro, S.; Crivellini, A.; Marini, F.; Cimarelli, A.; Capitanelli, L.; Mancinelli, A. Experimental and Numerical Analysis of the Hydrodynamics around a Vertical Cylinder in Waves. J. Mar. Sci. Eng.2019, 7, 453. [Google Scholar] [CrossRef][Green Version]
Flow3D User Manual, version 11.0.3; Flow Science, Inc.: Santa Fe, NM, USA, 2013.
Khosronejad, A.; Kang, S.; Sotiropoulos, F. Experimental and computational investigation of local scour around bridge piers. Adv. Water Resour.2012, 37, 73–85. [Google Scholar] [CrossRef]
Stahlmann, A. Experimental and Numerical Modeling of Scour at Foundation Structures for Offshore Wind Turbines. Ph.D. Thesis, Franzius-Institute for Hydraulic, Estuarine and Coastal Engineering, Leibniz Universität Hannover, Hannover, Germany, 2013. [Google Scholar]
Breusers, H.N.C.; Nicollet, G.; Shen, H. Local Scour Around Cylindrical Piers. J. Hydraul. Res.1977, 15, 211–252. [Google Scholar] [CrossRef]
Schendel, A.; Hildebrandt, A.; Goseberg, N.; Schlurmann, T. Processes and evolution of scour around a monopile induced by tidal currents. Coast. Eng.2018, 139, 65–84. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Hu, R.; Liu, H.; Leng, H.; Yu, P.; Wang, X. Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves. J. Mar. Sci. Eng.2021, 9, 886. https://doi.org/10.3390/jmse9080886
AMA Style
Hu R, Liu H, Leng H, Yu P, Wang X. Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves. Journal of Marine Science and Engineering. 2021; 9(8):886. https://doi.org/10.3390/jmse9080886Chicago/Turabian Style
Hu, Ruigeng, Hongjun Liu, Hao Leng, Peng Yu, and Xiuhai Wang. 2021. “Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves” Journal of Marine Science and Engineering 9, no. 8: 886. https://doi.org/10.3390/jmse9080886
Find Other Styles
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
For more information on the journal statistics, click here.
Multiple requests from the same IP address are counted as one view.
Alkistis Stergiopoulou1, Vassilios Stergiopoulos2 1Institut für Wasserwirtschaft, Hydrologie und Konstruktiven Wasserbau, B.O.K.U. University, Muthgasse 18, 1190 Vienna, (actually Senior Process Engineer at the VTU Engineering in Vienna, Zieglergasse 53/1/24, 1070 Vienna, Austria). 2 School of Pedagogical and Technological Education, Department of Civil Engineering Educators, ASPETE Campus, Eirini Station, 15122 Amarousio, Athens, Greece. Received 4 Jan. 2021; Received in revised form 8 Aug. 2021; Accepted 8 Aug. 2021; Available online 14 Aug. 2021
Abstract
This paper presents a short view of the first Archimedean Screw Turbines CFD modelling results, which were carried out within the recent research entitled “Rebirth of Archimedes in Greece: contribution to the study of hydraulic mechanics and hydrodynamic behavior of Archimedean cochlear waterwheels, for recovering the hydraulic potential of Greek natural and technical watercourses”. This CFD analysis, based to the Flow-3D code, concerns typical Tubular Archimedean Screw Turbines (TASTs) and shows some promising performances for such small hydropower systems harnessing the important unexploited hydraulic potential of natural and technical watercourses of Greece, of the order of several TWh / year and of a total installed capacity in the range of thousands MWs.
CFD; Flow-3D; TAST; Small Hydro; Renewable Energy; Greek Watercourses.
Figure 1. Photorealistic view of an inclined axis TAST (photo A. Stergiopoulou).
References.
[1] A. Stergiopoulou, Computational and experimental investigation of the hydrodynamic behaviour of screw hydro turbine, Ph.D. Thesis, NTUA, 2017. [2] B. Pelikan, A. Lashofer, Verbesserung der Strömungseigenschaften sowie Planungs-und Betriebsoptimierung von Wasserkraftschnecken, Research Project, BOKU University, Vienna, 2012. [3] G. Müller, J. Senior, Simplified theory of Archimedean screws, Journal of Hydraulic Research 47 (5) (2009) 666-669. [4] C. Rorres, The turn of the screw: Optimal design of an Archimedes screw, Journal of Hydraulic Engineering, 80 (2000) 72-80. [5] A. Stergiopoulou, V. Stergiopoulos, Return of Archimedes: Harnessing with new Archimedean spirals the hydraulic potential of the Greek watercourses, in: Proceedings of the Conference for Climate Change, Thessaloniki, 2009. [6] A. Stergiopoulou, V. Stergiopoulos, from the old Archimedean screw pumps to the new Archimedean screw turbines for hydropower production in Greece, in: Proceedings of CEMEPE Conference, Mykonos, June 21-26, 2009. International Journal of Energy and Environment (IJEE), Volume 12, Issue 1, 2021, pp.19-30 [7] V. Stergiopoulos, A. Stergiopoulou, E. Kalkani, Quo Vadis Archimedes Nowadays in Greece? Towards Modern Archimedean Turbines for Recovering Greek Small Hydropower Potential, in: Proceedings of 3rd International Scientific “Energy and Climate Change” Conference, Athens, 2010. [8] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Greece beyond the horizon of the era of transition: Archimedean screw hydropower development terra incognita, International Journal of Energy and Development, v.6, Issue 6, pp. 627-536, 2015. [9] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Experimental and theoretical research of zero head innovative horizontal axis Archimedean screw turbines, Journal of Energy and Development, v.6, Issue 5, pp. 471-478, 2015. [10] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Back to the Future: Rediscovering the Archimedean screws as modern turbines for harnessing Greek small hydropower potential, in: Proceedings of the Third International Conference CEMEPE 2011 & SECOTOX, Skiathos, 2011. [11] A. Stergiopoulou, V. Stergiopoulos, Educational Renewable Energy Screw Wheel Technologies for Pico Hydropower Generation, Modern Environmental Science and Engineering, v.4, No.5, pp. 439- 445, May 2018. [12] A. Stergiopoulou, V. Stergiopoulos, Educational Renewable Energy Screw Wheel Technologies for Pico Hydropower Generation, Modern Environmental Science and Engineering, v.4, No.5, pp. 439- 445, May 2018. [13] A. Stergiopoulou, V. Stergiopoulos, Towards an inventory of the archimedean small hydropower potential of Greece, INTERNATIONAL JOURNAL OF ENERGY AND ENVIRONMENT Volume 11, Issue 2, 2020 pp.137-144. [14] Flow Science, FLOW-3D Manual, 2013. [15] K. Versteeg and W. Malalasekera, An Introduction to Computational Fluid Dynamics, Pearson, 2007. [16] C. Hirsch, Numerical Computation of internal and external flows: The fundamentals of Computational Fluid dynamics, John Wiley & Sons, 2007. [17] A. Stergiopoulou, V. Stergiopoulos and E. Kalkani, An eagle’s CFD view of Studying Innovative Archimedean Screw Renewable Hydraulic Energy Systems, Proceedings of the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE) and SECOTOX Conference, Mykonos island, Greece, pp.454-460 June 24-28, 2013. [18] A. Stergiopoulou, V. Stergiopoulos, A., E. Kalkani, Computational Fluid Dynamics Study on a 3D Graphic Solid Model of Archimedean Screw Turbines, Fresenius Environmental Bulletin, vol.23- No1, 2014. [19] Α. Stergiopoulou, Kalkani E., “Towards a First C.F.D. Study of Innovative Archimedean Inclined Axis Hydropower Turbines”, International Journal of Engineering Research & Technology (IJERT), Vol. 2 Issue 9, September – 2013, pp. 193-199. [20] A. Stergiopoulou, V. Stergiopoulos, A first CFD study of small hydro energy recovery from the Attica water supply network, INTERNATIONAL JOURNAL OF ENERGY AND ENVIRONMENT, Volume 11, Issue 3, 2020 pp.157-166.
이 백서는 Flow-3D를 적용하여 다양한 흐름 배출 및 식생 시나리오가 흐름 속도(세로, 가로 및 수직 속도 포함)에 미치는 영향을 조사합니다.
실험적 측정을 통한 검증 후 식생직경, 식생높이, 유량방류량에 대한 민감도 분석을 수행하였다. 종방향 속도의 경우 흐름 구조에 가장 큰 영향을 미치는 것은 배출보다는 식생 직경에서 비롯됩니다.
그러나 식생 높이는 수직 분포의 변곡점을 결정합니다. 식생지 내 두 지점, 즉 상류와 하류의 횡속도를 비교하면 수심에 따른 대칭적인 패턴을 확인할 수 있다. 식생 지역의 가로 및 세로 유체 순환 패턴을 포함하여 흐름 또는 식생 시나리오와 관계없이 수직 속도에 대해서도 동일한 패턴이 관찰됩니다.
또한 식생의 직경이 클수록 이러한 패턴이 더 분명해집니다. 상부 순환은 초목 캐노피 근처에서 발생합니다. 식생지역의 가로방향과 세로방향의 순환에 관한 이러한 발견은 침수식생을 통한 3차원 유동구조를 밝혀준다.
This paper applies the Flow-3D to investigate the impacts of different flow discharge and vegetation scenarios on the flow velocity (including the longitudinal, transverse and vertical velocities). After the verification by using experimental measurements, a sensitivity analysis is conducted for the vegetation diameter, the vegetation height and the flow discharge. For the longitudinal velocity, the greatest impact on the flow structure originates from the vegetation diameter, rather than the discharge. The vegetation height, however, determines the inflection point of the vertical distribution. Comparing the transverse velocities at two positions in the vegetated area, i.e., the upstream and the downstream, a symmetric pattern is identified along the water depth. The same pattern is also observed for the vertical velocity regardless of the flow or vegetation scenario, including both transverse and vertical fluid circulation patterns in the vegetated area. Moreover, the larger the vegetation diameter is, the more evident these patterns become. The upper circulation occurs near the vegetation canopy. These findings regarding the circulations along the transverse and vertical directions in the vegetated region shed light on the 3-D flow structure through the submerged vegetation.
Key words
Submerged rigid vegetation
longitudinal velocity
transverse velocity
vertical velocity
open channel
References
Angelina A., Jordanova C. S. J. Experimental study of bed load transport through emergent vegetation [J]. Journal of Hydraulic Engineering, ASCE, 2003, 129(6): 474–478.ArticleGoogle Scholar
Li Y., Wang Y., Anim D. O. et al. Flow characteristics in different densities of submerged flexible vegetation from an open-channel flume study of artificial plants [J]. Geomorphology, 2014, 204: 314–324.ArticleGoogle Scholar
Bai F., Yang Z., Huai W. et al. A depth-averaged two dimensional shallow water model to simulate flow-rigid vegetation interactions [J]. Procedia Engineering, 2016, 154: 482–489.ArticleGoogle Scholar
Huai W. X., Song S., Han J. et al. Prediction of velocity distribution in straight open-channel flow with partial vegetation by singular perturbation method [J]. Applied Mathematics and Mechanics (Engilsh Editon), 2016, 37(10): 1315–1324.ArticleMathSciNetGoogle Scholar
Wang P. F., Wang C. Numerical model for flow through submerged vegetation regions in a shallow lake [J]. Journal of Hydrodynamics, 2011, 23(2): 170–178.ArticleGoogle Scholar
Wang W. J., Cui X. Y., Dong F. et al. Predictions of bulk velocity for open channel flow through submerged vegetation [J]. Journal of Hydrodynamics, 2020, 32(4): 795–799.ArticleGoogle Scholar
Zhang M., Li C. W., Shen Y. Depth-averaged modeling of free surface flows in open channels with emerged and submerged vegetation [J]. Applied Mathematical Modelling, 2013, 37(1–2): 540–553.ArticleMathSciNetGoogle Scholar
Huai W., Wang W., Hu Y. et al. Analytical model of the mean velocity distribution in an open channel with double-layered rigid vegetation [J]. Advances in Water Resources, 2014, 69: 106–113.ArticleGoogle Scholar
Panigrahi K., Khatua K. K. Prediction of velocity distribution in straight channel with rigid vegetation [J]. Aquatic Procedia, 2015, 4: 819–825.ArticleGoogle Scholar
Huai W. X., Zeng Y. H., Xu Z. G. et al. Three-layer model for vertical velocity distribution in open channel flow with submerged rigid vegetation [J]. Advances in Water Resources, 2009, 32(4): 487–492.ArticleGoogle Scholar
Chen S. C., Kuo Y. M., Li Y. H. Flow characteristics within different configurations of submerged flexible vegetation [J]. Journal of Hydrology, 2011, 398(1–2): 124–134.ArticleGoogle Scholar
Yagci O., Tschiesche U., Kabdasli M. S. The role of different forms of natural riparian vegetation on turbulence and kinetic energy characteristics [J]. Advances in Water Resources, 2010, 33(5): 601–614.ArticleGoogle Scholar
Wu F. S. Characteristics of flow resistance in open channels with non-submerged rigid vegetation [J]. Journal of Hydrodynamics, 2008, 20(2): 239–245.ArticleGoogle Scholar
Huai W., Hu Y., Zeng Y. et al. Velocity distribution for open channel flows with suspended vegetation [J]. Advances in Water Resources, 2012, 49: 56–61.ArticleGoogle Scholar
Pu J. H., Hussain A., Guo Y. K. et al. Submerged flexible vegetation impact on open channel flow velocity distribution: An analytical modelling study on drag and friction [J]. Water Science and Engineering, 2019, 12(2): 121–128.ArticleGoogle Scholar
Zhang M. L., Li C. W., Shen Y. M. A 3D non-linear k-ε turbulent model for prediction of flow and mass transport in channel with vegetation [J]. Applied Mathematical Modelling, 2010, 34(4): 1021–1031.ArticleMathSciNetGoogle Scholar
Anjum N., Tanaka N. Numerical investigation of velocity distribution of turbulent flow through vertically double-layered vegetation [J]. Water Science and Engineering, 2019, 12(4): 319–329.ArticleGoogle Scholar
Wang W., Huai W. X., Gao M. Numerical investigation of flow through vegetated multi-stage compound channel [J]. Journal of Hydrodynamics, 2014, 26(3): 467–473.ArticleGoogle Scholar
Ghani U., Anjum N., Pasha G. A. et al. Numerical investigation of the flow characteristics through discontinuous and layered vegetation patches of finite width in an open channel [J]. Environmental Fluid Mechanics, 2019, 19(6): 1469–1495.ArticleGoogle Scholar
Aydin M. C., Emiroglu M. E. Determination of capacity of labyrinth side weir by CFD [J]. Flow Measurement and Instrumentation, 2013, 29: 1–8.ArticleGoogle Scholar
Hao W. L., Wu W. Q., Zhu C. J. et al. Experimental study on vertical distribution of flow velocity in vegetated river channel [J]. Water Resources and Power, 2015, 33(2): 85–88(in Chinese).Google Scholar
Pietri L., Petroff A., Amielh M. et al. Turbulent flows interacting with varying density canopies [J]. Mécanique and Industries, 2009, 10(3–4): 181–185.ArticleGoogle Scholar
Li Y., Du W., Yu Z. et al. Impact of flexible emergent vegetation on the flow turbulence and kinetic energy characteristics in a flume experiment [J]. Journal of Hydro-environment Research, 2015, 9(3): 354–367.ArticleGoogle Scholar
Li W. Q., Wang D., Jiao J. L. et al. Effects of vegetation patch density on flow velocity characteristics in an open channel [J]. Journal of Hydrodynamics, 2018, 31(5): 1052–1059.ArticleGoogle Scholar
Langre E. D., Gutierrez A., Cossé J. On the scaling of drag reduction by reconfiguration in plants [J]. Comptes Rendus Mécanique, 2012, 340(1–2): 35–40.ArticleGoogle Scholar
Fathi-Maghadam M., Kouwen N. Nonrigid, nonsubmerged, vegetative roughness on floodplains [J]. Journal of Hydraulic Engineering, ASCE, 1997, 123(1): 51–57.ArticleGoogle Scholar
Liang D., Wu X. A random walk simulation of scalar mixing in flows through submerged vegetations [J]. Journal of Hydrodynamics, 2014, 26(3): 343–350.ArticleMathSciNetGoogle Scholar
Ghisalberti M., Nepf H. Mass transport in vegetated shear flows [J]. Environmental Fluid Mechanics, 2005, 5(6): 527–551.ArticleGoogle Scholar
Thermo-fluid modeling of influence of attenuated laser beam intensity profile on melt pool behavior in laser-assisted powder-based direct energy deposition
Mohammad Sattari, Amin Ebrahimi, Martin Luckabauer, Gert-willem R.B.E. Römer
A numerical framework based on computational fluid dynamics (CFD), using the finite volume method (FVM) and volume of fluid (VOF) technique is presented to investigate the effect of the laser beam intensity profile on melt pool behavior in laser-assisted powder-based directed energy deposition (L-DED). L-DED is an additive manufacturing (AM) process that utilizes a laser beam to fuse metal powder particles. To assure high-fidelity modeling, it was found that it is crucial to accurately model the interaction between the powder stream and the laser beam in the gas region above the substrate. The proposed model considers various phenomena including laser energy attenuation and absorption, multiple reflections of the laser rays, powder particle stream, particle-fluid interaction, temperature-dependent properties, buoyancy effects, thermal expansion, solidification shrinkage and drag, and Marangoni flow. The latter is induced by temperature and element-dependent surface tension. The model is validated using experimental results and highlights the importance of considering laser energy attenuation. Furthermore, the study investigates how the laser beam intensity profile affects melt pool size and shape, influencing the solidification microstructure and mechanical properties of the deposited material. The proposed model has the potential to optimize the L-DED process for a variety of materials and provides insights into the capability of numerical modeling for additive manufacturing optimization.
Mahdi Feizbahr,1Navid Tonekaboni,2Guang-Jun Jiang,3,4and Hong-Xia Chen3,4 Academic Editor: Mohammad Yazdi
Abstract
강을 따라 식생은 조도를 증가시키고 평균 유속을 감소시키며, 유동 에너지를 감소시키고 강 횡단면의 유속 프로파일을 변경합니다. 자연의 많은 운하와 강은 홍수 동안 초목으로 덮여 있습니다. 운하의 조도는 식물의 영향을 많이 받기 때문에 홍수시 유동저항에 큰 영향을 미친다. 식물로 인한 흐름에 대한 거칠기 저항은 흐름 조건과 식물에 따라 달라지므로 모델은 유속, 유속 깊이 및 수로를 따라 식생 유형의 영향을 고려하여 유속을 시뮬레이션해야 합니다. 총 48개의 모델을 시뮬레이션하여 근관의 거칠기 효과를 조사했습니다. 결과는 속도를 높임으로써 베드 속도를 감소시키는 식생의 영향이 무시할만하다는 것을 나타냅니다.
Abstract
Vegetation along the river increases the roughness and reduces the average flow velocity, reduces flow energy, and changes the flow velocity profile in the cross section of the river. Many canals and rivers in nature are covered with vegetation during the floods. Canal’s roughness is strongly affected by plants and therefore it has a great effect on flow resistance during flood. Roughness resistance against the flow due to the plants depends on the flow conditions and plant, so the model should simulate the current velocity by considering the effects of velocity, depth of flow, and type of vegetation along the canal. Total of 48 models have been simulated to investigate the effect of roughness in the canal. The results indicated that, by enhancing the velocity, the effect of vegetation in decreasing the bed velocity is negligible, while when the current has lower speed, the effect of vegetation on decreasing the bed velocity is obviously considerable.
1. Introduction
Considering the impact of each variable is a very popular field within the analytical and statistical methods and intelligent systems [1–14]. This can help research for better modeling considering the relation of variables or interaction of them toward reaching a better condition for the objective function in control and engineering [15–27]. Consequently, it is necessary to study the effects of the passive factors on the active domain [28–36]. Because of the effect of vegetation on reducing the discharge capacity of rivers [37], pruning plants was necessary to improve the condition of rivers. One of the important effects of vegetation in river protection is the action of roots, which cause soil consolidation and soil structure improvement and, by enhancing the shear strength of soil, increase the resistance of canal walls against the erosive force of water. The outer limbs of the plant increase the roughness of the canal walls and reduce the flow velocity and deplete the flow energy in vicinity of the walls. Vegetation by reducing the shear stress of the canal bed reduces flood discharge and sedimentation in the intervals between vegetation and increases the stability of the walls [38–41].
One of the main factors influencing the speed, depth, and extent of flood in this method is Manning’s roughness coefficient. On the other hand, soil cover [42], especially vegetation, is one of the most determining factors in Manning’s roughness coefficient. Therefore, it is expected that those seasonal changes in the vegetation of the region will play an important role in the calculated value of Manning’s roughness coefficient and ultimately in predicting the flood wave behavior [43–45]. The roughness caused by plants’ resistance to flood current depends on the flow and plant conditions. Flow conditions include depth and velocity of the plant, and plant conditions include plant type, hardness or flexibility, dimensions, density, and shape of the plant [46]. In general, the issue discussed in this research is the optimization of flood-induced flow in canals by considering the effect of vegetation-induced roughness. Therefore, the effect of plants on the roughness coefficient and canal transmission coefficient and in consequence the flow depth should be evaluated [47, 48].
Current resistance is generally known by its roughness coefficient. The equation that is mainly used in this field is Manning equation. The ratio of shear velocity to average current velocity is another form of current resistance. The reason for using the ratio is that it is dimensionless and has a strong theoretical basis. The reason for using Manning roughness coefficient is its pervasiveness. According to Freeman et al. [49], the Manning roughness coefficient for plants was calculated according to the Kouwen and Unny [50] method for incremental resistance. This method involves increasing the roughness for various surface and plant irregularities. Manning’s roughness coefficient has all the factors affecting the resistance of the canal. Therefore, the appropriate way to more accurately estimate this coefficient is to know the factors affecting this coefficient [51].
To calculate the flow rate, velocity, and depth of flow in canals as well as flood and sediment estimation, it is important to evaluate the flow resistance. To determine the flow resistance in open ducts, Manning, Chézy, and Darcy–Weisbach relations are used [52]. In these relations, there are parameters such as Manning’s roughness coefficient (n), Chézy roughness coefficient (C), and Darcy–Weisbach coefficient (f). All three of these coefficients are a kind of flow resistance coefficient that is widely used in the equations governing flow in rivers [53].
The three relations that express the relationship between the average flow velocity (V) and the resistance and geometric and hydraulic coefficients of the canal are as follows:where n, f, and c are Manning, Darcy–Weisbach, and Chézy coefficients, respectively. V = average flow velocity, R = hydraulic radius, Sf = slope of energy line, which in uniform flow is equal to the slope of the canal bed, = gravitational acceleration, and Kn is a coefficient whose value is equal to 1 in the SI system and 1.486 in the English system. The coefficients of resistance in equations (1) to (3) are related as follows:
Based on the boundary layer theory, the flow resistance for rough substrates is determined from the following general relation:where f = Darcy–Weisbach coefficient of friction, y = flow depth, Ks = bed roughness size, and A = constant coefficient.
On the other hand, the relationship between the Darcy–Weisbach coefficient of friction and the shear velocity of the flow is as follows:
By using equation (6), equation (5) is converted as follows:
Investigation on the effect of vegetation arrangement on shear velocity of flow in laboratory conditions showed that, with increasing the shear Reynolds number (), the numerical value of the ratio also increases; in other words the amount of roughness coefficient increases with a slight difference in the cases without vegetation, checkered arrangement, and cross arrangement, respectively [54].
Roughness in river vegetation is simulated in mathematical models with a variable floor slope flume by different densities and discharges. The vegetation considered submerged in the bed of the flume. Results showed that, with increasing vegetation density, canal roughness and flow shear speed increase and with increasing flow rate and depth, Manning’s roughness coefficient decreases. Factors affecting the roughness caused by vegetation include the effect of plant density and arrangement on flow resistance, the effect of flow velocity on flow resistance, and the effect of depth [45, 55].
One of the works that has been done on the effect of vegetation on the roughness coefficient is Darby [56] study, which investigates a flood wave model that considers all the effects of vegetation on the roughness coefficient. There are currently two methods for estimating vegetation roughness. One method is to add the thrust force effect to Manning’s equation [47, 57, 58] and the other method is to increase the canal bed roughness (Manning-Strickler coefficient) [45, 59–61]. These two methods provide acceptable results in models designed to simulate floodplain flow. Wang et al. [62] simulate the floodplain with submerged vegetation using these two methods and to increase the accuracy of the results, they suggested using the effective height of the plant under running water instead of using the actual height of the plant. Freeman et al. [49] provided equations for determining the coefficient of vegetation roughness under different conditions. Lee et al. [63] proposed a method for calculating the Manning coefficient using the flow velocity ratio at different depths. Much research has been done on the Manning roughness coefficient in rivers, and researchers [49, 63–66] sought to obtain a specific number for n to use in river engineering. However, since the depth and geometric conditions of rivers are completely variable in different places, the values of Manning roughness coefficient have changed subsequently, and it has not been possible to choose a fixed number. In river engineering software, the Manning roughness coefficient is determined only for specific and constant conditions or normal flow. Lee et al. [63] stated that seasonal conditions, density, and type of vegetation should also be considered. Hydraulic roughness and Manning roughness coefficient n of the plant were obtained by estimating the total Manning roughness coefficient from the matching of the measured water surface curve and water surface height. The following equation is used for the flow surface curve:where is the depth of water change, S0 is the slope of the canal floor, Sf is the slope of the energy line, and Fr is the Froude number which is obtained from the following equation:where D is the characteristic length of the canal. Flood flow velocity is one of the important parameters of flood waves, which is very important in calculating the water level profile and energy consumption. In the cases where there are many limitations for researchers due to the wide range of experimental dimensions and the variety of design parameters, the use of numerical methods that are able to estimate the rest of the unknown results with acceptable accuracy is economically justified.
FLOW-3D software uses Finite Difference Method (FDM) for numerical solution of two-dimensional and three-dimensional flow. This software is dedicated to computational fluid dynamics (CFD) and is provided by Flow Science [67]. The flow is divided into networks with tubular cells. For each cell there are values of dependent variables and all variables are calculated in the center of the cell, except for the velocity, which is calculated at the center of the cell. In this software, two numerical techniques have been used for geometric simulation, FAVOR™ (Fractional-Area-Volume-Obstacle-Representation) and the VOF (Volume-of-Fluid) method. The equations used at this model for this research include the principle of mass survival and the magnitude of motion as follows. The fluid motion equations in three dimensions, including the Navier–Stokes equations with some additional terms, are as follows:where are mass accelerations in the directions x, y, z and are viscosity accelerations in the directions x, y, z and are obtained from the following equations:
Shear stresses in equation (11) are obtained from the following equations:
The standard model is used for high Reynolds currents, but in this model, RNG theory allows the analytical differential formula to be used for the effective viscosity that occurs at low Reynolds numbers. Therefore, the RNG model can be used for low and high Reynolds currents.
Weather changes are high and this affects many factors continuously. The presence of vegetation in any area reduces the velocity of surface flows and prevents soil erosion, so vegetation will have a significant impact on reducing destructive floods. One of the methods of erosion protection in floodplain watersheds is the use of biological methods. The presence of vegetation in watersheds reduces the flow rate during floods and prevents soil erosion. The external organs of plants increase the roughness and decrease the velocity of water flow and thus reduce its shear stress energy. One of the important factors with which the hydraulic resistance of plants is expressed is the roughness coefficient. Measuring the roughness coefficient of plants and investigating their effect on reducing velocity and shear stress of flow is of special importance.
Roughness coefficients in canals are affected by two main factors, namely, flow conditions and vegetation characteristics [68]. So far, much research has been done on the effect of the roughness factor created by vegetation, but the issue of plant density has received less attention. For this purpose, this study was conducted to investigate the effect of vegetation density on flow velocity changes.
In a study conducted using a software model on three density modes in the submerged state effect on flow velocity changes in 48 different modes was investigated (Table 1).
Table 1
The studied models.
The number of cells used in this simulation is equal to 1955888 cells. The boundary conditions were introduced to the model as a constant speed and depth (Figure 1). At the output boundary, due to the presence of supercritical current, no parameter for the current is considered. Absolute roughness for floors and walls was introduced to the model (Figure 1). In this case, the flow was assumed to be nonviscous and air entry into the flow was not considered. After seconds, this model reached a convergence accuracy of .
Figure 1
The simulated model and its boundary conditions.
Due to the fact that it is not possible to model the vegetation in FLOW-3D software, in this research, the vegetation of small soft plants was studied so that Manning’s coefficients can be entered into the canal bed in the form of roughness coefficients obtained from the studies of Chow [69] in similar conditions. In practice, in such modeling, the effect of plant height is eliminated due to the small height of herbaceous plants, and modeling can provide relatively acceptable results in these conditions.
48 models with input velocities proportional to the height of the regular semihexagonal canal were considered to create supercritical conditions. Manning coefficients were applied based on Chow [69] studies in order to control the canal bed. Speed profiles were drawn and discussed.
Any control and simulation system has some inputs that we should determine to test any technology [70–77]. Determination and true implementation of such parameters is one of the key steps of any simulation [23, 78–81] and computing procedure [82–86]. The input current is created by applying the flow rate through the VFR (Volume Flow Rate) option and the output flow is considered Output and for other borders the Symmetry option is considered.
Simulation of the models and checking their action and responses and observing how a process behaves is one of the accepted methods in engineering and science [87, 88]. For verification of FLOW-3D software, the results of computer simulations are compared with laboratory measurements and according to the values of computational error, convergence error, and the time required for convergence, the most appropriate option for real-time simulation is selected (Figures 2 and 3 ).
Figure 2
Modeling the plant with cylindrical tubes at the bottom of the canal.
Figure 3
Velocity profiles in positions 2 and 5.
The canal is 7 meters long, 0.5 meters wide, and 0.8 meters deep. This test was used to validate the application of the software to predict the flow rate parameters. In this experiment, instead of using the plant, cylindrical pipes were used in the bottom of the canal.
The conditions of this modeling are similar to the laboratory conditions and the boundary conditions used in the laboratory were used for numerical modeling. The critical flow enters the simulation model from the upstream boundary, so in the upstream boundary conditions, critical velocity and depth are considered. The flow at the downstream boundary is supercritical, so no parameters are applied to the downstream boundary.
The software well predicts the process of changing the speed profile in the open canal along with the considered obstacles. The error in the calculated speed values can be due to the complexity of the flow and the interaction of the turbulence caused by the roughness of the floor with the turbulence caused by the three-dimensional cycles in the hydraulic jump. As a result, the software is able to predict the speed distribution in open canals.
2. Modeling Results
After analyzing the models, the results were shown in graphs (Figures 4–14 ). The total number of experiments in this study was 48 due to the limitations of modeling.
Flow velocity profiles for canals with a depth of 1 m and flow velocities of 3–3.3 m/s. Canal with a depth of 1 meter and a flow velocity of (a) 3 meters per second, (b) 3.1 meters per second, (c) 3.2 meters per second, and (d) 3.3 meters per second.
Figure 5
Canal diagram with a depth of 1 meter and a flow rate of 3 meters per second.
Figure 6
Canal diagram with a depth of 1 meter and a flow rate of 3.1 meters per second.
Figure 7
Canal diagram with a depth of 1 meter and a flow rate of 3.2 meters per second.
Figure 8
Canal diagram with a depth of 1 meter and a flow rate of 3.3 meters per second.
Flow velocity profiles for canals with a depth of 2 m and flow velocities of 4–4.3 m/s. Canal with a depth of 2 meters and a flow rate of (a) 4 meters per second, (b) 4.1 meters per second, (c) 4.2 meters per second, and (d) 4.3 meters per second.
Figure 10
Canal diagram with a depth of 2 meters and a flow rate of 4 meters per second.
Figure 11
Canal diagram with a depth of 2 meters and a flow rate of 4.1 meters per second.
Figure 12
Canal diagram with a depth of 2 meters and a flow rate of 4.2 meters per second.
Figure 13
Canal diagram with a depth of 2 meters and a flow rate of 4.3 meters per second.
Flow velocity profiles for canals with a depth of 3 m and flow velocities of 5–5.3 m/s. Canal with a depth of 2 meters and a flow rate of (a) 4 meters per second, (b) 4.1 meters per second, (c) 4.2 meters per second, and (d) 4.3 meters per second.
To investigate the effects of roughness with flow velocity, the trend of flow velocity changes at different depths and with supercritical flow to a Froude number proportional to the depth of the section has been obtained.
According to the velocity profiles of Figure 5, it can be seen that, with the increasing of Manning’s coefficient, the canal bed speed decreases.
According to Figures 5 to 8, it can be found that, with increasing the Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of the models 1 to 12, which can be justified by increasing the speed and of course increasing the Froude number.
According to Figure 10, we see that, with increasing Manning’s coefficient, the canal bed speed decreases.
According to Figure 11, we see that, with increasing Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of Figures 5–10, which can be justified by increasing the speed and, of course, increasing the Froude number.
With increasing Manning’s coefficient, the canal bed speed decreases (Figure 12). But this deceleration is more noticeable than the deceleration of the higher models (Figures 5–8 and 10, 11), which can be justified by increasing the speed and, of course, increasing the Froude number.
According to Figure 13, with increasing Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of Figures 5 to 12, which can be justified by increasing the speed and, of course, increasing the Froude number.
According to Figure 15, with increasing Manning’s coefficient, the canal bed speed decreases.
Figure 15
Canal diagram with a depth of 3 meters and a flow rate of 5 meters per second.
According to Figure 16, with increasing Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of the higher model, which can be justified by increasing the speed and, of course, increasing the Froude number.
Figure 16
Canal diagram with a depth of 3 meters and a flow rate of 5.1 meters per second.
According to Figure 17, it is clear that, with increasing Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of the higher models, which can be justified by increasing the speed and, of course, increasing the Froude number.
Figure 17
Canal diagram with a depth of 3 meters and a flow rate of 5.2 meters per second.
According to Figure 18, with increasing Manning’s coefficient, the canal bed speed decreases. But this deceleration is more noticeable than the deceleration of the higher models, which can be justified by increasing the speed and, of course, increasing the Froude number.
Figure 18
Canal diagram with a depth of 3 meters and a flow rate of 5.3 meters per second.
According to Figure 19, it can be seen that the vegetation placed in front of the flow input velocity has negligible effect on the reduction of velocity, which of course can be justified due to the flexibility of the vegetation. The only unusual thing is the unexpected decrease in floor speed of 3 m/s compared to higher speeds.
Comparison of velocity profiles with the same plant densities (depth 1 m). Comparison of velocity profiles with (a) plant densities of 25%, depth 1 m; (b) plant densities of 50%, depth 1 m; and (c) plant densities of 75%, depth 1 m.
According to Figure 20, by increasing the speed of vegetation, the effect of vegetation on reducing the flow rate becomes more noticeable. And the role of input current does not have much effect in reducing speed.
Comparison of velocity profiles with the same plant densities (depth 2 m). Comparison of velocity profiles with (a) plant densities of 25%, depth 2 m; (b) plant densities of 50%, depth 2 m; and (c) plant densities of 75%, depth 2 m.
According to Figure 21, it can be seen that, with increasing speed, the effect of vegetation on reducing the bed flow rate becomes more noticeable and the role of the input current does not have much effect. In general, it can be seen that, by increasing the speed of the input current, the slope of the profiles increases from the bed to the water surface and due to the fact that, in software, the roughness coefficient applies to the channel floor only in the boundary conditions, this can be perfectly justified. Of course, it can be noted that, due to the flexible conditions of the vegetation of the bed, this modeling can show acceptable results for such grasses in the canal floor. In the next directions, we may try application of swarm-based optimization methods for modeling and finding the most effective factors in this research [2, 7, 8, 15, 18, 89–94]. In future, we can also apply the simulation logic and software of this research for other domains such as power engineering [95–99].
Comparison of velocity profiles with the same plant densities (depth 3 m). Comparison of velocity profiles with (a) plant densities of 25%, depth 3 m; (b) plant densities of 50%, depth 3 m; and (c) plant densities of 75%, depth 3 m.
3. Conclusion
The effects of vegetation on the flood canal were investigated by numerical modeling with FLOW-3D software. After analyzing the results, the following conclusions were reached:(i)Increasing the density of vegetation reduces the velocity of the canal floor but has no effect on the velocity of the canal surface.(ii)Increasing the Froude number is directly related to increasing the speed of the canal floor.(iii)In the canal with a depth of one meter, a sudden increase in speed can be observed from the lowest speed and higher speed, which is justified by the sudden increase in Froude number.(iv)As the inlet flow rate increases, the slope of the profiles from the bed to the water surface increases.(v)By reducing the Froude number, the effect of vegetation on reducing the flow bed rate becomes more noticeable. And the input velocity in reducing the velocity of the canal floor does not have much effect.(vi)At a flow rate between 3 and 3.3 meters per second due to the shallow depth of the canal and the higher landing number a more critical area is observed in which the flow bed velocity in this area is between 2.86 and 3.1 m/s.(vii)Due to the critical flow velocity and the slight effect of the roughness of the horseshoe vortex floor, it is not visible and is only partially observed in models 1-2-3 and 21.(viii)As the flow rate increases, the effect of vegetation on the rate of bed reduction decreases.(ix)In conditions where less current intensity is passing, vegetation has a greater effect on reducing current intensity and energy consumption increases.(x)In the case of using the flow rate of 0.8 cubic meters per second, the velocity distribution and flow regime show about 20% more energy consumption than in the case of using the flow rate of 1.3 cubic meters per second.
Nomenclature
n:
Manning’s roughness coefficient
C:
Chézy roughness coefficient
f:
Darcy–Weisbach coefficient
V:
Flow velocity
R:
Hydraulic radius
g:
Gravitational acceleration
y:
Flow depth
Ks:
Bed roughness
A:
Constant coefficient
:
Reynolds number
∂y/∂x:
Depth of water change
S0:
Slope of the canal floor
Sf:
Slope of energy line
Fr:
Froude number
D:
Characteristic length of the canal
G:
Mass acceleration
:
Shear stresses.
Data Availability
All data are included within the paper.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was partially supported by the National Natural Science Foundation of China under Contract no. 71761030 and Natural Science Foundation of Inner Mongolia under Contract no. 2019LH07003.
References
H. Yu, L. Jie, W. Gui et al., “Dynamic Gaussian bare-bones fruit fly optimizers with abandonment mechanism: method and analysis,” Engineering with Computers, vol. 20, pp. 1–29, 2020.View at: Publisher Site | Google Scholar
X. Zhao, D. Li, B. Yang, C. Ma, Y. Zhu, and H. Chen, “Feature selection based on improved ant colony optimization for online detection of foreign fiber in cotton,” Applied Soft Computing, vol. 24, pp. 585–596, 2014.View at: Publisher Site | Google Scholar
J. Hu, H. Chen, A. A. Heidari et al., “Orthogonal learning covariance matrix for defects of grey wolf optimizer: insights, balance, diversity, and feature selection,” Knowledge-Based Systems, vol. 213, Article ID 106684, 2021.View at: Publisher Site | Google Scholar
C. Yu, M. Chen, K. Chen et al., “SGOA: annealing-behaved grasshopper optimizer for global tasks,” Engineering with Computers, vol. 4, pp. 1–28, 2021.View at: Publisher Site | Google Scholar
W. Shan, Z. Qiao, A. A. Heidari, H. Chen, H. Turabieh, and Y. Teng, “Double adaptive weights for stabilization of moth flame optimizer: balance analysis, engineering cases, and medical diagnosis,” Knowledge-Based Systems, vol. 8, Article ID 106728, 2020.View at: Google Scholar
J. Tu, H. Chen, J. Liu et al., “Evolutionary biogeography-based whale optimization methods with communication structure: towards measuring the balance,” Knowledge-Based Systems, vol. 212, Article ID 106642, 2021.View at: Publisher Site | Google Scholar
Y. Zhang, R. Liu, X. Wang et al., “Towards augmented kernel extreme learning models for bankruptcy prediction: algorithmic behavior and comprehensive analysis,” Neurocomputing, vol. 430, 2020.View at: Google Scholar
H.-L. Chen, G. Wang, C. Ma, Z.-N. Cai, W.-B. Liu, and S.-J. Wang, “An efficient hybrid kernel extreme learning machine approach for early diagnosis of Parkinson׳s disease,” Neurocomputing, vol. 184, pp. 131–144, 2016.View at: Publisher Site | Google Scholar
J. Xia, H. Chen, Q. Li et al., “Ultrasound-based differentiation of malignant and benign thyroid Nodules: an extreme learning machine approach,” Computer Methods and Programs in Biomedicine, vol. 147, pp. 37–49, 2017.View at: Publisher Site | Google Scholar
C. Li, L. Hou, B. Y. Sharma et al., “Developing a new intelligent system for the diagnosis of tuberculous pleural effusion,” Computer Methods and Programs in Biomedicine, vol. 153, pp. 211–225, 2018.View at: Publisher Site | Google Scholar
X. Xu and H.-L. Chen, “Adaptive computational chemotaxis based on field in bacterial foraging optimization,” Soft Computing, vol. 18, no. 4, pp. 797–807, 2014.View at: Publisher Site | Google Scholar
M. Wang, H. Chen, B. Yang et al., “Toward an optimal kernel extreme learning machine using a chaotic moth-flame optimization strategy with applications in medical diagnoses,” Neurocomputing, vol. 267, pp. 69–84, 2017.View at: Publisher Site | Google Scholar
L. Chao, K. Zhang, Z. Li, Y. Zhu, J. Wang, and Z. Yu, “Geographically weighted regression based methods for merging satellite and gauge precipitation,” Journal of Hydrology, vol. 558, pp. 275–289, 2018.View at: Publisher Site | Google Scholar
F. J. Golrokh, G. Azeem, and A. Hasan, “Eco-efficiency evaluation in cement industries: DEA malmquist productivity index using optimization models,” ENG Transactions, vol. 1, 2020.View at: Google Scholar
D. Zhao, L. Lei, F. Yu et al., “Chaotic random spare ant colony optimization for multi-threshold image segmentation of 2D Kapur entropy,” Knowledge-Based Systems, vol. 8, Article ID 106510, 2020.View at: Google Scholar
Y. Zhang, R. Liu, X. Wang, H. Chen, and C. Li, “Boosted binary Harris hawks optimizer and feature selection,” Engineering with Computers, vol. 517, pp. 1–30, 2020.View at: Publisher Site | Google Scholar
L. Hu, G. Hong, J. Ma, X. Wang, and H. Chen, “An efficient machine learning approach for diagnosis of paraquat-poisoned patients,” Computers in Biology and Medicine, vol. 59, pp. 116–124, 2015.View at: Publisher Site | Google Scholar
L. Shen, H. Chen, Z. Yu et al., “Evolving support vector machines using fruit fly optimization for medical data classification,” Knowledge-Based Systems, vol. 96, pp. 61–75, 2016.View at: Publisher Site | Google Scholar
X. Zhao, X. Zhang, Z. Cai et al., “Chaos enhanced grey wolf optimization wrapped ELM for diagnosis of paraquat-poisoned patients,” Computational Biology and Chemistry, vol. 78, pp. 481–490, 2019.View at: Publisher Site | Google Scholar
Y. Xu, H. Chen, J. Luo, Q. Zhang, S. Jiao, and X. Zhang, “Enhanced Moth-flame optimizer with mutation strategy for global optimization,” Information Sciences, vol. 492, pp. 181–203, 2019.View at: Publisher Site | Google Scholar
M. Wang and H. Chen, “Chaotic multi-swarm whale optimizer boosted support vector machine for medical diagnosis,” Applied Soft Computing Journal, vol. 88, Article ID 105946, 2020.View at: Publisher Site | Google Scholar
Y. Chen, J. Li, H. Lu, and P. Yan, “Coupling system dynamics analysis and risk aversion programming for optimizing the mixed noise-driven shale gas-water supply chains,” Journal of Cleaner Production, vol. 278, Article ID 123209, 2020.View at: Google Scholar
H. Tang, Y. Xu, A. Lin et al., “Predicting green consumption behaviors of students using efficient firefly grey wolf-assisted K-nearest neighbor classifiers,” IEEE Access, vol. 8, pp. 35546–35562, 2020.View at: Publisher Site | Google Scholar
H.-J. Ma and G.-H. Yang, “Adaptive fault tolerant control of cooperative heterogeneous systems with actuator faults and unreliable interconnections,” IEEE Transactions on Automatic Control, vol. 61, no. 11, pp. 3240–3255, 2015.View at: Google Scholar
H.-J. Ma and L.-X. Xu, “Decentralized adaptive fault-tolerant control for a class of strong interconnected nonlinear systems via graph theory,” IEEE Transactions on Automatic Control, vol. 66, 2020.View at: Google Scholar
H. J. Ma, L. X. Xu, and G. H. Yang, “Multiple environment integral reinforcement learning-based fault-tolerant control for affine nonlinear systems,” IEEE Transactions on Cybernetics, vol. 51, pp. 1–16, 2019.View at: Publisher Site | Google Scholar
J. Hu, M. Wang, C. Zhao, Q. Pan, and C. Du, “Formation control and collision avoidance for multi-UAV systems based on Voronoi partition,” Science China Technological Sciences, vol. 63, no. 1, pp. 65–72, 2020.View at: Publisher Site | Google Scholar
C. Zhang, H. Li, Y. Qian, C. Chen, and X. Zhou, “Locality-constrained discriminative matrix regression for robust face identification,” IEEE Transactions on Neural Networks and Learning Systems, vol. 99, pp. 1–15, 2020.View at: Publisher Site | Google Scholar
X. Zhang, D. Wang, Z. Zhou, and Y. Ma, “Robust low-rank tensor recovery with rectification and alignment,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 43, no. 1, pp. 238–255, 2019.View at: Google Scholar
X. Zhang, J. Wang, T. Wang, R. Jiang, J. Xu, and L. Zhao, “Robust feature learning for adversarial defense via hierarchical feature alignment,” Information Sciences, vol. 560, 2020.View at: Google Scholar
X. Zhang, R. Jiang, T. Wang, and J. Wang, “Recursive neural network for video deblurring,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 03, p. 1, 2020.View at: Publisher Site | Google Scholar
X. Zhang, T. Wang, J. Wang, G. Tang, and L. Zhao, “Pyramid channel-based feature attention network for image dehazing,” Computer Vision and Image Understanding, vol. 197-198, Article ID 103003, 2020.View at: Publisher Site | Google Scholar
X. Zhang, T. Wang, W. Luo, and P. Huang, “Multi-level fusion and attention-guided CNN for image dehazing,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 3, p. 1, 2020.View at: Publisher Site | Google Scholar
L. He, J. Shen, and Y. Zhang, “Ecological vulnerability assessment for ecological conservation and environmental management,” Journal of Environmental Management, vol. 206, pp. 1115–1125, 2018.View at: Publisher Site | Google Scholar
Y. Chen, W. Zheng, W. Li, and Y. Huang, “Large group Activity security risk assessment and risk early warning based on random forest algorithm,” Pattern Recognition Letters, vol. 144, pp. 1–5, 2021.View at: Publisher Site | Google Scholar
J. Hu, H. Zhang, Z. Li, C. Zhao, Z. Xu, and Q. Pan, “Object traversing by monocular UAV in outdoor environment,” Asian Journal of Control, vol. 25, 2020.View at: Google Scholar
P. Tian, H. Lu, W. Feng, Y. Guan, and Y. Xue, “Large decrease in streamflow and sediment load of Qinghai-Tibetan Plateau driven by future climate change: a case study in Lhasa River Basin,” Catena, vol. 187, Article ID 104340, 2020.View at: Publisher Site | Google Scholar
A. Stokes, C. Atger, A. G. Bengough, T. Fourcaud, and R. C. Sidle, “Desirable plant root traits for protecting natural and engineered slopes against landslides,” Plant and Soil, vol. 324, no. 1, pp. 1–30, 2009.View at: Publisher Site | Google Scholar
T. B. Devi, A. Sharma, and B. Kumar, “Studies on emergent flow over vegetative channel bed with downward seepage,” Hydrological Sciences Journal, vol. 62, no. 3, pp. 408–420, 2017.View at: Google Scholar
G. Ireland, M. Volpi, and G. Petropoulos, “Examining the capability of supervised machine learning classifiers in extracting flooded areas from Landsat TM imagery: a case study from a Mediterranean flood,” Remote Sensing, vol. 7, no. 3, pp. 3372–3399, 2015.View at: Publisher Site | Google Scholar
L. Goodarzi and S. Javadi, “Assessment of aquifer vulnerability using the DRASTIC model; a case study of the Dezful-Andimeshk Aquifer,” Computational Research Progress in Applied Science & Engineering, vol. 2, no. 1, pp. 17–22, 2016.View at: Google Scholar
K. Zhang, Q. Wang, L. Chao et al., “Ground observation-based analysis of soil moisture spatiotemporal variability across a humid to semi-humid transitional zone in China,” Journal of Hydrology, vol. 574, pp. 903–914, 2019.View at: Publisher Site | Google Scholar
L. De Doncker, P. Troch, R. Verhoeven, K. Bal, P. Meire, and J. Quintelier, “Determination of the Manning roughness coefficient influenced by vegetation in the river Aa and Biebrza river,” Environmental Fluid Mechanics, vol. 9, no. 5, pp. 549–567, 2009.View at: Publisher Site | Google Scholar
M. Fathi-Moghadam and K. Drikvandi, “Manning roughness coefficient for rivers and flood plains with non-submerged vegetation,” International Journal of Hydraulic Engineering, vol. 1, no. 1, pp. 1–4, 2012.View at: Google Scholar
F.-C. Wu, H. W. Shen, and Y.-J. Chou, “Variation of roughness coefficients for unsubmerged and submerged vegetation,” Journal of Hydraulic Engineering, vol. 125, no. 9, pp. 934–942, 1999.View at: Publisher Site | Google Scholar
M. K. Wood, “Rangeland vegetation-hydrologic interactions,” in Vegetation Science Applications for Rangeland Analysis and Management, vol. 3, pp. 469–491, Springer, 1988.View at: Publisher Site | Google Scholar
C. Wilson, O. Yagci, H.-P. Rauch, and N. Olsen, “3D numerical modelling of a willow vegetated river/floodplain system,” Journal of Hydrology, vol. 327, no. 1-2, pp. 13–21, 2006.View at: Publisher Site | Google Scholar
R. Yazarloo, M. Khamehchian, and M. R. Nikoodel, “Observational-computational 3d engineering geological model and geotechnical characteristics of young sediments of golestan province,” Computational Research Progress in Applied Science & Engineering (CRPASE), vol. 03, 2017.View at: Google Scholar
G. E. Freeman, W. H. Rahmeyer, and R. R. Copeland, “Determination of resistance due to shrubs and woody vegetation,” International Journal of River Basin Management, vol. 19, 2000.View at: Google Scholar
N. Kouwen and T. E. Unny, “Flexible roughness in open channels,” Journal of the Hydraulics Division, vol. 99, no. 5, pp. 713–728, 1973.View at: Publisher Site | Google Scholar
S. Hosseini and J. Abrishami, Open Channel Hydraulics, Elsevier, Amsterdam, Netherlands, 2007.
C. S. James, A. L. Birkhead, A. A. Jordanova, and J. J. O’Sullivan, “Flow resistance of emergent vegetation,” Journal of Hydraulic Research, vol. 42, no. 4, pp. 390–398, 2004.View at: Publisher Site | Google Scholar
F. Huthoff and D. Augustijn, “Channel roughness in 1D steady uniform flow: Manning or Chézy?,,” NCR-days, vol. 102, 2004.View at: Google Scholar
M. S. Sabegh, M. Saneie, M. Habibi, A. A. Abbasi, and M. Ghadimkhani, “Experimental investigation on the effect of river bank tree planting array, on shear velocity,” Journal of Watershed Engineering and Management, vol. 2, no. 4, 2011.View at: Google Scholar
A. Errico, V. Pasquino, M. Maxwald, G. B. Chirico, L. Solari, and F. Preti, “The effect of flexible vegetation on flow in drainage channels: estimation of roughness coefficients at the real scale,” Ecological Engineering, vol. 120, pp. 411–421, 2018.View at: Publisher Site | Google Scholar
S. E. Darby, “Effect of riparian vegetation on flow resistance and flood potential,” Journal of Hydraulic Engineering, vol. 125, no. 5, pp. 443–454, 1999.View at: Publisher Site | Google Scholar
V. Kutija and H. Thi Minh Hong, “A numerical model for assessing the additional resistance to flow introduced by flexible vegetation,” Journal of Hydraulic Research, vol. 34, no. 1, pp. 99–114, 1996.View at: Publisher Site | Google Scholar
T. Fischer-Antze, T. Stoesser, P. Bates, and N. R. B. Olsen, “3D numerical modelling of open-channel flow with submerged vegetation,” Journal of Hydraulic Research, vol. 39, no. 3, pp. 303–310, 2001.View at: Publisher Site | Google Scholar
U. Stephan and D. Gutknecht, “Hydraulic resistance of submerged flexible vegetation,” Journal of Hydrology, vol. 269, no. 1-2, pp. 27–43, 2002.View at: Publisher Site | Google Scholar
F. G. Carollo, V. Ferro, and D. Termini, “Flow resistance law in channels with flexible submerged vegetation,” Journal of Hydraulic Engineering, vol. 131, no. 7, pp. 554–564, 2005.View at: Publisher Site | Google Scholar
W. Fu-sheng, “Flow resistance of flexible vegetation in open channel,” Journal of Hydraulic Engineering, vol. S1, 2007.View at: Google Scholar
P.-f. Wang, C. Wang, and D. Z. Zhu, “Hydraulic resistance of submerged vegetation related to effective height,” Journal of Hydrodynamics, vol. 22, no. 2, pp. 265–273, 2010.View at: Publisher Site | Google Scholar
J. K. Lee, L. C. Roig, H. L. Jenter, and H. M. Visser, “Drag coefficients for modeling flow through emergent vegetation in the Florida Everglades,” Ecological Engineering, vol. 22, no. 4-5, pp. 237–248, 2004.View at: Publisher Site | Google Scholar
G. J. Arcement and V. R. Schneider, Guide for Selecting Manning’s Roughness Coefficients for Natural Channels and Flood Plains, US Government Printing Office, Washington, DC, USA, 1989.
Y. Ding and S. S. Y. Wang, “Identification of Manning’s roughness coefficients in channel network using adjoint analysis,” International Journal of Computational Fluid Dynamics, vol. 19, no. 1, pp. 3–13, 2005.View at: Publisher Site | Google Scholar
E. T. Engman, “Roughness coefficients for routing surface runoff,” Journal of Irrigation and Drainage Engineering, vol. 112, no. 1, pp. 39–53, 1986.View at: Publisher Site | Google Scholar
M. Feizbahr, C. Kok Keong, F. Rostami, and M. Shahrokhi, “Wave energy dissipation using perforated and non perforated piles,” International Journal of Engineering, vol. 31, no. 2, pp. 212–219, 2018.View at: Publisher Site | Google Scholar
M. Farzadkhoo, A. Keshavarzi, H. Hamidifar, and M. Javan, “Sudden pollutant discharge in vegetated compound meandering rivers,” Catena, vol. 182, Article ID 104155, 2019.View at: Publisher Site | Google Scholar
V. T. Chow, Open-channel Hydraulics, Mcgraw-Hill Civil Engineering Series, Chennai, TN, India, 1959.
X. Zhang, R. Jing, Z. Li, Z. Li, X. Chen, and C.-Y. Su, “Adaptive pseudo inverse control for a class of nonlinear asymmetric and saturated nonlinear hysteretic systems,” IEEE/CAA Journal of Automatica Sinica, vol. 8, no. 4, pp. 916–928, 2020.View at: Google Scholar
C. Zuo, Q. Chen, L. Tian, L. Waller, and A. Asundi, “Transport of intensity phase retrieval and computational imaging for partially coherent fields: the phase space perspective,” Optics and Lasers in Engineering, vol. 71, pp. 20–32, 2015.View at: Publisher Site | Google Scholar
C. Zuo, J. Sun, J. Li, J. Zhang, A. Asundi, and Q. Chen, “High-resolution transport-of-intensity quantitative phase microscopy with annular illumination,” Scientific Reports, vol. 7, no. 1, pp. 7654–7722, 2017.View at: Publisher Site | Google Scholar
B.-H. Li, Y. Liu, A.-M. Zhang, W.-H. Wang, and S. Wan, “A survey on blocking technology of entity resolution,” Journal of Computer Science and Technology, vol. 35, no. 4, pp. 769–793, 2020.View at: Publisher Site | Google Scholar
Y. Liu, B. Zhang, Y. Feng et al., “Development of 340-GHz transceiver front end based on GaAs monolithic integration technology for THz active imaging array,” Applied Sciences, vol. 10, no. 21, p. 7924, 2020.View at: Publisher Site | Google Scholar
J. Hu, H. Zhang, L. Liu, X. Zhu, C. Zhao, and Q. Pan, “Convergent multiagent formation control with collision avoidance,” IEEE Transactions on Robotics, vol. 36, no. 6, pp. 1805–1818, 2020.View at: Publisher Site | Google Scholar
M. B. Movahhed, J. Ayoubinejad, F. N. Asl, and M. Feizbahr, “The effect of rain on pedestrians crossing speed,” Computational Research Progress in Applied Science & Engineering (CRPASE), vol. 6, no. 3, 2020.View at: Google Scholar
A. Li, D. Spano, J. Krivochiza et al., “A tutorial on interference exploitation via symbol-level precoding: overview, state-of-the-art and future directions,” IEEE Communications Surveys & Tutorials, vol. 22, no. 2, pp. 796–839, 2020.View at: Publisher Site | Google Scholar
W. Zhu, C. Ma, X. Zhao et al., “Evaluation of sino foreign cooperative education project using orthogonal sine cosine optimized kernel extreme learning machine,” IEEE Access, vol. 8, pp. 61107–61123, 2020.View at: Publisher Site | Google Scholar
G. Liu, W. Jia, M. Wang et al., “Predicting cervical hyperextension injury: a covariance guided sine cosine support vector machine,” IEEE Access, vol. 8, pp. 46895–46908, 2020.View at: Publisher Site | Google Scholar
Y. Wei, H. Lv, M. Chen et al., “Predicting entrepreneurial intention of students: an extreme learning machine with Gaussian barebone harris hawks optimizer,” IEEE Access, vol. 8, pp. 76841–76855, 2020.View at: Publisher Site | Google Scholar
A. Lin, Q. Wu, A. A. Heidari et al., “Predicting intentions of students for master programs using a chaos-induced sine cosine-based fuzzy K-Nearest neighbor classifier,” Ieee Access, vol. 7, pp. 67235–67248, 2019.View at: Publisher Site | Google Scholar
Y. Fan, P. Wang, A. A. Heidari et al., “Rationalized fruit fly optimization with sine cosine algorithm: a comprehensive analysis,” Expert Systems with Applications, vol. 157, Article ID 113486, 2020.View at: Publisher Site | Google Scholar
E. Rodríguez-Esparza, L. A. Zanella-Calzada, D. Oliva et al., “An efficient Harris hawks-inspired image segmentation method,” Expert Systems with Applications, vol. 155, Article ID 113428, 2020.View at: Publisher Site | Google Scholar
S. Jiao, G. Chong, C. Huang et al., “Orthogonally adapted Harris hawks optimization for parameter estimation of photovoltaic models,” Energy, vol. 203, Article ID 117804, 2020.View at: Publisher Site | Google Scholar
Z. Xu, Z. Hu, A. A. Heidari et al., “Orthogonally-designed adapted grasshopper optimization: a comprehensive analysis,” Expert Systems with Applications, vol. 150, Article ID 113282, 2020.View at: Publisher Site | Google Scholar
A. Abbassi, R. Abbassi, A. A. Heidari et al., “Parameters identification of photovoltaic cell models using enhanced exploratory salp chains-based approach,” Energy, vol. 198, Article ID 117333, 2020.View at: Publisher Site | Google Scholar
M. Mahmoodi and K. K. Aminjan, “Numerical simulation of flow through sukhoi 24 air inlet,” Computational Research Progress in Applied Science & Engineering (CRPASE), vol. 03, 2017.View at: Google Scholar
F. J. Golrokh and A. Hasan, “A comparison of machine learning clustering algorithms based on the DEA optimization approach for pharmaceutical companies in developing countries,” ENG Transactions, vol. 1, 2020.View at: Google Scholar
H. Chen, A. A. Heidari, H. Chen, M. Wang, Z. Pan, and A. H. Gandomi, “Multi-population differential evolution-assisted Harris hawks optimization: framework and case studies,” Future Generation Computer Systems, vol. 111, pp. 175–198, 2020.View at: Publisher Site | Google Scholar
J. Guo, H. Zheng, B. Li, and G.-Z. Fu, “Bayesian hierarchical model-based information fusion for degradation analysis considering non-competing relationship,” IEEE Access, vol. 7, pp. 175222–175227, 2019.View at: Publisher Site | Google Scholar
J. Guo, H. Zheng, B. Li, and G.-Z. Fu, “A Bayesian approach for degradation analysis with individual differences,” IEEE Access, vol. 7, pp. 175033–175040, 2019.View at: Publisher Site | Google Scholar
M. M. A. Malakoutian, Y. Malakoutian, P. Mostafapour, and S. Z. D. Abed, “Prediction for monthly rainfall of six meteorological regions and TRNC (case study: north Cyprus),” ENG Transactions, vol. 2, no. 2, 2021.View at: Google Scholar
H. Arslan, M. Ranjbar, and Z. Mutlum, “Maximum sound transmission loss in multi-chamber reactive silencers: are two chambers enough?,,” ENG Transactions, vol. 2, no. 1, 2021.View at: Google Scholar
N. Tonekaboni, M. Feizbahr, N. Tonekaboni, G.-J. Jiang, and H.-X. Chen, “Optimization of solar CCHP systems with collector enhanced by porous media and nanofluid,” Mathematical Problems in Engineering, vol. 2021, Article ID 9984840, 12 pages, 2021.View at: Publisher Site | Google Scholar
Z. Niu, B. Zhang, J. Wang et al., “The research on 220GHz multicarrier high-speed communication system,” China Communications, vol. 17, no. 3, pp. 131–139, 2020.View at: Publisher Site | Google Scholar
B. Zhang, Z. Niu, J. Wang et al., “Four‐hundred gigahertz broadband multi‐branch waveguide coupler,” IET Microwaves, Antennas & Propagation, vol. 14, no. 11, pp. 1175–1179, 2020.View at: Publisher Site | Google Scholar
Z.-Q. Niu, L. Yang, B. Zhang et al., “A mechanical reliability study of 3dB waveguide hybrid couplers in the submillimeter and terahertz band,” Journal of Zhejiang University Science, vol. 1, no. 1, 1998.View at: Google Scholar
B. Zhang, D. Ji, D. Fang, S. Liang, Y. Fan, and X. Chen, “A novel 220-GHz GaN diode on-chip tripler with high driven power,” IEEE Electron Device Letters, vol. 40, no. 5, pp. 780–783, 2019.View at: Publisher Site | Google Scholar
M. Taleghani and A. Taleghani, “Identification and ranking of factors affecting the implementation of knowledge management engineering based on TOPSIS technique,” ENG Transactions, vol. 1, no. 1, 2020.View at: Google Scholar
The hydrodynamics of coral reefs strongly influences their biological functioning, impacting processes such as nutrient availability and uptake, recruitment success and bleaching. For example, coral reefs located in oligotrophic regions depend on upwelling for nutrient supply. Coral reefs at Sodwana Bay, located on the east coast of South Africa, are an example of high latitude marginal reefs. These reefs are subjected to complex hydrodynamic forcings due to the interaction between the strong Agulhas current and the highly variable topography of the region. In this study, we explore the reef scale hydrodynamics resulting from the bathymetry for two steady current scenarios at Two-Mile Reef (TMR) using a combination of field data and numerical simulations. The influence of tides or waves was not considered for this study as well as reef-scale roughness. Tilt current meters with onboard temperature sensors were deployed at selected locations within TMR. We used field observations to identify the dominant flow conditions on the reef for numerical simulations that focused on the hydrodynamics driven by mean currents. During the field campaign, southerly currents were the predominant flow feature with occasional flow reversals to the north. Northerly currents were associated with greater variability towards the southern end of TMR. Numerical simulations showed that Jesser Point was central to the development of flow features for both the northerly and southerly current scenarios. High current variability in the south of TMR during reverse currents is related to the formation of Kelvin-Helmholtz type shear instabilities along the outer edge of an eddy formed north of Jesser Point. Furthermore, downward vertical velocities were computed along the offshore shelf at TMR during southerly currents. Current reversals caused a change in vertical velocities to an upward direction due to the orientation of the bathymetry relative to flow directions.
Highlights
A predominant southerly current was measured at Two-Mile Reef with occasional reversals towards the north.
Field observations indicated that northerly currents are spatially varied along Two-Mile Reef.
Simulation of reverse currents show the formation of a separated flow due to interaction with Jesser Point with Kelvin–Helmholtz type shear instabilities along the seaward edge.
지금까지 Sodwana Bay에서 자세한 암초 규모 유체 역학을 모델링하려는 시도는 없었습니다. 이러한 모델의 결과는 규모가 있는 산호초 사이의 흐름이 산호초 건강에 어떤 영향을 미치는지 탐색하는 데 사용할 수 있습니다. 이 연구에서는 Sodwana Bay의 유체역학을 탐색하는 데 사용할 수 있는 LES 모델을 개발하기 위한 단계별 접근 방식을 구현합니다. 여기서 우리는 이 초기 단계에서 파도와 조수의 영향을 배제하면서 Agulhas 해류의 유체역학에 초점을 맞춥니다. 이 접근법은 흐름의 첫 번째 LES를 제시하고 Sodwana Bay의 산호초에서 혼합함으로써 향후 연구의 기초를 제공합니다.
Booij N, Ris RC, Holthuijsen LH (1999) A third-generation wave model for coastal regions: 1. Model description and validation. J Geophys Res Ocean 104(C4):7649–7666. https://doi.org/10.1029/98JC02622ArticleGoogle Scholar
Celliers L, Schleyer MH (2002) Coral bleaching on high-latitude marginal reefs at Sodwana Bay, South Africa. Mar Pollut Bull 44:1380–1387ArticleGoogle Scholar
Chen SC (2018) Performance assessment of FLOW-3D and XFlow in the numerical modelling of fish-bone type fishway hydraulics https://doi.org/10.15142/T3HH1J
Flow Science Inc (2018) FLOW-3D, Version 12.0 Users Manual. Santa Fe, NM, https://www.flow3d.com/
Flow Science Inc (2019) FLOW-3D, Version 12.0 [Computer Software]. Santa Fe, NM, https://www.flow3d.com/
Franco A, Moernaut J, Schneider-Muntau B, Strasser M, Gems B (2020) The 1958 Lituya Bay tsunami – pre-event bathymetry reconstruction and 3D numerical modelling utilising the computational fluid dynamics software Flow-3D. Nat Hazards Earth Syst Sci 20(8):2255–2279ArticleGoogle Scholar
Fringer OB, Gerritsen M, Street RL (2006) An unstructured-grid, finite-volume, nonhydrostatic, parallel coastal ocean simulator. Ocean Model 14(3):139–173ArticleGoogle Scholar
Hirt CW, Sicilian JM (1985) A porosity technique for the definition of obstacles in rectangular cell meshes. In: Proceedings of 4th International Conference on Ship Hydrodynamics https://ci.nii.ac.jp/naid/10009570543/en/
Hocker LO, Hruska MA (2004) Interleaving synchronous data and asynchronous data in a single data storage file
Lim A, Wheeler AJ, Price DM, O’Reilly L, Harris K, Conti L (2020) Influence of benthic currents on cold-water coral habitats: a combined benthic monitoring and 3D photogrammetric investigation. Sci Rep 10(1):19433. https://doi.org/10.1038/s41598-020-76446-yArticleGoogle Scholar
Morris T (2009) Physical oceanography of Sodwana Bay and its effect on larval transport and coral bleaching. PhD thesis, Cape Peninsula University of Technology
Pope SB (2001) Turbulent flows. Cambridge University Press, CambridgeGoogle Scholar
Porter SN (2009) Biogeography and potential factors regulating shallow subtidal reef communities in the Western Indian Ocean. PhD thesis, University of Cape Town
Porter SN, Schleyer MH (2019) Environmental variation and how its spatial structure influences the cross-shelf distribution of high-latitude coral communities in South Africa. Diversity. https://doi.org/10.3390/d11040057ArticleGoogle Scholar
Ramsay PJ, Mason TR (1990) Development of a type zoning model for Zululand coral reefs, Sodwana Bay, South Africa. J Coastal Res 6(4):829–852Google Scholar
Roberts H, Richardson J, Lagumbay R, Meselhe E, Ma Y (2013) Hydrodynamic and sediment transport modeling using FLOW-3D for siting and optimization of the LCA medium diversion at white ditch hydrodynamic and sediment transport modeling using FLOW-3D for siting and optimization of the LCA medium diversion at white D (December)
Roberts MJ, Ribbink AJ, Morris T, Berg MAVD, Engelbrecht DC, Harding RT (2006) Oceanographic environment of the Sodwana Bay coelacanths (Latimeria chalumnae), South Africa: coelacanth research. South Afr J Sci 102(9):435–443Google Scholar
Rogers JS, Monismith SG, Koweek DA, Torres WI, Dunbar RB (2016) Thermodynamics and hydrodynamics in an atoll reef system and their influence on coral cover. Limnol Oceanogr 61(6):2191–2206. https://doi.org/10.1002/lno.10365ArticleGoogle Scholar
Schleyer MH, Celliers L (2003) Coral dominance at the reef-sediment interface in marginal coral communities at Sodwana Bay, South Africa. Mar Freshw Res 54(8):967–972. https://doi.org/10.1071/MF02049ArticleGoogle Scholar
Schleyer MH, Porter SN (2018) Chapter One – drivers of soft and stony coral community distribution on the high-latitude coral reefs of South Africa. advances in marine biology, vol 80, Academic Press, pp 1–55, https://doi.org/10.1016/bs.amb.2018.09.001
Sebens KP, Grace SP, Helmuth B, Maney EJ Jr, Miles JS (1998) Water flow and prey capture by three scleractinian corals, Madracis mirabilis, Montastrea cavernosa and Porites porites, in a field enclosure. Mar Biol 131(2):347–360ArticleGoogle Scholar
Smagorinsky J (1963) General circulation experiments with the primitive equations. Mon Weather Rev 91(3):99–164ArticleGoogle Scholar
Stocking J, Laforsch C, Sigl R, Reidenbach M (2018) The role of turbulent hydrodynamics and surface morphology on heat and mass transfer in corals. J R Soc Interface 15:20180448. https://doi.org/10.1098/rsif.2018.0448ArticleGoogle Scholar
Wyatt ASJ, Lowe RJ, Humphries S, Waite AM (2010) Particulate nutrient fluxes over a fringing coral reef: relevant scales of phytoplankton production and mechanisms of supply. Mar Ecol Prog Ser 405:113–130ArticleGoogle Scholar
Melquisedec Cortés Zambrano*, Helmer Edgardo Monroy González, Wilson Enrique Amaya Tequia Faculty of Civil Engineering, Santo Tomas Tunja University. Address Av. Universitaria No. 45-202. Tunja – Boyacá – Colombia
Abstract
홍수는 지반이동 및 이동의 원인 중 하나이며, 급속한 도시화 및 도시화로 인해 이전보다 빈번하게 발생할 수 있다. 도시 배수 시스템의 특성은 집수 요소가 결정적인 역할을 하는 범람의 발생 및 범위를 정의할 수 있습니다. 이 문서는 7가지 유형의 화격자 유입구의 수력 유입 효율 및 배출 계수에 대한 수치 조사를 제시합니다. FLOW-3D® 시뮬레이터는 Q = 24, 34.1, 44, 100, 200 및 300 L/s의 유속에서 풀 스케일로 격자를 테스트하는 데 사용되며 종방향 기울기가 1.0인 실험 프로토타입의 구성을 유지합니다. %, 1.5% 및 2.0% 및 고정 횡단 경사, 총 126개 모델. 그 결과를 바탕으로 종류별 및 종단경사 조건에 따른 수력유입구 효율곡선과 토출계수를 구성하였다. 결과는 다른 조사에서 제안된 경험적 공식으로 조정되어 프로토타입의 물리적 테스트 결과를 검증하는 역할을 합니다.
Floods are one of the causes of ground movement and displacement, and due to rapid urbanization and urban growth may occur more frequently than before. The characteristics of an urban drainage system can define the occurrence and extent of flooding, where catchment elements have a determining role. This document presents the numerical investigation of the hydraulic inlet efficiency and the discharge coefficient of seven types of grate inlets. The FLOW-3D® simulator is used to test the gratings at a full scale, under flow rates of Q = 24, 34.1, 44, 100, 200 and 300 L/s, preserving the configuration of the experimental prototype with longitudinal slopes of 1.0%, 1.5% and 2.0% and a fixed cross slope, for a total of 126 models. Based on the results, hydraulic inlet efficiency curves and discharge coefficients are constructed for each type and a longitudinal slope condition. The results are adjusted with empirical formulations proposed in other investigations, serving to verify the results of physical testing of prototypes.
Fig. 1. Physical model of the experimental campaign (source: Chaparro Andrade and Abaunza Tabares, 2021)Fig. 2. Design of the grate inlet types studied: (a) R1, (b) R2, (c) R3, (d) R4, (e) R5, (f) R6, (g) R7 (source: based on geometries of Chaparro Andrade
and Abaunza Tabares, 2021)Fig. 4. Comparison between the results obtained during physical experimentation in prototype 7 and simulation results with FLOW-3D® (source:
made with FlowSight® and photographic record by Chaparro Andrade and Abaunza Tabares, 2021)Fig. 6. Example of the results of flow depth and velocity vectors in the xy plane, for a stable flow condition in a grate inlet type and free surface
configuration and flow regime, of some grating types (source: produced with FlowSight®)
References
Alia Md., S., and Sabtu, N. (2020). Comparison of Different Methodologies for Determining the Efficiency of Gully Inlets. In F. M. Nazri (Ed.), Proceedings of AICCE‘19: Transforming the Nation for a Sustainable Tomorrow (Vol. 53, pp. 1275-1284). Springer Nature Switzerland AG. https://doi.org/10.1007/978-3-030- 32816-0_99 Antunes do Carmo, J. S. (2020). Physical Modelling vs. Numerical Modelling: Complementarity and Learning. July. https://doi. org/10.20944/preprints202007.0753.v1 Aragón-Hernández, J. L. (2013). Modelación numérica integrada de los procesos hidráulicos en el drenaje urbano [Universidad Politécnica de Cataluña]. In Doctoral Tesis. https:// upcommons.upc.edu/handle/2117/95059?locale-attribute=es Argue, J. R., and Pezzaniti, D. (1996). How reliable are inlet (hydraulic) models at representing stormwater flow? Science of the Total Environment, 189-190, 355-359. https://doi.org/10.1016/0048-9697(96)05231-X Banco Mundial, O. (2019). Agua: Panorama general. https:// www.bancomundial.org/es/topic/water/overview Cárdenas-Quintero, M., Carvajal-Serna, L. F., and Marbello-Pérez, R. (2018). Evaluación numérica tridimensional de un sumidero de reja de fondo (Three-Dimensional Numerical Assessment of Grate Inlet). SSRN Electronic Journal, November. https://doi.org/10.2139/ssrn.3112980 Carvalho, R. F., Lopes, P., Leandro, J., and David, L. M. (2019). Numerical Research of Flows into Gullies with Different Outlet Locations. Water, 11(2), 794. https://doi.org/10.3390/ w11040794 Chaparro Andrade, F. G., and Abaunza Tabares, K. V. (2021). Importancia de los sumideros, su funcionamiento y diseño en redes de alcantarillado caso de estudio sector nororiental Tunja. Universidad Santo Tomás. Cortés Zambrano, M., Amaya Tequia, W. E., and Gamba Fernández, D. S. (2020). Implementation of the hydraulic modelling of urban drainage in the northeast sector, Tunja, Boyacá. Revista Facultad de Ingeniería Universidad de Antioquia. https://doi. org/10.17533/udea.redin.20200578 Cosco, C., Gómez, M., Russo, B., Tellez-Alvarez, J., Macchione, F., Costabile, P., and Costanzo, C. (2020). Discharge coefficients for specific grated inlets. Influence of the Froude number. Urban Water Journal, 17(7), 656-668. https://doi.org/10.1080/1573062X.2020.1811881 Despotovic, J., Plavsic, J., Stefanovic, N., and Pavlovic, D. (2005). Inefficiency of storm water inlets as a source of urban floods. Water Science and Technology, 51(2), 139-145. https://doi. org/10.2166/wst.2005.0041 Ellis, J. B., and Marsalek, J. (1996). Overview of urban drainage: Environmental impacts and concerns, means of mitigation and implementation policies. Journal of Hydraulic Research, 34(6), 723-732. https://doi.org/10.1080/00221689609498446 Fang, X., Jiang, S., and Alam, S. R. (2010). Numerical simulations of efficiency of curb-opening inlets. Journal of Hydraulic Engineering, 136(1), 62-66. https://doi.org/10.1061/(ASCE) HY.1943-7900.0000131 Faram, M. G., and Harwood, R. (2000). CFD for the Water Industry; The Role of CFD as a Tool for the Development of Wastewater Treatment Systems. Hydro International, 21-22. Faram, M. G., and Harwood, R. (2002). Assessment of the effectiveness of stormwater treatment chambers using computational fluid dynamics. Global Solutions for Urban Drainage, 40644(September 2002), 1-14. https://doi. org/10.1061/40644(2002)7 Flow Science, I. (2018). FLOW-3D® Version 12.0 Users Manual. In FLOW-3D [Computer software]. https://www.flow3d.com Flow Science, I. (2019). FLOW-3D® Version 12.0 [Computer software] (No. 12). https://www.flow3d.com Ghanbari, R., and Heidarnejad, M. (2020). Experimental and numerical analysis of flow hydraulics in triangular and rectangular piano key weirs. Water Science, 00(00), 1-7. https://doi.org/10. 1080/11104929.2020.1724649
Gómez, M., and Russo, B. (2005a). Comparative study of methodologies to determine inlet efficiency from test data. HEC-12 methodology vs UPC method. Water Resources Management, Algarve, Portugal., 80(October 2014), 623-632. https://doi. org/10.2495/WRM050621 Gómez, M., and Russo, B. (2005b). Comparative study among different methodologies to determine storm sewer inlet efficiency from test data. 10th International Conference on Urban Drainage, August, 21-26. https://www.researchgate.net/publication/255602448_Comparative_study_among_different_methodologies_to_determine_storm_sewer_inlet_efficiency_ from_test_data Gómez, M., Recasens, J., Russo, B., and Martínez-Gomariz, E. (2016). Assessment of inlet efficiency through a 3D simulation: Numerical and experimental comparison. Water Science and Technology, 74(8), 1926-1935. https://doi.org/10.2166/ wst.2016.326 Gómez, M., and Russo, B. (2011). Methodology to estimate hydraulic efficiency of drain inlets. Proceedings of the Institution of Civil Engineers: Water Management, 164(2), 81-90. https://doi. org/10.1680/wama.900070 Gómez Valentin, M. (2007). Hidrología urbana. In Hidrología Urbana (pp. 135-147). Instituto Flumen. Jakeman, A. J., Letcher, R. A., and Norton, J. P. (2006). Ten iterative steps in development and evaluation of environmental models. Environmental Modelling and Software, 21, 602-614. https://doi.org/10.1016/j.envsoft.2006.01.004 Jang, J. H., Hsieh, C. T., and Chang, T. H. (2019). The importance of gully flow modelling to urban flood simulation. Urban Water Journal, 16(5), 377-388. https://doi.org/10.1080/1573062X.2019.1669198 Kaushal, D. R., Thinglas, T., Tomita, Y., Kuchii, S., and Tsukamoto, H. (2012). Experimental investigation on optimization of invert trap configuration for sewer solid management. Powder Technology, 215-216, 1-14. https://doi.org/10.1016/j.powtec.2011.08.029 Khazaee, I., and Mohammadiun, M. (2010). Effects of flow field on open channel flow properties using numerical investigation and experimental comparison. International Journal of Energy and Environment, 1(6), 1083-1096. https://doi.org/10.1016/ S0031-9384(10)00122-8 Kleidorfer, M., Tscheikner-Gratl, F., Vonach, T., and Rauch, W. (2018). What can we learn from a 500-year event? Experiences from urban drainage in Austria. Water Science and Technology, 77(8), 2146-2154. https://doi.org/10.2166/wst.2018.138 Leitão, J. P., Simões, N. E., Pina, R. D., Ochoa-Rodriguez, S., Onof, C., and Sá Marques, A. (2017). Stochastic evaluation of the impact of sewer inlets‘ hydraulic capacity on urban pluvial flooding. Stochastic Environmental Research and Risk Assessment, 31(8), 1907-1922. https://doi.org/10.1007/s00477-016- 1283-x Lopes, P., Leandro, J., Carvalho, R. F., Russo, B., and Gómez, M. (2016). Assessment of the ability of a volume of fluid model to reproduce the efficiency of a continuous transverse gully with grate. Journal of Irrigation and Drainage Engineering, 142(10), 1-9. https://doi.org/10.1061/(ASCE)IR.1943-4774.0001058 Mohsin, M., and Kaushal, D. R. (2016). 3D CFD validation of invert trap efficiency for sewer solid management using VOF model. Water Science and Engineering, 9(2), 106-114. https://doi. org/10.1016/j.wse.2016.06.006 Palla, A., Colli, M., Candela, A., Aronica, G. T., and Lanza, L. G. (2018). Pluvial flooding in urban areas: the role of surface drainage efficiency. Journal of Flood Risk Management, 11, S663-S676. https://doi.org/10.1111/jfr3.12246 Russo, B. (2010). Design of surface drainage systems according to hazard criteria related to flooding of urban areas [Universitat Politècnica de Catalunya]. https://dialnet.unirioja.es/servlet/ tesis?codigo=258828 Sedano-Cruz, K., Carvajal-Escoar, Y., and Ávila Díaz, A. J. (2013). ANÁLISIS DE ASPECTOS QUE INCREMENTAN EL RIESGO DE INUNDACIONES EN COLOMBIA. Luna Azul, 37, 219-218. https://www.redalyc.org/articulo.oa?id=321729206014 Spaliviero, F., May, R. W. P., Escarameia, M. (2000). Spacing of road gullies. Hydraulic performance of BS EN 124 gully gratings. HR Walingford, 44(0). https://doi.org/10.13140/ RG.2.1.1344.0889 Téllez-Álvarez, J., Gómez, M., and Russo, B. (2020). Quantification of energy loss in two grated inlets under pressure. Water (Switzerland), 12(6). https://doi.org/10.3390/w12061601 Téllez Álvarez, J., Gómez, V., Russo, B., and Redondo, J. M. (2003). Performance assessment of numerical modelling for hydraulic efficiency of a grated inlet. 1, 6-8. https://doi.org/10.16309/j.cnki.issn.1007-1776.2003.03.004 Téllez Álvarez, J., Gómez Valentin, M., Paindelli, A., and Russo, B. (2017). ACTIVIDAD EXPERIMENTAL DE I+D+i EN INGENIERÍA HIDRÁULICA EN ESPAÑA. In L. J. Balairón Pérez and D. López Gómez (Eds.), Seminario 2017, Comunicaciones de las líneas prioritarias (pp. 41-43). Universitat Politècnica de València. https://doi.org/10.1017/CBO9781107415324.004 Téllez Álvarez, J., Gómez Valentin, M., and Russo, B. (2019). Modelling of Surcharge Flow Through Grated Inlet. In P. Gourbesville and G. Caignaert (Eds.), Advances in Hydroinformati-
cs. Springer, Singapore. https://doi.org/10.1007/978-981- 4451-42-0 UNDRR, I., and CRED, I. (2018). Pérdidas económicas, pobreza y Desastres 1998 – 2017 (Vol. 6, Issue 1). https://doi.org/10.12962/ j23373520.v6i1.22451 Vyzikas, T., and Greaves, D. (2018). Numerial Modelling. In D. Greaves and G. Iglesias (Eds.), Wave and Tidal Energy (pp. 289-363). John Wiley and Sons Ltd. https://doi. org/10.1002/9781119014492 Yakhot, V., and Orszag, S. A. (1986). Renormalization Group Analysis of Turbulence. I . Basic Theory. Journal of Scientific Computing, 1(1), 3-51. https://doi.org/10.1007/BF01061452 Yakhot, V., and Smith, L. M. (1992). The renormalization group, the ɛ-expansion and derivation of turbulence models. Journal of Scientific Computing, 7(l), 35-61. https://doi.org/10.1007/ BF01060210 Yazdanfar, Z., and Sharma, A. (2015). Urban drainage system planning and design – Challenges with climate change and urbanization: A review. Water Science and Technology, 72(2), 165-https://doi.org/10.2166/wst.2015.207
한국농어촌공사는 수리시설안전진단사업을 통하여 노후 및 기능 저하된 농업생산기반시설물에 대하여 정밀안전진단을 실시하여 사전에 재해, 재난을 대비하고 있습니다.
측수로형 여수로는 계획 홍수량이하의 홍수량이 유입시엔 안정적으로 방류가 일어나나 계획 홍수량 이상의 홍수량이 유입되면 물넘이에서 불완전 월류가 발생하며 방류량이 충분하지 않게 됩니다. 측수로형 여수로는 설계당시의 홍수량에 비해 늘어난 현재에 맞게 변경된 홍수량이 유입할 경우 물넘이에서 불완전월류가 발생하는지를 확인하게 됩니다.
현재 농어촌공사와 농어촌연구원, 수자원공사, 학계 등에서는 전 세계에서 오랜 기간 학계의 연구활동을 통한 수많은 논문 검증과 현장 사용을 통해 검증된 FLOW-3D 수치해석 프로그램을 이용하고 있습니다.
특히, 농어촌공사의 정밀안전진단을 실시할 때 설계홍수량의 저수지 유입 시 물넘이에서 불완전월류가 발생하는지를 확인하고, 불완전월류 발생 시 수위 상승 영향을 분석해 안전성 검토 후 문제가 발견되면 보수·보강 방안을 제시할 수 있는 대표적인 3차원 수치해석 프로그램을 사용하고 있습니다.
농어촌공사 정밀안전진단 업무 수행시 수치해석이 필요하십니까? 수치해석에 대해 궁금하신 사항이나 용역 의뢰가 필요하시면 언제든지 아래 연락처로 연락 주시기 바랍니다.
<담당자 연락처>
전화 : 02-2026-0455
Email : flow3d@stikorea.co.kr
당사에는 20년 이상 수치해석 수처리 분야의 수치해석 연구에 전념하고 있는 전문 연구인력과 다양한 기술적 경험과 전문 수치해석 용역 서비스를 제공하는 숙련된 기술팀이 준비되어 있습니다.
경주 저수지 붕괴 “많은 저수량에 따른 수압 탓”(속보) | 연합뉴스
저수지 정밀안전진단 수치해석 과업 예시
과업의 범위
3차원 수치해석을 통한 OO저수지의 측수로부 수면 검토
측수로 불완전 월류 발생 여부 및 제방 여유고 검토
수치해석 과업 세부내용
가능최대홍수량과 200년, 100년 빈도의 홍수량에 대해 각각의 측수로부 3차원 수치해석
경계조건
가. 수위
만수위
홍수위 – 100년 빈도 – 200년 빈도 – 가능최대홍수량(PMF)
나. 홍수량
100년 빈도의 홍수량
200년 빈도의 홍수량
가능최대홍수량(PMF)
저수지 수위별 방류량 검토 및 제방 여유고 검토
경계조건에 대해 측수로부 물넘이 수면 형상 검토
수위별 방류량을 제공된 수리계산값과 수치해석 결과값을 비교하여 방류 능력 검토
수위에 따른 물넘이 수위를 검토하여 제방 여유고 검토
※ 수위별 수리계산값은 발주처에서 제공
성과물
100년빈도, 200년빈도 및 가능최대홍수량(PMF) 유입에 따른 측수로부 불완전 월류 여부로 인한 제방 여유고 안정성 검토
가능최대홍수량(PMF)을 고려 할 경우 검증된 3차원 수치해석 모델 Data 구축
과업보고서, 보고서 원본 파일 및 PDF 파일, 수치해석 원본 입력 파일 및 결과 파일
기타 ※ 모든 성과물은 CD 및 이동저장장치에 별도 저장하여 납품
보고서 예
OO저수지 측수로 및 주변 지형 형상을 반영하고자 지형 및 구조물 등에 대한 차원 3 CAD , FLOW-3D 3 형상을 작성하였으며 수치모형을 이용하여 구룡저수지 차원 형상에 대 한 측수로 수위별 유동상황을 재현하고자 수치해석을 수행하였다. 금회 수치해석 결과 수위조 , EL.224.82m 건 이상에서는 측수로부 방류능력이 부족하여 불완전 월류가 발생하는 것으로 나타났으며 수리계산 대비 수치해석의 , 오차율은 대략 -16.2% +7 ∼ 2.8% 정도로 나타났다. 수치해석 결과로 산정한 수위별 방류량 관계식은 아래와 같다. y = 3.2583E+00x4 – 2.9427E+03x3 – 9.9663E+05x2 – 1.5001E+08x + 8.4674E+09 여기서 는 저수지수위 는 방류량을 나타낸다 x , y
측수로 수위 방류량 및 불완전월류율
수위 방류량 데이터
수위 방류량 데이터
측수로 물넘이 수위별 해석 결과
측수로 물넘이 수위별 해석 결과
측수로 측벽 수위 분포
측수로 측벽 수위 분포
측수로 측벽 여유고
측수로 측벽은 저수지 수위 에서부터 여유고가 부족하여 월류 발생 EL. 225.32m 예상
FLOW-3D 소프트웨어 제품군의 모든 제품은 2023R1에서 IT 관련 개선 사항을 받았습니다. FLOW-3D 2023R1은 이제 Windows 11 및 RHEL 8을 지원합니다. 누락된 종속성을 보고하도록 Linux 설치 프로그램이 개선되었으며 더 이상 루트 수준 권한이 필요하지 않으므로 설치가 더 쉽고 안전해집니다. 또한 워크플로를 자동화한 사용자를 위해 입력 파일 변환기에 명령줄 인터페이스를 추가하여 스크립트 환경에서도 워크플로가 업데이트된 입력 파일로 작동하는지 확인할 수 있습니다.
확장된 PQ 2 분석
제조에 사용되는 유압 시스템은 PQ 2 곡선을 사용하여 모델링할 수 있습니다. 장치의 세부 사항을 건너뛰고 흐름에 미치는 영향을 포함하기 위해 질량-운동량 소스 또는 속도 경계 조건을 사용하여 유압 시스템을 근사화하는 것이 편리한 단순화인 경우가 많습니다. 기존 PQ 2 분석 모델을 확장하여 이러한 유형의 기하학적 단순화를 허용하면서도 여전히 현실적인 결과를 제공합니다. 이것은 시뮬레이션 시간과 모델 복잡성의 감소로 해석됩니다.
FLOW-3D 2022R2 의 새로운 기능
FLOW-3D 2022R2 제품군 의 출시와 함께 Flow Science는 워크스테이션과 FLOW-3D 의 HPC 버전 을 통합하여 단일 노드 CPU 구성에서 다중 구성에 이르기까지 모든 유형의 하드웨어 아키텍처를 활용할 수 있는 단일 솔버 엔진을 제공합니다. 노드 병렬 고성능 컴퓨팅 실행. 추가 개발에는 점탄성 흐름을 위한 새로운 로그 구조 텐서 방법, 지속적인 솔버 속도 성능 개선, 고급 냉각 채널 및 팬텀 구성 요소 제어, 향상된 연행 공기 기능이 포함됩니다.
통합 솔버
FLOW-3D 제품을 단일 통합 솔버로 마이그레이션하여 로컬 워크스테이션 또는 고성능 컴퓨팅 하드웨어 환경에서 원활하게 실행했습니다.
많은 사용자가 노트북이나 로컬 워크스테이션에서 모델을 실행하지만 고성능 컴퓨팅 클러스터에서 더 큰 모델을 실행합니다. 2022R2 릴리스에서는 통합 솔버를 통해 사용자가 HPC 솔루션에서 OpenMP/MPI 하이브리드 병렬화의 동일한 이점을 활용하여 워크스테이션 및 노트북에서 실행할 수 있습니다.
점점 더 많은 수의 CPU 코어를 사용하는 성능 확장의 예OpenMP/MPI 하이브리드 병렬화를 위한 메시 분해의 예
솔버 성능 개선
멀티 소켓 워크스테이션
멀티 소켓 워크스테이션은 이제 매우 일반적이며 대규모 시뮬레이션을 실행할 수 있습니다. 새로운 통합 솔버를 통해 이러한 유형의 하드웨어를 사용하는 사용자는 일반적으로 HPC 클러스터 구성에서만 사용할 수 있었던 OpenMP/MPI 하이브리드 병렬화를 활용하여 모델을 실행할 수 있는 성능 이점을 볼 수 있습니다.
낮은 수준의 루틴으로 벡터화 및 메모리 액세스 개선
대부분의 테스트 사례에서 10%에서 20% 정도의 성능 향상이 관찰되었으며 일부 사례에서는 20%를 초과하는 런타임 이점이 있었습니다.
정제된 체적 대류 안정성 한계
시간 단계 안정성 한계는 모델 런타임의 주요 동인입니다. 2022R2에서는 새로운 시간 단계 안정성 한계인 3D 대류 안정성 한계를 숫자 위젯에서 사용할 수 있습니다. 실행 중이고 대류가 제한된(cx, cy 또는 cz 제한) 모델의 경우 새 옵션은 30% 정도의 일반적인 속도 향상을 보여주었습니다.
압력 솔버 프리 컨디셔너
경우에 따라 까다로운 흐름 구성의 경우 과도한 압력 솔버 반복으로 인해 실행 시간이 길어질 수 있습니다. 어려운 경우 2022R2에서는 모델이 너무 많이 반복될 때 FLOW-3D가 자동으로 새로운 프리 컨디셔너를 활성화하여 압력 수렴을 돕습니다. 테스트의 런타임이 1.9배에서 335배까지 빨라졌습니다!
점탄성 유체에 대한 로그 형태 텐서 방법
점탄성 유체에 대한 새로운 솔버 옵션을 사용자가 사용할 수 있으며 특히 높은 Weissenberg 수치에 효과적입니다.
로그 구조 텐서 솔루션을 사용하여 점탄성 흐름에 대한 높은 Weissenberg 수에서 개선된 솔루션의 예. Courtesy MF Tome, et al., J. Non-Newton. 체액. 기계 175-176 (2012) 44–54
활성 시뮬레이션 제어 확장
능동 시뮬레이션 제어 기능은 연속 주조 및 적층 제조 응용 프로그램과 주조 및 기타 여러 열 관리 응용 프로그램에 사용되는 냉각 채널에 일반적으로 사용되는 팬텀 개체를 포함하도록 확장되었습니다.
융합 증착 모델링 애플리케이션을 위한 동적 열 제어의 예산업용 탱크 적용을 위한 동적 냉각 채널 제어의 예연속 주조 애플리케이션을 위한 팬텀 물체 속도 제어의 예
연행 공기 기능 개선
디퓨저 및 유사한 산업용 기포 흐름 응용 분야의 경우 이제 대량 공급원을 사용하여 물 기둥에 공기를 도입할 수 있습니다. 또한 혼입 공기 및 용존 산소의 난류 확산에 대한 기본값이 업데이트되었으며 매우 낮은 공기 농도에 대한 모델 정확도가 향상되었습니다.
Fatemehsadat Mirshafiee1, Emad Shahbazi 2, Mohadeseh Safi 3, Rituraj Rituraj 4,* 1Department of Electrical and Computer Engineering, K.N. Toosi University of Technology, Tehran 1999143344 , Iran 2Department of Mechatronic, Amirkabir University of Technology, Tehran 158754413, Iran 3Department of Mechatronic, Electrical and Computer Engineering, University of Tehran, Tehran 1416634793, Iran 4 Faculty of Informatics, Obuda University, 1023, Budapest, Hungary
Correspondence: rituraj88@stud.uni-obuda.hu
ABSTRACT
본 연구는 지속가능한 에너지 변환기의 전력 및 수소 발생 모델링을 위한 데이터 기반 방법론을 제안합니다. 파고와 풍속을 달리하여 파고와 수소생산을 예측합니다.
또한 이 연구는 파도에서 수소를 추출할 수 있는 가능성을 강조하고 장려합니다. FLOW-3D 소프트웨어 시뮬레이션에서 추출한 데이터와 해양 특수 테스트의 실험 데이터를 사용하여 두 가지 데이터 기반 학습 방법의 비교 분석을 수행합니다.
결과는 수소 생산의 양은 생성된 전력의 양에 비례한다는 것을 보여줍니다. 제안된 재생 에너지 변환기의 신뢰성은 지속 가능한 스마트 그리드 애플리케이션으로 추가로 논의됩니다.
This study proposes a data-driven methodology for modeling power and hydrogen generation of a sustainable energy converter. The wave and hydrogen production at different wave heights and wind speeds are predicted. Furthermore, this research emphasizes and encourages the possibility of extracting hydrogen from ocean waves. By using the extracted data from FLOW-3D software simulation and the experimental data from the special test in the ocean, the comparison analysis of two data-driven learning methods is conducted. The results show that the amount of hydrogen production is proportional to the amount of generated electrical power. The reliability of the proposed renewable energy converter is further discussed as a sustainable smart grid application.
Key words
Cavity, Combustion efficiency, hydrogen fuel, Computational Fluent and Gambit.
Figure 1. The process of power and hydrogen production with Searaser.Figure 2. The cross-section A-A of the two essential parts of a SearaserFigure 3. Different parts of a Searaser; 1) Buoy 2) Chamber 3) Valves 4) Generator 5) Anchor systemFigure 4. The boundary conditions of the control volumeFigure 5. The wind velocity during the period of the experimental test
REFERENCES
Kalbasi, R., Jahangiri, M., Dehshiri, S.J.H., Dehshiri, S.S.H., Ebrahimi, S., Etezadi, Z.A.S. and Karimipour, A., 2021. Finding the best station in Belgium to use residential-scale solar heating, one-year dynamic simulation with considering all system losses: economic analysis of using ETSW. Sustainable Energy Technologies and Assessments, 45, p.101097.
Megura M, Gunderson R. Better poison is the cure? Critically examining fossil fuel companies, climate change framing, and corporate sustainability reports. Energy Research & Social Science. 2022 Mar 1;85:102388.
Holechek JL, Geli HM, Sawalhah MN, Valdez R. A global assessment: can renewable energy replace fossil fuels by 2050?. Sustainability. 2022 Jan;14(8):4792.
Ahmad M, Kumar A, Ranjan R. Recent Developments of Tidal Energy as Renewable Energy: An Overview. River and Coastal Engineering. 2022:329-43.
Amini E, Mehdipour H, Faraggiana E, Golbaz D, Mozaffari S, Bracco G, Neshat M. Optimization of hydraulic power take-off system settings for point absorber wave energy converter. Renewable Energy. 2022 Jun 4.
Claywell, R., Nadai, L., Felde, I., Ardabili, S. 2020. Adaptive neuro-fuzzy inference system and a multilayer perceptron model trained with grey wolf optimizer for predicting solar diffuse fraction. Entropy, 22(11), p.1192.
McLeod I, Ringwood JV. Powering data buoys using wave energy: a review of possibilities. Journal of Ocean Engineering and Marine Energy. 2022 Jun 20:1-6.
Olsson G. Water interactions: A systemic view: Why we need to comprehend the water-climate-energy-food-economics-lifestyle connections.
Malkowska A, Malkowski A. Green Energy in the Political Debate. InGreen Energy 2023 (pp. 17-39). Springer, Cham.
Mayon R, Ning D, Ding B, Sergiienko NY. Wave energy converter systems–status and perspectives. InModelling and Optimisation of Wave Energy Converters (pp. 3-58). CRC Press.
Available online at: https://www.offshore-energy.biz/uk-ecotricity-introduces-wave-power-device-searaser/ (9/27/2022)
Mousavi SM, et al.,. Deep learning for wave energy converter modeling using long short-term memory. Mathematics. 2021 Apr 15;9(8):871.
Mega V. The Energy Race to Decarbonisation. InHuman Sustainable Cities 2022 (pp. 105-141). Springer, Cham.
Li R, Tang BJ, Yu B, Liao H, Zhang C, Wei YM. Cost-optimal operation strategy for integrating large scale of renewable energy in China’s power system: From a multi-regional perspective. Applied Energy. 2022 Nov 1;325:119780.
Ardabili S., Abdolalizadeh L., Mako C., Torok B., Systematic Review of Deep Learning and Machine Learning for Building Energy, Frontiers in Energy Research, 10, 2022.
Penalba M, Aizpurua JI, Martinez-Perurena A, Iglesias G. A data-driven long-term metocean data forecasting approach for the design of marine renewable energy systems. Renewable and Sustainable Energy Reviews. 2022 Oct 1;167:112751.
Torabi, M., Hashemi, S., Saybani, M.R., 2019. A Hybrid clustering and classification technique for forecasting short‐term energy consumption. Environmental progress & sustainable energy, 38(1), pp.66-76.
Rivera FP, Zalamea J, Espinoza JL, Gonzalez LG. Sustainable use of spilled turbinable energy in Ecuador: Three different energy storage systems. Renewable and Sustainable Energy Reviews. 2022 Mar 1;156:112005.
Raza SA, Jiang J. Mathematical foundations for balancing single-phase residential microgrids connected to a three-phase distribution system. IEEE Access. 2022 Jan 6;10:5292-303.
Takach M, Sarajlić M, Peters D, Kroener M, Schuldt F, von Maydell K. Review of Hydrogen Production Techniques from Water Using Renewable Energy Sources and Its Storage in Salt Caverns. Energies. 2022 Feb 15;15(4):1415.
Lv Z, Li W, Wei J, Ho F, Cao J, Chen X. Autonomous Chemistry Enabling Environment-Adaptive Electrochemical Energy Storage Devices. CCS Chemistry. 2022 Jul 7:1-9.
Dehghan Manshadi, Mahsa, Milad Mousavi, M. Soltani, Amir Mosavi, and Levente Kovacs. 2022. “Deep Learning for Modeling an Offshore Hybrid Wind–Wave Energy System” Energies 15, no. 24: 9484. https://doi.org/10.3390/en15249484
Ishaq H, Dincer I, Crawford C. A review on hydrogen production and utilization: Challenges and opportunities. International Journal of Hydrogen Energy. 2022 Jul 22;47(62):26238-64.
Maguire JF, Woodcock LV. On the Thermodynamics of Aluminum Cladding Oxidation: Water as the Catalyst for Spontaneous Combustion. Journal of Failure Analysis and Prevention. 2022 Sep 10:1-5.
Mohammadi, M. R., Hadavimoghaddam, F., Pourmahdi, M., Atashrouz, S., Munir, M. T., Hemmati-Sarapardeh, A., … & Mohaddespour, A. (2021). Modeling hydrogen solubility in hydrocarbons using extreme gradient boosting and equations of state. Scientific reports, 11(1).
Ma S, Qin J, Xiu X, Wang S. Design and performance evaluation of an underwater hybrid system of fuel cell and battery. Energy Conversion and Management. 2022 Jun 15;262:115672.
Ahamed R, McKee K, Howard I. A Review of the Linear Generator Type of Wave Energy Converters’ Power Take-Off Systems. Sustainability. 2022 Jan;14(16):9936.
Nejad, H.D., Nazari, M., Nazari, M., Mardan, M.M.S., 2022. Fuzzy State-Dependent Riccati Equation (FSDRE) Control of the Reverse Osmosis Desalination System With Photovoltaic Power Supply. IEEE Access, 10, pp.95585-95603.
Zou S, Zhou X, Khan I, Weaver WW, Rahman S. Optimization of the electricity generation of a wave energy converter using deep reinforcement learning. Ocean Engineering. 2022 Jan 15;244:110363.
Wu J, Qin L, Chen N, Qian C, Zheng S. Investigation on a spring-integrated mechanical power take-off system for wave energy conversion purpose. Energy. 2022 Apr 15;245:123318.
Papini G, Dores Piuma FJ, Faedo N, Ringwood JV, Mattiazzo G. Nonlinear Model Reduction by Moment-Matching for a Point Absorber Wave Energy Conversion System. Journal of Marine Science and Engineering. 2022 May;10(5):656.
Forbush DD, Bacelli G, Spencer SJ, Coe RG, Bosma B, Lomonaco P. Design and testing of a free floating dual flap wave energy converter. Energy. 2022 Feb 1;240:122485.
Rezaei, M.A., 2022. A New Hybrid Cascaded Switched-Capacitor Reduced Switch Multilevel Inverter for Renewable Sources and Domestic Loads. IEEE Access, 10, pp.14157-14183.
Lin Z, Cheng L, Huang G. Electricity consumption prediction based on LSTM with attention mechanism. IEEJ Transactions on Electrical and Electronic Engineering. 2020;15(4):556-562.
Tavoosi, J., Mohammadzadeh, A., Pahlevanzadeh, B., Kasmani, M.B., 2022. A machine learning approach for active/reactive power control of grid-connected doubly-fed induction generators. Ain Shams Engineering Journal, 13(2), p.101564.
Ghalandari, M., 2019. Flutter speed estimation using presented differential quadrature method formulation. Engineering Applications of Computational Fluid Mechanics, 13(1), pp.804-810.
Li Z, Bouscasse B, Ducrozet G, Gentaz L, Le Touzé D, Ferrant P. Spectral wave explicit navier-stokes equations for wavestructure interactions using two-phase computational fluid dynamics solvers. Ocean Engineering. 2021 Feb 1;221:108513.
Zhou Y. Ocean energy applications for coastal communities with artificial intelligencea state-of-the-art review. Energy and AI. 2022 Jul 29:100189.
Miskati S, Farin FM. Performance evaluation of wave-carpet in wave energy extraction at different coastal regions: an analytical approach (Doctoral dissertation, Department of Mechanical and Production Engineering).
Gu C, Li H. Review on Deep Learning Research and Applications in Wind and Wave Energy. Energies. 2022 Feb 17;15(4):1510.
Aazami, R., 2022. Optimal Control of an Energy-Storage System in a Microgrid for Reducing Wind-Power Fluctuations. Sustainability, 14(10), p.6183.
Kabir M, Chowdhury MS, Sultana N, Jamal MS, Techato K. Ocean renewable energy and its prospect for developing economies. InRenewable Energy and Sustainability 2022 Jan 1 (pp. 263-298). Elsevier.
Babajani A, Jafari M, Hafezisefat P, Mirhosseini M, Rezania A, Rosendahl L. Parametric study of a wave energy converter (Searaser) for Caspian Sea. Energy Procedia. 2018 Aug 1;147:334-42.
He J. Coherence and cross-spectral density matrix analysis of random wind and wave in deep water. Ocean Engineering. 2020;197:106930
Ijadi Maghsoodi, A., 2018. Renewable energy technology selection problem using integrated h-swara-multimoora approach. Sustainability, 10(12), p.4481.
Band, S.S., Ardabili, S., Sookhak, M., Theodore, A., Elnaffar, S., Moslehpour, M., Csaba, M., Torok, B., Pai, H.T., 2022. When Smart Cities Get Smarter via Machine Learning: An In-depth Literature Review. IEEE Access.
Shamshirband, S., Rabczuk, T., Nabipour, N. and Chau, K.W., 2020. Prediction of significant wave height; comparison between nested grid numerical model, and machine learning models of artificial neural networks, extreme learning and support vector machines. Engineering Applications of Computational Fluid Mechanics, 14(1), pp.805-817.
Liu, Z., Mohammadzadeh, A., Turabieh, H., Mafarja, M., 2021. A new online learned interval type-3 fuzzy control system for solar energy management systems. IEEE Access, 9, pp.10498-10508.
Bavili, R.E., Mohammadzadeh, A., Tavoosi, J., Mobayen, S., Assawinchaichote, W., Asad, J.H. 2021. A New Active Fault Tolerant Control System: Predictive Online Fault Estimation. IEEE Access, 9, pp.118461-118471.
Akbari, E., Teimouri, A.R., Saki, M., Rezaei, M.A., Hu, J., Band, S.S., Pai, H.T., 2022. A Fault-Tolerant Cascaded SwitchedCapacitor Multilevel Inverter for Domestic Applications in Smart Grids. IEEE Access.
Band, S.S., Ardabili, S., 2022. Feasibility of soft computing techniques for estimating the long-term mean monthly wind speed. Energy Reports, 8, pp.638-648.
Tavoosi, J., Mohammadzadeh, A., Pahlevanzadeh, B., Kasmani, M.B., 2022. A machine learning approach for active/reactive power control of grid-connected doubly-fed induction generators. Ain Shams Engineering Journal, 13(2), p.101564.
Ponnusamy, V. K., Kasinathan, P., Madurai Elavarasan, R., Ramanathan, V., Anandan, R. K., Subramaniam, U., … & Hossain, E. A Comprehensive Review on Sustainable Aspects of Big Data Analytics for the Smart Grid. Sustainability, 2021; 13(23), 13322.
Ahmad, T., Zhang, D., Huang, C., Zhang, H., Dai, N., Song, Y., & Chen, H. Artificial intelligence in sustainable energy industry: Status Quo, challenges and opportunities. Journal of Cleaner Production, 2021; 289, 125834.
Wang, G., Chao, Y., Cao, Y., Jiang, T., Han, W., & Chen, Z. A comprehensive review of research works based on evolutionary game theory for sustainable energy development. Energy Reports, 2022; 8, 114-136.
Iranmehr H., Modeling the Price of Emergency Power Transmission Lines in the Reserve Market Due to the Influence of Renewable Energies, Frontiers in Energy Research, 9, 2022
Farmanbar, M., Parham, K., Arild, Ø., & Rong, C. A widespread review of smart grids towards smart cities. Energies, 2019; 12(23), 4484.
Quartier, N., Crespo, A. J., Domínguez, J. M., Stratigaki, V., & Troch, P. Efficient response of an onshore Oscillating Water Column Wave Energy Converter using a one-phase SPH model coupled with a multiphysics library. Applied Ocean Research, 2021; 115, 102856.
Mahmoodi, K., Nepomuceno, E., & Razminia, A. Wave excitation force forecasting using neural networks. Energy, 2022; 247, 123322.
Wang, H., Alattas, K.A., 2022. Comprehensive review of load forecasting with emphasis on intelligent computing approaches. Energy Reports, 8, pp.13189-13198.
Clemente, D., Rosa-Santos, P., & Taveira-Pinto, F. On the potential synergies and applications of wave energy converters: A review. Renewable and Sustainable Energy Reviews, 2021; 135, 110162.
Felix, A., V. Hernández-Fontes, J., Lithgow, D., Mendoza, E., Posada, G., Ring, M., & Silva, R. Wave energy in tropical regions: deployment challenges, environmental and social perspectives. Journal of Marine Science and Engineering, 2019; 7(7), 219.
Farrok, O., Ahmed, K., Tahlil, A. D., Farah, M. M., Kiran, M. R., & Islam, M. R. Electrical power generation from the oceanic wave for sustainable advancement in renewable energy technologies. Sustainability, 2020; 12(6), 2178.
Guo, B., & Ringwood, J. V. A review of wave energy technology from a research and commercial perspective. IET Renewable Power Generation, 2021; 15(14), 3065-3090.
López-Ruiz, A., Bergillos, R. J., Lira-Loarca, A., & Ortega-Sánchez, M. A methodology for the long-term simulation and uncertainty analysis of the operational lifetime performance of wave energy converter arrays. Energy, 2018; 153, 126-135.
Safarian, S., Saryazdi, S. M. E., Unnthorsson, R., & Richter, C. Artificial neural network integrated with thermodynamic equilibrium modeling of downdraft biomass gasification-power production plant. Energy, 2020; 213, 118800.
Kushwah, S. An oscillating water column (OWC): the wave energy converter. Journal of The Institution of Engineers (India): Series C, 2021; 102(5), 1311-1317.
Pap, J., Mako, C., Illessy, M., Kis, N., 2022. Modeling Organizational Performance with Machine Learning. Journal of Open Innovation: Technology, Market, and Complexity, 8(4), p.177.
Pap, J., Mako, C., Illessy, M., Dedaj, Z., Ardabili, S., Torok, B., 2022. Correlation Analysis of Factors Affecting Firm Performance and Employees Wellbeing: Application of Advanced Machine Learning Analysis. Algorithms, 15(9), p.300.
Alanazi, A., 2022. Determining Optimal Power Flow Solutions Using New Adaptive Gaussian TLBO Method. Applied Sciences, 12(16), p.7959.
Shakibjoo, A.D., Moradzadeh, M., Din, S.U., 2021. Optimized Type-2 Fuzzy Frequency Control for Multi-Area Power Systems. IEEE access, 10, pp.6989-7002.
Zhang, G., 2021. Solar radiation estimation in different climates with meteorological variables using Bayesian model averaging and new soft computing models. Energy Reports, 7, pp.8973-8996.
Cao, Y., Raise, A., Mohammadzadeh, A., Rathinasamy, S., 2021. Deep learned recurrent type-3 fuzzy system: Application for renewable energy modeling/prediction. Energy Reports, 7, pp.8115-8127.
Tavoosi, J., Suratgar, A.A., Menhaj, M.B., 2021. Modeling renewable energy systems by a self-evolving nonlinear consequent part recurrent type-2 fuzzy system for power prediction. Sustainability, 13(6), p.3301.
Bourouis, S., Band, S.S., 2022. Meta-Heuristic Algorithm-Tuned Neural Network for Breast Cancer Diagnosis Using Ultrasound Images. Frontiers in Oncology, 12, p.834028.
Mosavi, A.H., Mohammadzadeh, A., Rathinasamy, S., Zhang, C., Reuter, U., Levente, K. and Adeli, H., 2022. Deep learning fuzzy immersion and invariance control for type-I diabetes. Computers in Biology and Medicine, 149, p.105975.
Almutairi, K., Algarni, S., Alqahtani, T., Moayedi, H., 2022. A TLBO-Tuned Neural Processor for Predicting Heating Load in Residential Buildings. Sustainability, 14(10), p.5924.
Ahmad, Z., Zhong, H., 2020. Machine learning modeling of aerobic biodegradation for azo dyes and hexavalent chromium. Mathematics, 8(6), p.913.
Mosavi, A., Shokri, M., Mansor, Z., Qasem, S.N., Band, S.S. and Mohammadzadeh, A., 2020. Machine learning for modeling the singular multi-pantograph equations. Entropy, 22(9), p.1041.
Ardabili, S., 2019, September. Deep learning and machine learning in hydrological processes climate change and earth systems a systematic review. In International conference on global research and education (pp. 52-62). Springer, Cham.
Moayedi, H., (2021). Suggesting a stochastic fractal search paradigm in combination with artificial neural network for early prediction of cooling load in residential buildings. Energies, 14(6), 1649.
Rezakazemi, M., et al., 2019. ANFIS pattern for molecular membranes separation optimization. Journal of Molecular Liquids, 274, pp.470-476.
Mosavi, A., Faghan, Y., Ghamisi, P., Duan, P., Ardabili, S.F., Salwana, E. and Band, S.S., 2020. Comprehensive review of deep reinforcement learning methods and applications in economics. Mathematics, 8(10), p.1640.
Samadianfard, S., Jarhan, S., Salwana, E., 2019. Support vector regression integrated with fruit fly optimization algorithm for river flow forecasting in Lake Urmia Basin. Water, 11(9), p.1934.
Moayedi, H., (2021). Double-target based neural networks in predicting energy consumption in residential buildings. Energies, 14(5), 1331.
Mohammadzadeh S, D., Kazemi, S.F., 2019. Prediction of compression index of fine-grained soils using a gene expression programming model. Infrastructures, 4(2), p.26.
Karballaeezadeh, N., Mohammadzadeh S, D., Shamshirband, S., Hajikhodaverdikhan, P., 2019. Prediction of remaining service life of pavement using an optimized support vector machine (case study of Semnan–Firuzkuh road). Engineering Applications of Computational Fluid Mechanics, 13(1), pp.188-198.
Rezaei, M. Et al., (2022). Adaptation of A Real-Time Deep Learning Approach with An Analog Fault Detection Technique for Reliability Forecasting of Capacitor Banks Used in Mobile Vehicles. IEEE Access v. 21 pp. 89-99.
Khakian, R., et al., (2020). Modeling nearly zero energy buildings for sustainable development in rural areas. Energies, 13(10), 2593.
인공어초(Artificial Reef, ARs)는 연안 어업 자원을 복원하고 생태 환경을 복원하기 위한 핵심 인공 구조물 중 하나입니다. 그러나 많은 AR이 세굴로 인해 안정성과 기능을 상실한 것으로 밝혀졌다.
AR의 기능적 효과를 보장하기 위해서는 서로 다른 흐름 조건에서 세굴로 인한 매장과 같은 AR의 불안정성을 연구하는 것이 매우 중요합니다.
FLOW-3D에 의해 확립된 3차원 수치 모델은 정상류에서 AR 주변의 국부 세굴 특성을 연구하는 데 사용됩니다. RNG k-ε 난류 모델로 닫힌 RANS 방정식은 하나의 AR 주변의 안정적인 유동장을 시뮬레이션하기 위해 설정됩니다.
시뮬레이션 결과는 이전 실험 결과와 비교되었으며 좋은 일치를 보여줍니다. 그 다음에, 세굴 특성, 평형 세굴 깊이 및 최대 세굴 체적에 대한 AR의 개구수 및 입사각의 영향을 조사하였다. 결과는 개구수가 증가함에 따라 세굴 깊이와 세굴 부피가 감소함을 나타냅니다.
또한 수치적 결과를 바탕으로 AR의 개구수가 평형 세굴깊이와 최대 세굴량에 미치는 영향에 대한 실증식을 제시하였다. 입사각의 변화는 AR의 가장 상류 코너에서 베드 전단 응력의 변화에 영향을 미칠 것입니다. 베드 전단 응력이 클수록 세굴이 더 강해집니다.
본 연구는 증강현실의 최적화된 공학적 설계 및 구축을 위한 이론적 지원과 실질적인 지침을 제공할 것이다. 결과는 개구수가 증가함에 따라 세굴 깊이와 세굴 부피가 감소함을 나타냅니다. 또한 수치적 결과를 바탕으로 AR의 개구수가 평형 세굴깊이와 최대 세굴량에 미치는 영향에 대한 실증식을 제시하였다.
입사각의 변화는 AR의 가장 상류 코너에서 베드 전단 응력의 변화에 영향을 미칠 것입니다. 베드 전단 응력이 클수록 세굴이 더 강해집니다. 본 연구는 증강현실의 최적화된 공학적 설계 및 구축을 위한 이론적 지원과 실질적인 지침을 제공할 것이다.
결과는 개구수가 증가함에 따라 세굴 깊이와 세굴 부피가 감소함을 나타냅니다. 또한 수치적 결과를 바탕으로 AR의 개구수가 평형 세굴깊이와 최대 세굴량에 미치는 영향에 대한 실증식을 제시하였다. 입사각의 변화는 AR의 가장 상류 코너에서 베드 전단 응력의 변화에 영향을 미칠 것입니다.
베드 전단 응력이 클수록 세굴이 더 강해집니다. 본 연구는 증강현실의 최적화된 공학적 설계 및 구축을 위한 이론적 지원과 실질적인 지침을 제공할 것이다. 입사각의 변화는 AR의 가장 상류 코너에서 베드 전단 응력의 변화에 영향을 미칠 것입니다.
베드 전단 응력이 클수록 세굴이 더 강해집니다. 본 연구는 증강현실의 최적화된 공학적 설계 및 구축을 위한 이론적 지원과 실질적인 지침을 제공할 것이다. 입사각의 변화는 AR의 가장 상류 코너에서 베드 전단 응력의 변화에 영향을 미칠 것입니다. 베드 전단 응력이 클수록 세굴이 더 강해집니다.
본 연구는 증강현실의 최적화된 공학적 설계 및 구축을 위한 이론적 지원과 실질적인 지침을 제공할 것이다.
Numerical Simulation of Local Scour Around Square Artificial Reef
Artificial reefs (ARs) are one of the key man-made constructs to restore the offshore fishery resources and recover the ecological environment. However, it is found that many ARs lost their stability and function due to scour. In order to ensure the functional effect of ARs, it is of great significance to study the instability of ARs, like burying caused by scour in different flow conditions. The three-dimensional numerical model established by FLOW-3D is used to study the local scour characteristics around the AR in steady currents. The RANS equations, closed with the RNG k-ε turbulence model, are established for simulating a stable flow field around one AR. The simulation results are compared with previous experimental results and shows good agreement. Then, the effect of the opening number and the incident angles of ARs on the scour characteristics, the equilibrium scour depth and maximum scour volume are investigated. The results indicate that the scour depth and scour volume decrease with the increasing opening number. Moreover, the empirical equations of the effect of the opening number of the AR on the equilibrium scour depth and maximum scour volume are proposed based on the numerical results. The change of the incident angles will affect the change of bed shear stress at the most upstream corner of the AR. The greater bed shear stress results in a more intense scour. This study will provide theoretical support, and practical guidance for the optimized engineering design and construction of ARs.
선박의 동력 요구 사항을 설계할 때 고려해야 할 가장 중요한 요소는 선박 저항 또는 선박에 작용하는 항력입니다. 항력을 극복하는 데 필요한 동력이 추진 시스템의 ‘손실’에 기여하기 때문에 추진 시스템을 설계하는 동안 선박 저항을 추정하는 것이 중요합니다. 선박 저항을 계산하는 세 가지 주요 방법이 있습니다:
Holtrop-Mennen(HM) 방법과 같은 통계적 방법, 수치 분석 또는 CFD(전산 유체 역학) 시뮬레이션 및 모델 테스트, 즉 예인 탱크에서 축소된 모델 테스트. 설계 단계 초기에는 기본 선박 매개변수만 사용할 수 있을 때 HM 방법과 같은 통계 모델만 사용할 수 있습니다.
수치 해석/CFD 시뮬레이션 및 모델 테스트는 선박의 완전한 3D 설계가 완료된 경우에만 수행할 수 있습니다. 본 논문은 Flow-3D 소프트웨어 패키지를 사용하여 CFD 시뮬레이션을 사용하여 잔잔한 수상 선박 저항을 예측하는 것을 목표로 합니다.
롤온/롤오프 승객(RoPax) 페리에 대한 사례 연구를 조사했습니다. 선박 저항은 다양한 선박 속도에서 계산되었습니다. 메쉬는 모든 CFD 시뮬레이션의 결과에 영향을 미치기 때문에 메쉬 민감도를 확인하기 위해 여러 개의 메쉬가 사용되었습니다. 시뮬레이션의 결과를 HM 방법의 추정치와 비교했습니다.
시뮬레이션 결과는 낮은 선박 속도에 대한 HM 방법과 잘 일치했습니다. 더 높은 선속을 위한 HM 방법에 비해 결과의 차이가 상당히 컸다. 선박 저항 분석을 수행하는 Flow-3D의 기능이 시연되었습니다.
While designing the power requirements of a ship, the most important factor to be considered is the ship resistance, or the sea drag forces acting on the ship. It is important to have an estimate of the ship resistance while designing the propulsion system since the power required to overcome the sea drag forces contribute to ‘losses’ in the propulsion system. There are three main methods to calculate ship resistance: Statistical methods like the Holtrop-Mennen (HM) method, numerical analysis or CFD (Computational Fluid Dynamics) simulations, and model testing, i.e. scaled model tests in towing tanks. At the start of the design stage, when only basic ship parameters are available, only statistical models like the HM method can be used. Numerical analysis/ CFD simulations and model tests can be performed only when the complete 3D design of the ship is completed. The present paper aims at predicting the calm water ship resistance using CFD simulations, using the Flow-3D software package. A case study of a roll-on/roll-off passenger (RoPax) ferry was investigated. Ship resistance was calculated at various ship speeds. Since the mesh affects the results in any CFD simulation, multiple meshes were used to check the mesh sensitivity. The results from the simulations were compared with the estimate from the HM method. The results from simulations agreed well with the HM method for low ship speeds. The difference in the results was considerably high compared to the HM method for higher ship speeds. The capability of Flow-3D to perform ship resistance analysis was demonstrated.
Figure 1: Simplified ship geometryFigure 3: Wave pattern at sea surface at 20 knots (10.29 m/s) for mesh 1Figure 4: Ship Resistance (kN) vs Ship Speed (knots)
Publisher
International Society of Multiphysics
Citation
Deshpande SR, Sundsbø P, Das S. Ship resistance analysis using CFD simulations in Flow-3D. The International Journal of Multiphysics. 2020;14(3):227-236
REFERENCES
[1] K. Min and S. Kang, “Study on the form factor and full-scale ship resistance prediction method,” Journal of Marine Science and Technology, vol. 15, pp. 108-118, June 2010. [2] A. Molland, S. Turnock and D. Hudson, “Ship Resistance and Propulsion” Second Edition. In Ship Resistance and Propulsion: Practical Estimation of Ship Propulsive Power (pp. 12-69), August 2017, Cambridge University Press. [3] K. Niklas and H. Pruszko, “Full-scale CFD simulations for the determination of ship resistance as a rational, alternative method to towing tank experiments,” Ocean Engineering, vol. 190, October 2019. [4] A. Elkafas, M. Elgohary and A. Zeid, “Numerical study on the hydrodynamic drag force of a container ship model,” Alexandria Engineering Journal, vol. 58, no. 3, pp. 849-859, September 2019. [5] J. Holtrop and G. Mennen, “An approximate power prediction method,” International Shipbuilding Progress, vol. 29, no. 335, pp. 166-170, July 1982. [6] E. Bøckmann and S. Steen, “Model test and simulation of a ship with wavefoils,” Applied Ocean research, vol. 57, pp. 8-18, April 2016. [7] K. Atreyapurapu, B. Tallapragada and K. Voonna, “Simulation of a Free Surface Flow over a Container Vessel Using CFD,” International Journal of Engineering Trends and Technology (IJETT), vol. 18, no. 7, pp. 334-339, December 2014. [8] J. Petersen, D. Jacobsen and O. Winther, “Statistical modelling for ship propulsion efficiency,” Journal of Marine Science and Technology, vol. 17, pp. 30-39, December 2011. [9] H. Versteeg and W. Malalasekera, An introduction to computational fluid dynamics: the finite volume method (second edition), Harlow, England: Pearson Education Ltd, 2007. [10]C. Hirth and B. Nichols, “Volume of fluid (VOF) method for the dynamics of free boundaries,” Journal of Computational Physics, vol. 39, no. 1, pp. 201-225, January 1981. [11] A. Nordli and H. Khawaja, “Comparison of Explicit Method of Solution for CFD Euler Problems using MATLAB® and FORTRAN 77,” International Journal of Multiphysics, vol. 13, no. 2, 2019. [12] FLOW-3D® Version 12.0 User’s Manual (2018). FLOW-3D [Computer software]. Santa Fe, NM: Flow Science, Inc. https://www.flow3d.com. [13] D. McCluskey and A. Holdø, “Optimizing the hydrocyclone for ballast water treatment using computational fluid dynamics,” International Journal of Multiphysics, vol. 3, no. 3, 2009. [14]M. Breuer, D. Lakehal and W. Rodi, “Flow around a Surface Mounted Cubical Obstacle: Comparison of Les and Rans-Results,” Computation of Three-Dimensional Complex Flows. Notes on Numerical Fluid Mechanics, vol. 49, p. 1996. [15] G. Wei, “A Fixed-Mesh Method for General Moving Objects in Fluid Flow”, Modern Physics Letters B, vol. 19, no. 28, pp. 1719-1722, 2005. [16]J. Michell, “The wave-resistance of a ship,” The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, Vols. 45, 1898, no. 272, pp. 106-123, May 2009.
1 Institut für Wasserwirtschaft, Hydrologie und Konstruktiven Wasserbau, B.O.K.U. University, Muthgasse 18, 1190 Vienna, (actually Senior Process Engineer at the VTU Engineering in Vienna, Zieglergasse 53/1/24, 1070 Vienna, Austria).2 School of Pedagogical and Technological Education, Department of Civil Engineering Educators, ASPETE Campus, Eirini Station, 15122 Amarousio, Athens, Greece.
Abstract
이 논문은 최초의 아르키메데스 나사 터빈 CFD 모델링 결과에 대한 간략한 견해를 제시하며, 이는 “그리스에서 아르키메데스의 부활: 수리 역학 및 아르키메데스 달팽이관 물레방아의 유체역학적 거동 연구에 대한 기여”라는 제목의 최근 연구에서 수행되었습니다. 그리스 자연 및 기술 수로의 수력 잠재력”. Flow-3D 코드를 기반으로 하는 이 CFD 분석은 일반적인 TAST(Tubular Archimedean Screw Turbines)와 관련이 있으며 몇 TWh 정도의 그리스 자연 및 기술 수로의 중요한 미개발 수력 잠재력을 활용하는 연간 및 수천 MW 범위의 총 설치 용량인 소규모 수력 발전 시스템에 대한 몇 가지 유망한 성능을 보여줍니다.
This paper presents a short view of the first Archimedean Screw Turbines CFD modelling results, which were carried out within the recent research entitled “Rebirth of Archimedes in Greece: contribution to the study of hydraulic mechanics and hydrodynamic behavior of Archimedean cochlear waterwheels, for recovering the hydraulic potential of Greek natural and technical watercourses”. This CFD analysis, based to the Flow-3D code, concerns typical Tubular Archimedean Screw Turbines (TASTs) and shows some promising performances for such small hydropower systems harnessing the important unexploited hydraulic potential of natural and technical watercourses of Greece, of the order of several TWh / year and of a total installed capacity in the range of thousands MWs.
Keywords
CFD; Flow-3D; TAST; Small Hydro; Renewable Energy; Greek Watercourses.
Figure 1. Photorealistic view of an inclined axis TAST (photo A. Stergiopoulou).Figure 3. The spectrum of all the screw axis orientation cases.Figure 4. Creation of the 3bladed Archimedean Screw with SolidworksFigure 6. “Meshing & Geometry” tab Operations (Flow 3-D).Figure 7. Comparison of Archimedean screw power performances P(W) for Q = 0.15 m3
/s and 0.30m3
/s
and angles of orientation 22ο & 32ο
.Figure 12. Various performances of the Archimedean Screw (MKE/Mean Kinetic Energy, Torque,
Turbulent Kinetic Energy, Turbulent Dissipation) for flow discharge Q = 0.45 m3
/s and an angle of
orientation θ = 32ο
References
[1] A. Stergiopoulou, Computational and experimental investigation of the hydrodynamic behaviour of screw hydro turbine, Ph.D. Thesis, NTUA, 2017. [2] B. Pelikan, A. Lashofer, Verbesserung der Strömungseigenschaften sowie Planungs-und Betriebsoptimierung von Wasserkraftschnecken, Research Project, BOKU University, Vienna, 2012. [3] G. Müller, J. Senior, Simplified theory of Archimedean screws, Journal of Hydraulic Research 47 (5) (2009) 666-669. [4] C. Rorres, The turn of the screw: Optimal design of an Archimedes screw, Journal of Hydraulic Engineering, 80 (2000) 72-80. [5] A. Stergiopoulou, V. Stergiopoulos, Return of Archimedes: Harnessing with new Archimedean spirals the hydraulic potential of the Greek watercourses, in: Proceedings of the Conference for Climate Change, Thessaloniki, 2009. [6] A. Stergiopoulou, V. Stergiopoulos, from the old Archimedean screw pumps to the new Archimedean screw turbines for hydropower production in Greece, in: Proceedings of CEMEPE Conference, Mykonos, June 21-26, 2009.
[7] V. Stergiopoulos, A. Stergiopoulou, E. Kalkani, Quo Vadis Archimedes Nowadays in Greece? Towards Modern Archimedean Turbines for Recovering Greek Small Hydropower Potential, in: Proceedings of 3rd International Scientific “Energy and Climate Change” Conference, Athens, 2010. [8] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Greece beyond the horizon of the era of transition: Archimedean screw hydropower development terra incognita, International Journal of Energy and Development, v.6, Issue 6, pp. 627-536, 2015. [9] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Experimental and theoretical research of zero head innovative horizontal axis Archimedean screw turbines, Journal of Energy and Development, v.6, Issue 5, pp. 471-478, 2015. [10] A. Stergiopoulou, V. Stergiopoulos, E. Κalkani, Back to the Future: Rediscovering the Archimedean screws as modern turbines for harnessing Greek small hydropower potential, in: Proceedings of the Third International Conference CEMEPE 2011 & SECOTOX, Skiathos, 2011. [11] A. Stergiopoulou, V. Stergiopoulos, Educational Renewable Energy Screw Wheel Technologies for Pico Hydropower Generation, Modern Environmental Science and Engineering, v.4, No.5, pp. 439- 445, May 2018. [12] A. Stergiopoulou, V. Stergiopoulos, Educational Renewable Energy Screw Wheel Technologies for Pico Hydropower Generation, Modern Environmental Science and Engineering, v.4, No.5, pp. 439- 445, May 2018. [13] A. Stergiopoulou, V. Stergiopoulos, Towards an inventory of the archimedean small hydropower potential of Greece, INTERNATIONAL JOURNAL OF ENERGY AND ENVIRONMENT Volume 11, Issue 2, 2020 pp.137-144. [14] Flow Science, FLOW-3D Manual, 2013. [15] K. Versteeg and W. Malalasekera, An Introduction to Computational Fluid Dynamics, Pearson, 2007. [16] C. Hirsch, Numerical Computation of internal and external flows: The fundamentals of Computational Fluid dynamics, John Wiley & Sons, 2007. [17] A. Stergiopoulou, V. Stergiopoulos and E. Kalkani, An eagle’s CFD view of Studying Innovative Archimedean Screw Renewable Hydraulic Energy Systems, Proceedings of the 4th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE) and SECOTOX Conference, Mykonos island, Greece, pp.454-460 June 24-28, 2013. [18] A. Stergiopoulou, V. Stergiopoulos, A., E. Kalkani, Computational Fluid Dynamics Study on a 3D Graphic Solid Model of Archimedean Screw Turbines, Fresenius Environmental Bulletin, vol.23- No1, 2014. [19] Α. Stergiopoulou, Kalkani E., “Towards a First C.F.D. Study of Innovative Archimedean Inclined Axis Hydropower Turbines”, International Journal of Engineering Research & Technology (IJERT), Vol. 2 Issue 9, September – 2013, pp. 193-199. [20] A. Stergiopoulou, V. Stergiopoulos, A first CFD study of small hydro energy recovery from the Attica water supply network, INTERNATIONAL JOURNAL OF ENERGY AND ENVIRONMENT, Volume 11, Issue 3, 2020 pp.157-166.
결합된 Bi-level 메타휴리스틱 접근법을 사용한 해양 재생 에너지 변환기의 설계 최적화
Erfan Amini a1, Mahdieh Nasiri b1, Navid Salami Pargoo a, Zahra Mozhgani c, Danial Golbaz d, Mehrdad Baniesmaeil e, Meysam Majidi Nezhad f, Mehdi Neshat gj, Davide Astiaso Garcia h, Georgios Sylaios i
Abstract
In recent years, there has been an increasing interest in renewable energies in view of the fact that fossil fuels are the leading cause of catastrophic environmental consequences. Ocean wave energy is a renewable energy source that is particularly prevalent in coastal areas. Since many countries have tremendous potential to extract this type of energy, a number of researchers have sought to determine certain effective factors on wave converters’ performance, with a primary emphasis on ambient factors. In this study, we used metaheuristic optimization methods to investigate the effects of geometric factors on the performance of an Oscillating Surge Wave Energy Converter (OSWEC), in addition to the effects of hydrodynamic parameters. To do so, we used CATIA software to model different geometries which were then inserted into a numerical model developed in Flow3D software. A Ribed-surface design of the converter’s flap is also introduced in this study to maximize wave-converter interaction. Besides, a Bi-level Hill Climbing Multi-Verse Optimization (HCMVO) method was also developed for this application. The results showed that the converter performs better with greater wave heights, flap freeboard heights, and shorter wave periods. Additionally, the added ribs led to more wave-converter interaction and better performance, while the distance between the flap and flume bed negatively impacted the performance. Finally, tracking the changes in the five-dimensional objective function revealed the optimum value for each parameter in all scenarios. This is achieved by the newly developed optimization algorithm, which is much faster than other existing cutting-edge metaheuristic approaches.
Keywords
Wave Energy Converter
OSWEC
Hydrodynamic Effects
Geometric Design
Metaheuristic Optimization
Multi-Verse Optimizer
1. Introduction
The increase in energy demand, the limitations of fossil fuels, as well as environmental crises, such as air pollution and global warming, are the leading causes of calling more attention to harvesting renewable energy recently [1], [2], [3]. While still in its infancy, ocean wave energy has neither reached commercial maturity nor technological convergence. In recent decades, remarkable progress has been made in the marine energy domain, which is still in the early stage of development, to improve the technology performance level (TPL) [4], [5]and technology readiness level (TRL) of wave energy converters (WECs). This has been achieved using novel modeling techniques [6], [7], [8], [9], [10], [11], [12], [13], [14] to gain the following advantages [15]: (i) As a source of sustainable energy, it contributes to the mix of energy resources that leads to greater diversity and attractiveness for coastal cities and suppliers. [16] (ii) Since wave energy can be exploited offshore and does not require any land, in-land site selection would be less expensive and undesirable visual effects would be reduced. [17] (iii) When the best layout and location of offshore site are taken into account, permanent generation of energy will be feasible (as opposed to using solar energy, for example, which is time-dependent) [18].
In general, the energy conversion process can be divided into three stages in a WEC device, including primary, secondary, and tertiary stages [19], [20]. In the first stage of energy conversion, which is the subject of this study, the wave power is converted to mechanical power by wave-structure interaction (WSI) between ocean waves and structures. Moreover, the mechanical power is transferred into electricity in the second stage, in which mechanical structures are coupled with power take-off systems (PTO). At this stage, optimal control strategies are useful to tune the system dynamics to maximize power output [10], [13], [12]. Furthermore, the tertiary energy conversion stage revolves around transferring the non-standard AC power into direct current (DC) power for energy storage or standard AC power for grid integration [21], [22]. We discuss only the first stage regardless of the secondary and tertiary stages. While Page 1 of 16 WECs include several categories and technologies such as terminators, point absorbers, and attenuators [15], [23], we focus on oscillating surge wave energy converters (OSWECs) in this paper due to its high capacity for industrialization [24].
Over the past two decades, a number of studies have been conducted to understand how OSWECs’ structures and interactions between ocean waves and flaps affect converters performance. Henry et al.’s experiment on oscillating surge wave energy converters is considered as one of the most influential pieces of research [25], which demonstrated how the performance of oscillating surge wave energy converters (OSWECs) is affected by seven different factors, including wave period, wave power, flap’s relative density, water depth, free-board of the flap, the gap between the tubes, gap underneath the flap, and flap width. These parameters were assessed in their two models in order to estimate the absorbed energy from incoming waves [26], [27]. In addition, Folly et al. investigated the impact of water depth on the OSWECs performance analytically, numerically, and experimentally. According to this and further similar studies, the average annual incident wave power is significantly reduced by water depth. Based on the experimental results, both the surge wave force and the power capture of OSWECs increase in shallow water [28], [29]. Following this, Sarkar et al. found that under such circumstances, the device that is located near the coast performs much better than those in the open ocean [30]. On the other hand, other studies are showing that the size of the converter, including height and width, is relatively independent of the location (within similar depth) [31]. Subsequently, Schmitt et al. studied OSWECs numerically and experimentally. In fact, for the simulation of OSWEC, OpenFOAM was used to test the applicability of Reynolds-averaged Navier-Stokes (RANS) solvers. Then, the experimental model reproduced the numerical results with satisfying accuracy [32]. In another influential study, Wang et al. numerically assessed the effect of OSWEC’s width on their performance. According to their findings, as converter width increases, its efficiency decreases in short wave periods while increases in long wave periods [33]. One of the main challenges in the analysis of the OSWEC is the coupled effect of hydrodynamic and geometric variables. As a result, numerous cutting-edge geometry studies have been performed in recent years in order to find the optimal structure that maximizes power output and minimizes costs. Garcia et al. reviewed hull geometry optimization studies in the literature in [19]. In addition, Guo and Ringwood surveyed geometric optimization methods to improve the hydrodynamic performance of OSWECs at the primary stage [14]. Besides, they classified the hull geometry of OSWECs based on Figure 1. Subsequently, Whittaker et al. proposed a different design of OSWEC called Oyster2. There have been three examples of different geometries of oysters with different water depths. Based on its water depth, they determined the width and height of the converter. They also found that in the constant wave period the less the converter’s width, the less power captures the converter has [34]. Afterward, O’Boyle et al. investigated a type of OSWEC called Oyster 800. They compared the experimental and numerical models with the prototype model. In order to precisely reproduce the shape, mass distribution, and buoyancy properties of the prototype, a 40th-scale experimental model has been designed. Overall, all the models were fairly accurate according to the results [35].
Inclusive analysis of recent research avenues in the area of flap geometry has revealed that the interaction-based designs of such converters are emerging as a novel approach. An initiative workflow is designed in the current study to maximizing the wave energy extrication by such systems. To begin with, a sensitivity analysis plays its role of determining the best hydrodynamic values for installing the converter’s flap. Then, all flap dimensions and characteristics come into play to finalize the primary model. Following, interactive designs is proposed to increase the influence of incident waves on the body by adding ribs on both sides of the flap as a novel design. Finally, a new bi-level metaheuristic method is proposed to consider the effects of simultaneous changes in ribs properties and other design parameters. We hope this novel approach will be utilized to make big-scale projects less costly and justifiable. The efficiency of the method is also compared with four well known metaheuristic algorithms and out weight them for this application.
This paper is organized as follows. First, the research methodology is introduced by providing details about the numerical model implementation. To that end, we first introduced the primary model’s geometry and software details. That primary model is later verified with a benchmark study with regard to the flap angle of rotation and water surface elevation. Then, governing equations and performance criteria are presented. In the third part of the paper, we discuss the model’s sensitivity to lower and upper parts width (we proposed a two cross-sectional design for the flap), bottom elevation, and freeboard. Finally, the novel optimization approach is introduced in the final part and compared with four recent metaheuristic algorithms.
2. Numerical Methods
In this section, after a brief introduction of the numerical software, Flow3D, boundary conditions are defined. Afterwards, the numerical model implementation, along with primary model properties are described. Finally, governing equations, as part of numerical process, are discussed.
2.1. Model Setup
FLOW-3D is a powerful and comprehensive CFD simulation platform for studying fluid dynamics. This software has several modules to solve many complex engineering problems. In addition, modeling complex flows is simple and effective using FLOW-3D’s robust meshing capabilities [36]. Interaction between fluid and moving objects might alter the computational range. Dynamic meshes are used in our modeling to take these changes into account. At each time step, the computational node positions change in order to adapt the meshing area to the moving object. In addition, to choose mesh dimensions, some factors are taken into account such as computational accuracy, computational time, and stability. The final grid size is selected based on the detailed procedure provided in [37]. To that end, we performed grid-independence testing on a CFD model using three different mesh grid sizes of 0.01, 0.015, and 0.02 meters. The problem geometry and boundary conditions were defined the same, and simulations were run on all three grids under the same conditions. The predicted values of the relevant variable, such as velocity, was compared between the grids. The convergence behavior of the numerical solution was analyzed by calculating the relative L2 norm error between two consecutive grids. Based on the results obtained, it was found that the grid size of 0.02 meters showed the least error, indicating that it provided the most accurate and reliable solution among the three grids. Therefore, the grid size of 0.02 meters was selected as the optimal spatial resolution for the mesh grid.
In this work, the flume dimensions are 10 meters long, 0.1 meters wide, and 2.2 meters high, which are shown in figure2. In addition, input waves with linear characteristics have a height of 0.1 meters and a period of 1.4 seconds. Among the linear wave methods included in this software, RNGk-ε and k- ε are appropriate for turbulence model. The research of Lopez et al. shows that RNGk- ε provides the most accurate simulation of turbulence in OSWECs [21]. We use CATIA software to create the flap primary model and other innovative designs for this project. The flap measures 0.1 m x 0.65 m x 0.360 m in x, y and z directions, respectively. In Figure 3, the primary model of flap and its dimensions are shown. In this simulation, five boundaries have been defined, including 1. Inlet, 2. Outlet, 3. Converter flap, 4. Bed flume, and 5. Water surface, which are shown in figure 2. Besides, to avoid wave reflection in inlet and outlet zones, Flow3D is capable of defining some areas as damping zones, the length of which has to be one to one and a half times the wavelength. Therefore, in the model, this length is considered equal to 2 meters. Furthermore, there is no slip in all the boundaries. In other words, at every single time step, the fluid velocity is zero on the bed flume, while it is equal to the flap velocity on the converter flap. According to the wave theory defined in the software, at the inlet boundary, the water velocity is called from the wave speed to be fed into the model.
2.2. Verification
In the current study, we utilize the Schmitt experimental model as a benchmark for verification, which was developed at the Queen’s University of Belfast. The experiments were conducted on the flap of the converter, its rotation, and its interaction with the water surface. Thus, the details of the experiments are presented below based up on the experimental setup’s description [38]. In the experiment, the laboratory flume has a length of 20m and a width of 4.58m. Besides, in order to avoid incident wave reflection, a wave absorption source is devised at the end of the left flume. The flume bed, also, includes two parts with different slops. The flap position and dimensions of the flume can be seen in Figure4. In addition, a wave-maker with 6 paddles is installed at one end. At the opposite end, there is a beach with wire meshes. Additionally, there are 6 indicators to extract the water level elevation. In the flap model, there are three components: the fixed support structure, the hinge, and the flap. The flap measures 0.1m x 0.65m x 0.341m in x, y and z directions, respectively. In Figure5, the details are given [32]. The support structure consists of a 15 mm thick stainless steel base plate measuring 1m by 1.4m, which is screwed onto the bottom of the tank. The hinge is supported by three bearing blocks. There is a foam centerpiece on the front and back of the flap which is sandwiched between two PVC plates. Enabling changes of the flap, three metal fittings link the flap to the hinge. Moreover, in this experiment, the selected wave is generated based on sea wave data at scale 1:40. The wave height and the wave period are equal to 0.038 (m) and 2.0625 (s), respectively, which are tantamount to a wave with a period of 13 (s) and a height of 1.5 (m).
Two distinct graphs illustrate the numerical and experi-mental study results. Figure6 and Figure7 are denoting the angle of rotation of flap and surface elevation in computational and experimental models, respectively. The two figures roughly represent that the numerical and experimental models are a good match. However, for the purpose of verifying the match, we calculated the correlation coefficient (C) and root mean square error (RMSE). According to Figure6, correlation coefficient and RMSE are 0.998 and 0.003, respectively, and in Figure7 correlation coefficient and RMSE are respectively 0.999 and 0.001. Accordingly, there is a good match between the numerical and empirical models. It is worth mentioning that the small differences between the numerical and experimental outputs may be due to the error of the measuring devices and the calibration of the data collection devices.
Including continuity equation and momentum conserva- tion for incompressible fluid are given as [32], [39]:(1)
where P represents the pressure, g denotes gravitational acceleration, u represents fluid velocity, and Di is damping coefficient. Likewise, the model uses the same equation. to calculate the fluid velocity in other directions as well. Considering the turbulence, we use the two-equation model of RNGK- ε. These equations are:
(3)��t(��)+����(����)=����[�eff�������]+��-��and(4)���(��)+����(����)=����[�eff�������]+�1�∗����-��2��2�Where �2� and �1� are constants. In addition, �� and �� represent the turbulent Prandtl number of � and k, respectively.
�� also denote the production of turbulent kinetic energy of k under the effect of velocity gradient, which is calculated as follows:(5)��=�eff[�����+�����]�����(6)�eff=�+��(7)�eff=�+��where � is molecular viscosity,�� represents turbulence viscosity, k denotes kinetic energy, and ∊∊ is energy dissipation rate. The values of constant coefficients in the two-equation RNGK ∊-∊ model is as shown in the Table 1[40].Table 2.
Table 1. Constant coefficients in RNGK-∊ model
Factors
�
�0
�1
�2
��
��
��
Quantity
0.012
4.38
1.42
1.68
1.39
1.39
0.084
Table 2. Flap properties
Joint height (m)
0.476
Height of the center of mass (m)
0.53
Weight (Kg)
10.77
It is worth mentioning that the volume of fluid method is used to separate water and air phases in this software [41]. Below is the equation of this method [40].(8)����+����(���)=0where α and 1 − α are portion of water phase and air phase, respectively. As a weighting factor, each fluid phase portion is used to determine the mixture properties. Finally, using the following equations, we calculate the efficiency of converters [42], [34], [43]:(9)�=14|�|2�+�2+(�+�a)2(�n2-�2)2where �� represents natural frequency, I denotes the inertia of OSWEC, Ia is the added inertia, F is the complex wave force, and B denotes the hydrodynamic damping coefficient. Afterward, the capture factor of the converter is calculated by [44]:(10)��=�1/2��2����gw where �� represents the capture factor, which is the total efficiency of device per unit length of the wave crest at each time step [15], �� represent the dimensional amplitude of the incident wave, w is the flap’s width, and Cg is the group velocity of the incident wave, as below:(11)��=��0·121+2�0ℎsinh2�0ℎwhere �0 denotes the wave number, h is water depth, and H is the height of incident waves.
According to previous sections ∊,����-∊ modeling is used for all models simulated in this section. For this purpose, the empty boundary condition is used for flume walls. In order to preventing wave reflection at the inlet and outlet of the flume, the length of wave absorption is set to be at least one incident wavelength. In addition, the structured mesh is chosen, and the mesh dimensions are selected in two distinct directions. In each model, all grids have a length of 2 (cm) and a height of 1 (cm). Afterwards, as an input of the software for all of the models, we define the time step as 0.001 (s). Moreover, the run time of every simulation is 30 (s). As mentioned before, our primary model is Schmitt model, and the flap properties is given in table2. For all simulations, the flume measures 15 meters in length and 0.65 meters in width, and water depth is equal to 0.335 (m). The flap is also located 7 meters from the flume’s inlet.
Finally, in order to compare the results, the capture factor is calculated for each simulation and compared to the primary model. It is worth mentioning that capture factor refers to the ratio of absorbed wave energy to the input wave energy.
According to primary model simulation and due to the decreasing horizontal velocity with depth, the wave crest has the highest velocity. Considering the fact that the wave’s orbital velocity causes the flap to move, the contact between the upper edge of the flap and the incident wave can enhance its performance. Additionally, the numerical model shows that the dynamic pressure decreases as depth increases, and the hydrostatic pressure increases as depth increases.
To determine the OSWEC design, it is imperative to understand the correlation between the capture factor, wave period, and wave height. Therefore, as it is shown in Figure8, we plot the change in capture factor over the variations in wave period and wave height in 3D and 2D. In this diagram, the first axis features changes in wave period, the second axis displays changes in wave height, and the third axis depicts changes in capture factor. According to our wave properties in the numerical model, the wave period and wave height range from 2 to 14 seconds and 2 to 8 meters, respectively. This is due to the fact that the flap does not oscillate if the wave height is less than 2 (m), and it does not reverse if the wave height is more than 8 (m). In addition, with wave periods more than 14 (s), the wavelength would be so long that it would violate the deep-water conditions, and with wave periods less than 2 (s), the flap would not oscillate properly due to the shortness of wavelength. The results of simulation are shown in Figure 8. As it can be perceived from Figure 8, in a constant wave period, the capture factor is in direct proportion to the wave height. It is because of the fact that waves with more height have more energy to rotate the flap. Besides, in a constant wave height, the capture factor increases when the wave period increases, until a given wave period value. However, the capture factor falls after this point. These results are expected since the flap’s angular displacement is not high in lower wave periods, while the oscillating motion of that is not fast enough to activate the power take-off system in very high wave periods.
As is shown in Figure 9, we plot the change in capture factor over the variations in wave period (s) and water depth (m) in 3D. As it can be seen in this diagram, the first axis features changes in water depth (m), the second axis depicts the wave period (s), and the third axis displays OSWEC’s capture factor. The wave period ranges from 0 to 10 seconds based on our wave properties, which have been adopted from Schmitt’s model, while water depth ranges from 0 to 0.5 meters according to the flume and flap dimensions and laboratory limitations. According to Figure9, for any specific water depth, the capture factor increases in a varying rate when the wave period increases, until a given wave period value. However, the capture factor falls steadily after this point. In fact, the maximum capture factor occurs when the wave period is around 6 seconds. This trend is expected since, in a specific water depth, the flap cannot oscillate properly when the wavelength is too short. As the wave period increases, the flap can oscillate more easily, and consequently its capture factor increases. However, the capture factor drops in higher wave periods because the wavelength is too large to move the flap. Furthermore, in a constant wave period, by changing the water depth, the capture factor does not alter. In other words, the capture factor does not depend on the water depth when it is around its maximum value.
3. Sensitivity Analysis
Based on previous studies, in addition to the flap design, the location of the flap relative to the water surface (freeboard) and its elevation relative to the flume bed (flap bottom elevation) play a significant role in extracting energy from the wave energy converter. This study measures the sensitivity of the model to various parameters related to the flap design including upper part width of the flap, lower part width of the flap, the freeboard, and the flap bottom elevation. Moreover, as a novel idea, we propose that the flap widths differ in the lower and upper parts. In Figure10, as an example, a flap with an upper thickness of 100 (mm) and a lower thickness of 50 (mm) and a flap with an upper thickness of 50 (mm) and a lower thickness of 100 (mm) are shown. The influence of such discrepancy between the widths of the upper and lower parts on the interaction between the wave and the flap, or in other words on the capture factor, is evaluated. To do so, other parameters are remained constant, such as the freeboard, the distance between the flap and the flume bed, and the wave properties.
In Figure11, models are simulated with distinct upper and lower widths. As it is clear in this figure, the first axis depicts the lower part width of the flap, the second axis indicates the upper part width of the flap, and the colors represent the capture factor values. Additionally, in order to consider a sufficient range of change, the flap thickness varies from half to double the value of the primary model for each part.
According to this study, the greater the discrepancy in these two parts, the lower the capture factor. It is on account of the fact that when the lower part of the flap is thicker than the upper part, and this thickness difference in these two parts is extremely conspicuous, the inertia against the motion is significant at zero degrees of rotation. Consequently, it is difficult to move the flap, which results in a low capture factor. Similarly, when the upper part of the flap is thicker than the lower part, and this thickness difference in these two parts is exceedingly noticeable, the inertia is so great that the flap can not reverse at the maximum degree of rotation. As the results indicate, the discrepancy can enhance the performance of the converter if the difference between these two parts is around 20%. As it is depicted in the Figure11, the capture factor reaches its own maximum amount, when the lower part thickness is from 5 to 6 (cm), and the upper part thickness is between 6 and 7 (cm). Consequently, as a result of this discrepancy, less material will be used, and therefore there will be less cost.
As illustrated in Figure12, this study examines the effects of freeboard (level difference between the flap top and water surface) and the flap bottom elevation (the distance between the flume bed and flap bottom) on the converter performance. In this diagram, the first axis demonstrates the freeboard and the second axis on the left side displays the flap bottom elevation, while the colors indicate the capture factor. In addition, the feasible range of freeboard is between -15 to 15 (cm) due to the limitation of the numerical model, so that we can take the wave slamming and the overtopping into consideration. Additionally, based on the Schmitt model and its scaled model of 1:40 of the base height, the flap bottom should be at least 9 (cm) high. Since the effect of surface waves is distributed over the depth of the flume, it is imperative to maintain a reasonable flap height exposed to incoming waves. Thus, the maximum flap bottom elevation is limited to 19 (cm). As the Figure12 pictures, at constant negative values of the freeboard, the capture factor is in inverse proportion with the flap bottom elevation, although slightly.
Furthermore, at constant positive values of the freeboard, the capture factor fluctuates as the flap bottom elevation decreases while it maintains an overall increasing trend. This is on account of the fact that increasing the flap bottom elevation creates turbulence flow behind the flap, which encumbers its rotation, as well as the fact that the flap surface has less interaction with the incoming waves. Furthermore, while keeping the flap bottom elevation constant, the capture factor increases by raising the freeboard. This is due to the fact that there is overtopping with adverse impacts on the converter performance when the freeboard is negative and the flap is under the water surface. Besides, increasing the freeboard makes the wave slam more vigorously, which improves the converter performance.
Adding ribs to the flap surface, as shown in Figure13, is a novel idea that is investigated in the next section. To achieve an optimized design for the proposed geometry of the flap, we determine the optimal number and dimensions of ribs based on the flap properties as our decision variables in the optimization process. As an example, Figure13 illustrates a flap with 3 ribs on each side with specific dimensions.
Figure14 shows the flow velocity field around the flap jointed to the flume bed. During the oscillation of the flap, the pressure on the upper and lower surfaces of the flap changes dynamically due to the changing angle of attack and the resulting change in the direction of fluid flow. As the flap moves upwards, the pressure on the upper surface decreases, and the pressure on the lower surface increases. Conversely, as the flap moves downwards, the pressure on the upper surface increases, and the pressure on the lower surface decreases. This results in a cyclic pressure variation around the flap. Under certain conditions, the pressure field around the flap can exhibit significant variations in magnitude and direction, forming vortices and other flow structures. These flow structures can affect the performance of the OSWEC by altering the lift and drag forces acting on the flap.