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

Performance Comparison of NAMI DANCE and FLOW-3D® Models in Tsunami Propagation, Inundation and Currents using NTHMP Benchmark Problems

NTHMP 벤치마크 문제를 사용하여 쓰나미 전파, 침수 및 해류에서 NAMI DANCE 및 FLOW-3D® 모델의 성능 비교

Pure and Applied Geophysics volume 176, pages3115–3153 (2019)Cite this article

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
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

References

  • Allan, J. C., Komar, P. D., Ruggiero, P., & Witter, R. (2012). The March 2011 Tohoku tsunami and its impacts along the U.S. West Coast. Journal of Coastal Research, 28(5), 1142–1153. https://doi.org/10.2112/jcoastres-d-11-00115.1.Article Google Scholar 
  • 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.Article Google 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.Article Google 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.Article Google 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.Article Google 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.Article Google 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.Article Google Scholar 
  • http://nws.weather.gov/nthmp/documents/nthmpWorkshopProcMerged.pdf.
  • http://nws.weather.gov/nthmp/documents/NTHMP_Currents_Workshop_Report.pdf.
  • 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.Article Google 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., Borrero, J., Son, S., Wilson, R., & Miller, K. (2014). Assessment of the tsunami-induced current hazard. Geophysical Research Letters, 41, 2048–2055. https://doi.org/10.1002/2013GL058680.Article 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.Article Google 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.Article Google 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.Article Google 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.Article Google 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.Article Google Scholar 
  • NOAA Center for Tsunami Research: Tsunami Run-up onto a Complex Three-dimensional Beach; Monai Valley. (n.d). Retrieved from: https://nctr.pmel.noaa.gov/benchmark/Laboratory/Laboratory_MonaiValley/.
  • 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.Article Google 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.Article Google 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.Article Google Scholar 
  • Shuto, N. (1991). Numerical simulation of Tsunamis—its present and near future. Natural Hazards, 4, 171–191.Article Google 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.Article Google Scholar 
  • Tolkova, E. (2014). Land-water boundary treatment for a tsunami model with dimensional splitting. Pure and Applied Geophysics, 171(9), 2289–2314.Article Google 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.Article Google 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.Article Google 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 

Download references

Acknowledgements

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

  1. Deniz Velioglu SogutPresent address: 1212 Computer Science, Department of Civil Engineering, Stony Brook University, Stony Brook, NY, 11794, USA

Authors and Affiliations

  1. Middle East Technical University, 06800, Ankara, TurkeyDeniz Velioglu Sogut & Ahmet Cevdet Yalciner

Corresponding author

Correspondence to Deniz Velioglu Sogut.

Ethics declarations

Conflicts of Interest

The authors declare no conflict of interest.

Reprints and Permissions

About this article

Cite this article

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

Download citation

  • Received22 December 2017
  • Revised16 May 2018
  • Accepted24 May 2018
  • Published07 June 2018
  • Issue Date01 July 2019
  • DOIhttps://doi.org/10.1007/s00024-018-1907-9

Keywords

  • Tsunami
  • depth-averaged shallow water
  • Reynolds-averaged Navier–Stokes
  • benchmarking
  • NAMI DANCE
  • FLOW-3D®
Figure 11. Sketch of scour mechanism around USAF under random waves.

Scour Characteristics and Equilibrium Scour Depth Prediction around Umbrella Suction Anchor Foundation under Random Waves

by Ruigeng Hu 1,Hongjun Liu 2,Hao Leng 1,Peng Yu 3 andXiuhai Wang 1,2,*

1College of Environmental Science and Engineering, Ocean University of China, Qingdao 266000, China

2Key Lab of Marine Environment and Ecology (Ocean University of China), Ministry of Education, Qingdao 266000, China

3Qingdao Geo-Engineering Survering Institute, Qingdao 266100, China

*Author to whom correspondence should be addressed.

J. Mar. Sci. Eng. 20219(8), 886; https://doi.org/10.3390/jmse9080886

Received: 6 July 2021 / Revised: 8 August 2021 / Accepted: 13 August 2021 / Published: 17 August 2021

(This article belongs to the Section Ocean Engineering)

Download 

Abstract

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.

Keywords: 

scournumerical investigationrandom wavesequilibrium scour depthKC number

1. Introduction

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].

Jmse 09 00886 g001 550

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 θKC 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.

S/D=1.3(1−exp([−m(KC−6)])�/�=1.3(1−exp(−�(��−6))(2)

where, m = 0.03 for linear waves.

S/D=1.3(1−exp([−0.02(KC−4)])�/�=1.3(1−exp(−0.02(��−4))(3)

S/D=1.3γKwaveKhw�/�=1.3��wave�ℎw(4)

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.

S/D=1.5[tanh(hwD)]KwaveKhw�/�=1.5tanh(ℎw�)�wave�ℎw(5)

where, hw is water depth.

S/D=0.0753(θθcr−−−√−0.5)0.69KC0.68�/�=0.0753(��cr−0.5)0.69��0.68(6)

where, θ is shields parameter, θcr is critical shields parameter.

S/D=2.5(1−0.5u/uc)xrelxrel=xeff/(1+xeff)xeff=0.03(1−0.35ucr/u)(KC−6)⎫⎭⎬⎪⎪�/�=2.5(1−0.5�/��)��������=����/(1+����)����=0.03(1−0.35�cr/�)(��−6)(7)

where, u is near-bed orbital velocity amplitude, uc is critical velocity corresponding the onset of sediment motion.

S/D=1.3{1−exp[−0.03(KC2lnn+36)1/2−6]}�/�=1.31−exp−0.03(��2ln�+36)1/2−6(8)

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:

∂u∂t+1VF(uAx∂u∂x+vAy∂u∂y+wAz∂u∂z)=−1ρf∂p∂x+Gx+fx∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(9)

∂v∂t+1VF(uAx∂v∂x+vAy∂v∂y+wAz∂v∂z)=−1ρf∂p∂y+Gy+fy∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(10)

∂w∂t+1VF(uAx∂w∂x+vAy∂w∂y+wAz∂w∂z)=−1ρf∂p∂z+Gz+fz∂�∂�+1��(���∂�∂�+���∂�∂�+���∂�∂�)=−1�f∂�∂�+��+��(11)

where, VF is the volume fraction; uv, and w are the velocity components in xyz 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 = xyz).

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].

∂kT∂T+1VF(uAx∂kT∂x+vAy∂kT∂y+wAz∂kT∂z)=PT+GT+DiffkT−εkT∂��∂�+1��(���∂��∂�+���∂��∂�+���∂��∂�)=��+��+������−���(12)

∂εT∂T+1VF(uAx∂εT∂x+vAy∂εT∂y+wAz∂εT∂z)=CDIS1εTkT(PT+CDIS3GT)+Diffε−CDIS2ε2TkT∂��∂�+1��(���∂��∂�+���∂��∂�+���∂��∂�)=����1����(��+����3��)+�����−����2��2��(13)

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 VFAiCDIS1CDIS2 and CDIS3 are dimensionless parameters, and CDIS1CDIS3 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]:

ulift,i=αinsd0.3∗(θ−θcr)1.5∥g∥di(ρi−ρf)ρf−−−−−−−−−−−−√�lift,i=�����*0.3(�−�cr)1.5���(��−�f)�f(14)

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.

d∗=di(∥g∥ρf(ρi−ρf)μ2f)1/3�*=��(��f(��−�f)�f2)1/3(15)

where μf is the fluid dynamic viscosity.

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]

θcr=0.31+1.2d∗+0.055[1−exp(−0.02d∗)]�cr=0.31+1.2�*+0.0551−exp(−0.02�*)(17)

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]:

usettling,i=νfdi[(10.362+1.049d3∗)0.5−10.36]�settling,�=�f��(10.362+1.049�*3)0.5−10.36(18)

where νf is the fluid kinematic viscosity.

2.3.2. Bed Load Transport

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]:

ubedload,i=qb,iδicb,ifb�bedload,�=�b,����b,��b(19)

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.

qb,i=8[∥g∥(ρi−ρfρf)d3i]1/2�b,�=8�(��−�f�f)��31/2(20)

δi=0.3d0.7∗(θθcr−1)0.5di��=0.3�*0.7(��cr−1)0.5��(21)

2.3.3. Suspended Load Transport

Through the following transport equation, the suspended sediment concentration could be acquired.

∂Cs,i∂t+∇(us,iCs,i)=∇∇(DfCs,i)∂�s,�∂�+∇(�s,��s,�)=∇∇(�f�s,�)(22)

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 iDf is the diffusivity.

The velocity of sand i in the multiple species could be obtained from the following equation:

us,i=u¯¯+usettling,ics,i�s,�=�¯+�settling,��s,�(23)

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.

Jmse 09 00886 g002 550

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.

Table

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.

Jmse 09 00886 g003 550

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]:

S(ω)=αg2ω5exp[−54(ωpω)4]γexp[−(ω−ωp)22σ2ω2p]�(�)=��2�5exp−54(�p�)4�exp−(�−�p)22�2�p2(25)

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.

Table

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.

Jmse 09 00886 g004 550

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.

Jmse 09 00886 g005 550

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.

Jmse 09 00886 g006 550

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.

Jmse 09 00886 g007 550

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.

Jmse 09 00886 g008 550

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.

Jmse 09 00886 g009 550

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.

Jmse 09 00886 g010a 550
Jmse 09 00886 g010b 550

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.

Jmse 09 00886 g011 550

Figure 11. Sketch of scour mechanism around USAF under random waves.

Jmse 09 00886 g012 550

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.

Jmse 09 00886 g013 550

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

Table 4. Uwm,s and KC for case 1~9.

Table

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.

Jmse 09 00886 g014 550

Figure 14. Sketch of the position where the Seq was evaluated.

Jmse 09 00886 g015a 550
Jmse 09 00886 g015b 550

Figure 15. Comparison of the equilibrium scour depth between the present model and the model of Raaijmakers and Rudolph [34]: (aKCrms,sKCrms,a; (bKCrms,pKCrms,m; (cKCrms,n = 1/10KCrms,n = 1/5; (dKCs,sKCs,a; (eKCs,pKCs,m; (fKCs,n = 1/10KCs,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

S′eq/D=1.95[tanh(hD)](1−exp(−0.012KCs,p))+ΔS/D�eq′/�=1.95tanh(ℎ�)(1−exp(−0.012��s,p))+∆�/�(31)

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.

Jmse 09 00886 g016 550

Figure 16. Comparison of Seq between the simulating results and the predicting values by Equation (31).

Jmse 09 00886 g017 550

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.

Jmse 09 00886 g018 550

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).

ΔSeq/D=0.052*exp(−n/6.566)+0.068∆�eq/�=0.052*exp(−�/6.566)+0.068(32)

S′eq¯/D=S′eq/D+0.052*exp(−n/6.566)+0.068�eq′¯/�=�eq′/�+0.052*exp(−�/6.566)+0.068(33)

Jmse 09 00886 g019 550

Figure 19. The fitting line between ∆S′and n.

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.

Jmse 09 00886 g020 550

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.

Jmse 09 00886 g021 550

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.

Uw=1T/4∫0T/4Uwmsin(t/T)dt=2πUwm�w=1�/4∫0�/4�wmsin(�/�)��=2��wm(35)

Jmse 09 00886 g022 550

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 AB and C are constant.

Jmse 09 00886 g023 550

Figure 23. The fitting curve between Seq/D and Fr.

Jmse 09 00886 g024 550

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.

Jmse 09 00886 g025 550

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

  1. Sumer, B.M.; Fredsøe, J.; Christiansen, N. Scour Around Vertical Pile in Waves. J. Waterw. Port. Coast. Ocean Eng. 1992118, 15–31. [Google Scholar] [CrossRef]
  2. 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]
  3. 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. 201372, 20–38. [Google Scholar] [CrossRef]
  4. 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. 2018138, 132–151. [Google Scholar] [CrossRef]
  5. 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. 2018140, 042001. [Google Scholar] [CrossRef]
  6. Li, H.; Liu, H.; Liu, S. Dynamic analysis of umbrella suction anchor foundation embedded in seabed for offshore wind turbines. Géoméch. Energy Environ. 201710, 12–20. [Google Scholar] [CrossRef]
  7. 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. 2019172, 118–123. [Google Scholar] [CrossRef]
  8. 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. Energies 201912, 1709. [Google Scholar] [CrossRef][Green Version]
  9. 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. 20208, 417. [Google Scholar] [CrossRef]
  10. Sørensen, S.P.H.; Ibsen, L.B. Assessment of foundation design for offshore monopiles unprotected against scour. Ocean Eng. 201363, 17–25. [Google Scholar] [CrossRef]
  11. Prendergast, L.; Gavin, K.; Doherty, P. An investigation into the effect of scour on the natural frequency of an offshore wind turbine. Ocean Eng. 2015101, 1–11. [Google Scholar] [CrossRef][Green Version]
  12. 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. 20219, 297. [Google Scholar] [CrossRef]
  13. 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. 2020202, 106701. [Google Scholar] [CrossRef]
  14. 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. 2020213, 107696. [Google Scholar] [CrossRef]
  15. 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. 1997332, 41–70. [Google Scholar] [CrossRef]
  16. Sumer, B.M.; Fredsøe, J. Scour around Pile in Combined Waves and Current. J. Hydraul. Eng. 2001127, 403–411. [Google Scholar] [CrossRef]
  17. 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]
  18. 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. 2015106, 42–72. [Google Scholar] [CrossRef]
  19. Qi, W.; Gao, F. Equilibrium scour depth at offshore monopile foundation in combined waves and current. Sci. China Ser. E Technol. Sci. 201457, 1030–1039. [Google Scholar] [CrossRef][Green Version]
  20. Larsen, B.E.; Fuhrman, D.R.; Baykal, C.; Sumer, B.M. Tsunami-induced scour around monopile foundations. Coast. Eng. 2017129, 36–49. [Google Scholar] [CrossRef][Green Version]
  21. 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. 2018144, 04018018. [Google Scholar] [CrossRef]
  22. 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. 2020161, 103751. [Google Scholar] [CrossRef]
  23. Fazeres-Ferradosa, T.; Taveira-Pinto, F.; Romão, X.; Reis, M.; das Neves, L. Reliability assessment of offshore dynamic scour protections using copulas. Wind. Eng. 201843, 506–538. [Google Scholar] [CrossRef]
  24. 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. 2020158, 103671. [Google Scholar] [CrossRef]
  25. Tavouktsoglou, N.S.; Harris, J.M.; Simons, R.R.; Whitehouse, R.J.S. Equilibrium Scour-Depth Prediction around Cylindrical Structures. J. Waterw. Port. Coast. Ocean Eng. 2017143, 04017017. [Google Scholar] [CrossRef][Green Version]
  26. Ettema, R.; Melville, B.; Barkdoll, B. Scale Effect in Pier-Scour Experiments. J. Hydraul. Eng. 1998124, 639–642. [Google Scholar] [CrossRef]
  27. Umeda, S. Scour Regime and Scour Depth around a Pile in Waves. J. Coast. Res. Spec. Issue 201164, 845–849. [Google Scholar]
  28. Umeda, S. Scour process around monopiles during various phases of sea storms. J. Coast. Res. 2013165, 1599–1604. [Google Scholar] [CrossRef]
  29. 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. 2017122, 87–107. [Google Scholar] [CrossRef][Green Version]
  30. Miles, J.; Martin, T.; Goddard, L. Current and wave effects around windfarm monopile foundations. Coast. Eng. 2017121, 167–178. [Google Scholar] [CrossRef][Green Version]
  31. Miozzi, M.; Corvaro, S.; Pereira, F.A.; Brocchini, M. Wave-induced morphodynamics and sediment transport around a slender vertical cylinder. Adv. Water Resour. 2019129, 263–280. [Google Scholar] [CrossRef]
  32. 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. 2019189, 106302. [Google Scholar] [CrossRef]
  33. 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]
  34. 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]
  35. Khalfin, I.S. Modeling and calculation of bed score around large-diameter vertical cylinder under wave action. Water Resour. 200734, 357. [Google Scholar] [CrossRef][Green Version]
  36. 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. 201158, 986–991. [Google Scholar] [CrossRef]
  37. Myrhaug, D.; Rue, H. Scour below pipelines and around vertical piles in random waves. Coast. Eng. 200348, 227–242. [Google Scholar] [CrossRef]
  38. 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. 200936, 605–616. [Google Scholar] [CrossRef]
  39. Myrhaug, D.; Ong, M.C. Random wave-induced onshore scour characteristics around submerged breakwaters using a stochastic method. Ocean Eng. 201037, 1233–1238. [Google Scholar] [CrossRef]
  40. 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. 201373, 106–114. [Google Scholar] [CrossRef]
  41. Yakhot, V.; Orszag, S.A. Renormalization group analysis of turbulence. I. Basic theory. J. Sci. Comput. 19861, 3–51. [Google Scholar] [CrossRef]
  42. Yakhot, V.; Smith, L.M. The renormalization group, the e-expansion and derivation of turbulence models. J. Sci. Comput. 19927, 35–61. [Google Scholar] [CrossRef]
  43. Mastbergen, D.R.; Berg, J.V.D. Breaching in fine sands and the generation of sustained turbidity currents in submarine canyons. Sedimentology 200350, 625–637. [Google Scholar] [CrossRef]
  44. Soulsby, R. Dynamics of Marine Sands; Thomas Telford Ltd.: London, UK, 1998. [Google Scholar] [CrossRef]
  45. Van Rijn, L.C. Sediment Transport, Part I: Bed Load Transport. J. Hydraul. Eng. 1984110, 1431–1456. [Google Scholar] [CrossRef][Green Version]
  46. Zhang, Q.; Zhou, X.-L.; Wang, J.-H. Numerical investigation of local scour around three adjacent piles with different arrangements under current. Ocean Eng. 2017142, 625–638. [Google Scholar] [CrossRef]
  47. 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]
  48. 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. 201657, 114–124. [Google Scholar] [CrossRef]
  49. 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. 201483, 243–258. [Google Scholar] [CrossRef]
  50. 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. 20197, 453. [Google Scholar] [CrossRef][Green Version]
  51. Flow3D User Manual, version 11.0.3; Flow Science, Inc.: Santa Fe, NM, USA, 2013.
  52. Khosronejad, A.; Kang, S.; Sotiropoulos, F. Experimental and computational investigation of local scour around bridge piers. Adv. Water Resour. 201237, 73–85. [Google Scholar] [CrossRef]
  53. 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]
  54. Breusers, H.N.C.; Nicollet, G.; Shen, H. Local Scour Around Cylindrical Piers. J. Hydraul. Res. 197715, 211–252. [Google Scholar] [CrossRef]
  55. Schendel, A.; Hildebrandt, A.; Goseberg, N.; Schlurmann, T. Processes and evolution of scour around a monopile induced by tidal currents. Coast. Eng. 2018139, 65–84. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

      

MDPI and ACS 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. J. Mar. Sci. Eng. 20219, 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.

Study on Hydrodynamic Performance of Unsymmetrical Double Vertical Slotted Barriers

침수된 강성 식생을 갖는 개방 수로 흐름의 특성에 대한 3차원 수치 시뮬레이션

A 3-D numerical simulation of the characteristics of open channel flows with submerged rigid vegetation

Journal of Hydrodynamics volume 33, pages833–843 (2021)Cite this article

Abstract

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

  1. 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.Article Google Scholar 
  2. 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.Article Google Scholar 
  3. 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.Article Google Scholar 
  4. 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.Article MathSciNet Google Scholar 
  5. 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.Article Google Scholar 
  6. 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.Article Google Scholar 
  7. 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.Article MathSciNet Google Scholar 
  8. 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.Article Google Scholar 
  9. Panigrahi K., Khatua K. K. Prediction of velocity distribution in straight channel with rigid vegetation [J]. Aquatic Procedia, 2015, 4: 819–825.Article Google Scholar 
  10. 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.Article Google Scholar 
  11. 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.Article Google Scholar 
  12. 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.Article Google Scholar 
  13. Wu F. S. Characteristics of flow resistance in open channels with non-submerged rigid vegetation [J]. Journal of Hydrodynamics, 2008, 20(2): 239–245.Article Google Scholar 
  14. 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.Article Google Scholar 
  15. 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.Article Google Scholar 
  16. 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.Article MathSciNet Google Scholar 
  17. 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.Article Google Scholar 
  18. 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.Article Google Scholar 
  19. 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.Article Google Scholar 
  20. Aydin M. C., Emiroglu M. E. Determination of capacity of labyrinth side weir by CFD [J]. Flow Measurement and Instrumentation, 2013, 29: 1–8.Article Google Scholar 
  21. 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 
  22. 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.Article Google Scholar 
  23. 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.Article Google Scholar 
  24. 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.Article Google Scholar 
  25. 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.Article Google Scholar 
  26. Fathi-Maghadam M., Kouwen N. Nonrigid, nonsubmerged, vegetative roughness on floodplains [J]. Journal of Hydraulic Engineering, ASCE, 1997, 123(1): 51–57.Article Google Scholar 
  27. 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.Article MathSciNet Google Scholar 
  28. Ghisalberti M., Nepf H. Mass transport in vegetated shear flows [J]. Environmental Fluid Mechanics, 2005, 5(6): 527–551.Article Google Scholar 
Figure 5. Schematic view of flap and support structure [32]

Design Optimization of Ocean Renewable Energy Converter Using a Combined Bi-level Metaheuristic Approach

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

1Introduction

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.

2.1Model 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.

2.2Verification

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������
Quantity0.0124.381.421.681.391.390.084

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.

3Sensitivity 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.

4Design Optimization

We consider optimizing the design parameters of the flap of converter using a nature-based swarm optimization method, that fall in the category of metaheuristic algorithms [45]. Accordingly, we choose four state-of-the-art algorithms to perform an optimization study. Then, based on their performances to achieve the highest capture factor, one of them will be chosen to be combined with the Hill Climb algorithm to carry out a local search. Therefore, in the remainder of this section, we discuss the search process of each algorithm and visualize their performance and convergence curve as they try to find the best values for decision variables.

4.1. Metaheuristic Approaches

As the first considered algorithm, the Gray Wolf Optimizer (GWO) algorithm simulates the natural leadership and hunting performance of gray wolves which tend to live in colonies. Hunters must obey the alpha wolf, the leader, who is responsible for hunting. Then, the beta wolf is at the second level of the gray wolf hierarchy. A subordinate of alpha wolf, beta stands under the command of the alpha. At the next level in this hierarchy, there are the delta wolves. They are subordinate to the alpha and beta wolves. This category of wolves includes scouts, sentinels, elders, hunters, and caretakers. In this ranking, omega wolves are at the bottom, having the lowest level and obeying all other wolves. They are also allowed to eat the prey just after others have eaten. Despite the fact that they seem less important than others, they are really central to the pack survival. Since, it has been shown that without omega wolves, the entire pack would experience some problems like fighting, violence, and frustration. In this simulation, there are three primary steps of hunting including searching, surrounding, and finally attacking the prey. Mathematically model of gray wolves’ hunting technique and their social hierarchy are applied in determined by optimization. this study. As mentioned before, gray wolves can locate their prey and surround them. The alpha wolf also leads the hunt. Assuming that the alpha, beta, and delta have more knowledge about prey locations, we can mathematically simulate gray wolf hunting behavior. Hence, in addition to saving the top three best solutions obtained so far, we compel the rest of the search agents (also the omegas) to adjust their positions based on the best search agent. Encircling behavior can be mathematically modeled by the following equations: [46].(12)�→=|�→·��→(�)-�→(�)|(13)�→(�+1)=��→(�)-�→·�→(14)�→=2.�2→(15)�→=2�→·�1→-�→Where �→indicates the position vector of gray wolf, ��→ defines the vector of prey, t indicates the current iteration, and �→and �→are coefficient vectors. To force the search agent to diverge from the prey, we use �→ with random values greater than 1 or less than -1. In addition, C→ contains random values in the range [0,2], and �→ 1 and �2→ are random vectors in [0,1]. The second considered technique is the Moth Flame Optimizer (MFO) algorithm. This method revolves around the moths’ navigation mechanism, which is realized by positioning themselves and maintaining a fixed angle relative to the moon while flying. This effective mechanism helps moths to fly in a straight path. However, when the source of light is artificial, maintaining an angle with the light leads to a spiral flying path towards the source that causes the moth’s death [47]. In MFO algorithm, moths and flames are both solutions. The moths are actual search agents that fly in hyper-dimensional space by changing their position vectors, and the flames are considered pins that moths drop when searching the search space [48]. The problem’s variables are the position of moths in the space. Each moth searches around a flame and updates it in case of finding a better solution. The fitness value is the return value of each moth’s fitness (objective) function. The position vector of each moth is passed to the fitness function, and the output of the fitness function is assigned to the corresponding moth. With this mechanism, a moth never loses its best solution [49]. Some attributes of this algorithm are as follows:

  • •It takes different values to converge moth in any point around the flame.
  • •Distance to the flame is lowered to be eventually minimized.
  • •When the position gets closer to the flame, the updated positions around the flame become more frequent.

As another method, the Multi-Verse Optimizer is based on a multiverse theory which proposes there are other universes besides the one in which we all live. According to this theory, there are more than one big bang in the universe, and each big bang leads to the birth of a new universe [50]. Multi-Verse Optimizer (MVO) is mainly inspired by three phenomena in cosmology: white holes, black holes, and wormholes. A white hole has never been observed in our universe, but physicists believe the big bang could be considered a white hole [51]. Black holes, which behave completely in contrast to white holes, attract everything including light beams with their extremely high gravitational force [52]. In the multiverse theory, wormholes are time and space tunnels that allow objects to move instantly between any two corners of a universe (or even simultaneously from one universe to another) [53]. Based on these three concepts, mathematical models are designed to perform exploration, exploitation, and local search, respectively. The concept of white and black holes is implied as an exploration phase, while the concept of wormholes is considered as an exploitation phase by MVO. Additionally, each solution is analogous to a universe, and each variable in the solution represents an object in that universe. Furthermore, each solution is assigned an inflation rate, and the time is used instead of iterations. Following are the universe rules in MVO:

  • •The possibility of having white hole increases with the inflation rate.
  • •The possibility of having black hole decreases with the inflation rate.
  • •Objects tend to pass through black holes more frequently in universes with lower inflation rates.
  • •Regardless of inflation rate, wormholes may cause objects in universes to move randomly towards the best universe. [54]

Modeling the white/black hole tunnels and exchanging objects of universes mathematically was accomplished by using the roulette wheel mechanism. With every iteration, the universes are sorted according to their inflation rates, then, based on the roulette wheel, the one with the white hole is selected as the local extremum solution. This is accomplished through the following steps:

Assume that

(16)���=����1<��(��)����1≥��(��)

Where ��� represents the jth parameter of the ith universe, Ui indicates the ith universe, NI(Ui) is normalized inflation rate of the ith universe, r1 is a random number in [0,1], and j xk shows the jth parameter of the kth universe selected by a roulette wheel selection mechanism [54]. It is assumed that wormhole tunnels always exist between a universe and the best universe formed so far. This mechanism is as follows:(17)���=if�2<���:��+���×((���-���)×�4+���)�3<0.5��-���×((���-���)×�4+���)�3≥0.5����:���where Xj indicates the jth parameter of the best universe formed so far, TDR and WEP are coefficients, where Xj indicates the jth parameter of the best universelbjshows the lower bound of the jth variable, ubj is the upper bound of the jth variable, and r2, r3, and r4 are random numbers in [1][54].

Finally, one of the newest optimization algorithms is WOA. The WOA algorithm simulates the movement of prey and the whale’s discipline when looking for their prey. Among several species, Humpback whales have a specific method of hunting [55]. Humpback whales can recognize the location of prey and encircle it before hunting. The optimal design position in the search space is not known a priori, and the WOA algorithm assumes that the best candidate solution is either the target prey or close to the optimum. This foraging behavior is called the bubble-net feeding method. Two maneuvers are associated with bubbles: upward spirals and double loops. A unique behavior exhibited only by humpback whales is bubble-net feeding. In fact, The WOA algorithm starts with a set of random solutions. At each iteration, search agents update their positions for either a randomly chosen search agent or the best solution obtained so far [56][55]. When the best search agent is determined, the other search agents will attempt to update their positions toward that agent. It is important to note that humpback whales swim around their prey simultaneously in a circular, shrinking circle and along a spiral-shaped path. By using a mathematical model, the spiral bubble-net feeding maneuver is optimized. The following equation represents this behavior:(18)�→(�+1)=�′→·�bl·cos(2��)+�∗→(�)

Where:(19)�′→=|�∗→(�)-�→(�)|

X→(t+ 1) indicates the distance of the it h whale to the prey (best solution obtained so far),� is a constant for defining the shape of the logarithmic spiral, l is a random number in [−1, 1], and dot (.) is an element-by-element multiplication [55].

Comparing the four above-mentioned methods, simulations are run with 10 search agents for 400 iterations. In Figure 15, there are 20 plots the optimal values of different parameters in optimization algorithms. The five parameters of this study are freeboard, bottom elevations, number of ribs on the converter, rib thickness, and rib Height. The optimal value for each was found by optimization algorithms, naming WOA, MVO, MFO, and GWO. By looking through the first row, the freeboard parameter converges to its maximum possible value in the optimization process of GWO after 300 iterations. Similarly, MFO finds the same result as GWO. In contrast, the freeboard converges to its minimum possible value in MVO optimizing process, which indicates positioning the converter under the water. Furthermore, WOA found the optimal value of freeboard as around 0.02 after almost 200 iterations. In the second row, the bottom elevation is found at almost 0.11 (m) in all algorithms; however, the curves follow different trends in each algorithm. The third row shows the number of ribs, where results immediately reveal that it should be over 4. All algorithms coincide at 5 ribs as the optimal number in this process. The fourth row displays the trends of algorithms to find optimal rib thickness. MFO finds the optimal value early and sets it to around 0.022, while others find the same value in higher iterations. Finally, regarding the rib height, MVO, MFO, and GWO state that the optimal value is 0.06 meters, but WOA did not find a higher value than 0.039.

4.2. HCMVO Bi-level Approach

Despite several strong search characteristics of MVO and its high performance in various optimization problems, it suffers from a few deficiencies in local and global search mechanisms. For instance, it is trapped in the local optimum when wormholes stochastically generate many solutions near the best universe achieved throughout iterations, especially in solving complex multimodal problems with high dimensions [57]. Furthermore, MVO needs to be modified by an escaping strategy from the local optima to enhance the global search abilities. To address these shortages, we propose a fast and effective meta-algorithm (HCMVO) to combine MVO with a Random-restart hill-climbing local search. This meta-algorithm uses MVO on the upper level to develop global tracking and provide a range of feasible and proper solutions. The hill-climbing algorithm is designed to develop a comprehensive neighborhood search around the best-found solution proposed by the upper-level (MVO) when MVO is faced with a stagnation issue or falling into a local optimum. The performance threshold is formulated as follows.(20)Δ����THD=∑�=1�����TH��-����TH��-1�where BestTHDis the best-found solution per generation, andM is related to the domain of iterations to compute the average performance of MVO. If the proposed best solution by the local search is better than the initial one, the global best of MVO will be updated. HCMVO iteratively runs hill climbing when the performance of MVO goes down, each time with an initial condition to prepare for escaping such undesirable situations. In order to get a better balance between exploration and exploitation, the search step size linearly decreases as follows:(21)��=��-����Ma�iter��+1where iter and Maxiter are the current iteration and maximum number of evaluation, respectively. �� stands for the step size of the neighborhood search. Meanwhile, this strategy can improve the convergence rate of MVO compared with other algorithms.

Algorithm 1 shows the technical details of the proposed optimization method (HCMVO). The initial solution includes freeboard (�), bottom elevation (�), number of ribs (Nr), rib thickness (�), and rib height(�).

5. Conclusion

The high trend of diminishing worldwide energy resources has entailed a great crisis upon vulnerable societies. To withstand this effect, developing renewable energy technologies can open doors to a more reliable means, among which the wave energy converters will help the coastal residents and infrastructure. This paper set out to determine the optimized design for such devices that leads to the highest possible power output. The main goal of this research was to demonstrate the best design for an oscillating surge wave energy converter using a novel metaheuristic optimization algorithm. In this regard, the methodology was devised such that it argued the effects of influential parameters, including wave characteristics, WEC design, and interaction criteria.

To begin with, a numerical model was developed in Flow 3D software to simulate the response of the flap of a wave energy converter to incoming waves, followed by a validation study based upon a well-reputed experimental study to verify the accuracy of the model. Secondly, the hydrodynamics of the flap was investigated by incorporating the turbulence. The effect of depth, wave height, and wave period are also investigated in this part. The influence of two novel ideas on increasing the wave-converter interaction was then assessed: i) designing a flap with different widths in the upper and lower part, and ii) adding ribs on the surface of the flap. Finally, four trending single-objective metaheuristic optimization methods

Empty CellAlgorithm 1: Hill Climb Multiverse Optimization
01:procedure HCMVO
02:�=30,�=5▹���������������������������������
03:�=〈F1,B1,N,R,H1〉,…〈FN,B2,N,R,HN〉⇒lb1N⩽�⩽ubN
04:Initialize parameters�ER,�DR,�EP,Best�,���ite��▹Wormhole existence probability (WEP)
05:��=����(��)
06:��=Normalize the inflation rate��
07:for iter in[1,⋯,���iter]do
08:for�in[1,⋯,�]do
09:Update�EP,�DR,Black����Index=�
10:for���[1,⋯,�]��
11:�1=����()
12:if�1≤��(��)then
13:White HoleIndex=Roulette�heelSelection(-��)
14:�(Black HoleIndex,�)=��(White HoleIndex,�)
15:end if
16:�2=����([0,�])
17:if�2≤�EPthen
18:�3=����(),�4=����()
19:if�3<0.5then
20:�1=((��(�)-��(�))�4+��(�))
21:�(�,�)=Best�(�)+�DR�
22:else
23:�(�,�)=Best�(�)-�DR�
24:end if
25:end if
26:end for
27:end for
28:�HD=����([�1,�2,⋯,�Np])
29:Bes�TH�itr=����HD
30:ΔBestTHD=∑�=1�BestTII��-BestTII��-1�
31:ifΔBestTHD<��then▹Perform hill climbing local search
32:BestTHD=����-�lim��������THD
33:end if
34:end for
35:return�,BestTHD▹Final configuration
36:end procedure

The implementation details of the hill-climbing algorithm applied in HCMPA can be seen in Algorithm 2. One of the critical parameters isg, which denotes the resolution of the neighborhood search around the proposed global best by MVO. If we set a small step size for hill-climbing, the convergence speed will be decreased. On the other hand, a large step size reinforces the exploration ability. Still, it may reduce the exploitation ability and in return increase the act of jumping from a global optimum or surfaces with high-potential solutions. Per each decision variable, the neighborhood search evaluates two different direct searches, incremental or decremental. After assessing the generated solutions, the best candidate will be selected to iterate the search algorithm. It is noted that the hill-climbing algorithm should not be applied in the initial iteration of the optimization process due to the immense tendency for converging to local optima. Meanwhile, for optimizing largescale problems, hill-climbing is not an appropriate selection. In order to improve understanding of the proposed hybrid optimization algorithm’s steps, the flowchart of HCMVO is designed and can be seen in Figure 16.

Figure 17 shows the observed capture factor (which is the absorbed energy with respect to the available energy) by each optimization algorithm from iterations 1 to 400. The algorithms use ten search agents in their modified codes to find the optimal solutions. While GWO and MFO remain roughly constant after iterations 54 and 40, the other three algorithms keep improving the capture factor. In this case, HCMVO and MVO worked very well in the optimizing process with a capture factor obtained by the former as 0.594 and by the latter as 0.593. MFO almost found its highest value before the iteration 50, which means the exploration part of the algorithm works out well. Similarly, HCMVO does the same. However, it keeps finding the better solution during the optimization process until the last iteration, indicating the strong exploitation part of the algorithm. GWO reveals a weakness in exploration and exploitation because not only does it evoke the least capture factor value, but also the curve remains almost unchanged throughout 350 iterations.

Figure 18 illustrates complex interactions between the five optimization parameters and the capture factor for HCMVO (a), MPA (b), and MFO (c) algorithms. The first interesting observation is that there is a high level of nonlinear relationships among the setting parameters that can make a multi-modal search space. The dark blue lines represent the best-found configuration throughout the optimisation process. Based on both HCMVO (a) and MVO (b), we can infer that the dark blue lines concentrate in a specific range, showing the high convergence ability of both HCMVO and MVO. However, MFO (c) could not find the exact optimal range of the decision variables, and the best-found solutions per generation distribute mostly all around the search space.

Empty CellAlgorithm 1: Hill Climb Multiverse Optimization
01:procedure HCMVO
02:Initialization
03:Initialize the constraints��1�,��1�
04:�1�=Mi�1�+���1�/�▹Compute the step size,�is search resolution
05:So�1=〈�,�,�,�,�〉▹���������������
06:�������1=����So�1▹���������ℎ���������
07:Main loop
08:for iter≤���ita=do
09:���=���±��
10:while�≤���(Sol1)do
11:���=���+�,▹����ℎ���ℎ��������ℎ
12:fitness��iter=�������
13:t = t+1
14:end while
15:〈�����,������max〉=����������
16:���itev=���Inde�max▹�������ℎ�������������������������������ℎ�������
17:��=��-����Max��+1▹�����������������
18:end for
19:return���iter,����
20:end procedure

were utilized to illuminate the optimum values of the design parameters, and the best method was chosen to develop a new algorithm that performs both local and global search methods.

The correlation between hydrodynamic parameters and the capture factor of the converter was supported by the results. For any given water depth, the capture factor increases as the wave period increases, until a certain wave period value (6 seconds) is reached, after which the capture factor gradually decreases. It is expected since the flap cannot oscillate effectively when the wavelength is too short for a certain water depth. Conversely, when the wavelength is too long, the capture factor decreases. Furthermore, under a constant wave period, increasing the water depth does not affect the capture factor. Regarding the sensitivity analysis, the study found that increasing the flap bottom elevation causes turbulence flow behind the flap and limitation of rotation, which leads to less interaction with the incoming waves. Furthermore, while keeping the flap bottom elevation constant, increasing the freeboard improves the capture factor. Overtopping happens when the freeboard is negative and the flap is below the water surface, which has a detrimental influence on converter performance. Furthermore, raising the freeboard causes the wave impact to become more violent, which increases converter performance.

In the last part, we discussed the search process of each algorithm and visualized their performance and convergence curves as they try to find the best values for decision variables. Among the four selected metaheuristic algorithms, the Multi-verse Optimizer proved to be the most effective in achieving the best answer in terms of the WEC capture factor. However, the MVO needed modifications regarding its escape approach from the local optima in order to improve its global search capabilities. To overcome these constraints, we presented a fast and efficient meta-algorithm (HCMVO) that combines MVO with a Random-restart hill-climbing local search. On a higher level, this meta-algorithm employed MVO to generate global tracking and present a range of possible and appropriate solutions. Taken together, the results demonstrated that there is a significant degree of nonlinearity among the setup parameters that might result in a multimodal search space. Since MVO was faced with a stagnation issue or fell into a local optimum, we constructed a complete neighborhood search around the best-found solution offered by the upper level. In sum, the newly-developed algorithm proved to be highly effective for the problem compared to other similar optimization methods. The strength of the current findings may encourage future investigation on design optimization of wave energy converters using developed geometry as well as the novel approach.

CRediT authorship contribution statement

Erfan Amini: Conceptualization, Methodology, Validation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Mahdieh Nasiri: Conceptualization, Methodology, Validation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Navid Salami Pargoo: Writing – original draft, Writing – review & editing. Zahra Mozhgani: Conceptualization, Methodology. Danial Golbaz: Writing – original draft. Mehrdad Baniesmaeil: Writing – original draft. Meysam Majidi Nezhad: . Mehdi Neshat: Supervision, Conceptualization, Writing – original draft, Writing – review & editing, Visualization. Davide Astiaso Garcia: Supervision. Georgios Sylaios: Supervision.

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.

Acknowledgement

This research has been carried out within ILIAD (Inte-grated Digital Framework for Comprehensive Maritime Data and Information Services) project that received funding from the European Union’s H2020 programme.

Data availability

Data will be made available on request.

References

Effect of tailwater depth on non-cohesive earth dam failure due to overtopping

Effect of tailwater depth on non-cohesive earth dam failure due to overtopping

범람으로 인한 비점착성 흙댐 붕괴에 대한 테일워터 깊이의 영향

ShaimaaAmanaMohamedAbdelrazek RezkbRabieaNasrc

Abstract

본 연구에서는 범람으로 인한 토사댐 붕괴에 대한 테일워터 깊이의 영향을 실험적으로 조사하였다. 테일워터 깊이의 네 가지 다른 값을 검사합니다. 각 실험에 대해 댐 수심 측량 프로파일의 진화, 고장 기간, 침식 체적 및 유출 수위곡선을 관찰하고 기록합니다.

결과는 tailwater 깊이를 늘리면 고장 시간이 최대 57% 감소하고 상대적으로 침식된 마루 높이가 최대 77.6% 감소한다는 것을 보여줍니다. 또한 상대 배수 깊이가 3, 4, 5인 경우 누적 침식 체적의 감소는 각각 23, 36.5 및 75%인 반면 최대 유출량의 감소는 각각 7, 14 및 17.35%입니다.

실험 결과는 침식 과정을 복제할 때 Flow 3D 소프트웨어의 성능을 평가하는 데 활용됩니다. 수치 모델은 비응집성 흙댐의 침식 과정을 성공적으로 시뮬레이션합니다.

The influence of tailwater depth on earth dam failure due to overtopping is investigated experimentally in this work. Four different values of tailwater depths are examined. For each experiment, the evolution of the dam bathymetry profile, the duration of failure, the eroded volume, and the outflow hydrograph are observed and recorded. The results reveal that increasing the tailwater depth reduces the time of failure by up to 57% and decreases the relative eroded crest height by up to 77.6%. In addition, for relative tailwater depths equal to 3, 4, and 5, the reduction in the cumulative eroded volume is 23, 36.5, and 75%, while the reduction in peak discharge is 7, 14, and 17.35%, respectively. The experimental results are utilized to evaluate the performance of the Flow 3D software in replicating the erosion process. The numerical model successfully simulates the erosion process of non-cohesive earth dams.

Keywords

Earth dam, Eroded volume, Flow 3D model, Non-cohesive soil, Overtopping failure, Tailwater depth

Notation

d50

Mean partical diameterWc

Optimum water contentZo

Dam height (cm)do

Tailwater depth (cm)Zeroded

Eroded height of the dam measured at distance of 0.7 m from the dam heel (cm)t

Total time of failure (sec)t1

Time of crest width erosion (sec)Zcrest

The crest height (cm)Vtotal

Total volume of the dam (m3)Veroded

Cumulative eroded volume (m3)RMSE

The statistical variable root- mean- square errord

Degree of agreement indexyu.s.

The upstream water depth (cm)yd.s

The downstream water depth (cm)H

Water surface elevation over sharp crested weir (cm)Q

Outflow discharge (liter/sec)Qpeak

Peak discharge (liter/sec)

1. Introduction

Earth dams are compacted structures composed of natural materials that are usually mined or quarried from local locations. The failures of the earth dams have proven to be deadly, destructive, and costly. According to People’s Daily, two earthen dams, Yong’an Dam and Xinfa Dam located in Hulun Buir City in North China’s Inner Mongolia failed on 2021, due to a surge in the water level of the Nuomin River caused by heavy rain. The dam breach affected 16,660 people, flooded 325,622 mu of farmland (21708.1 ha), and destroyed 22 bridges, 124 culverts, and 15.6 km of roadways. Also, the failure of south fork dam (earth and rock fill dam) near Johnstown on 1889 is considered the worst U.S dam disaster in terms of loss of life. The dam was overtopped and washed away due to unexpected heavy rains, releasing 20 million tons of water which destroyed Johnstown and resulted in 2209 deaths, [1][2]. Piping or shear sliding, failure due to natural factors, and failure due to overtopping are all possible causes of earth dam failure. However, overtopping failure is the most frequent cause of dam failure. According to The International Committee on Large Dams (ICOLD, 1995), and [3], more than one-third of the total known dam failures were caused by dam overtopping.

Overtopping occurs as the result of insufficient flood design or freeboard in some cases. Extreme rainstorms can cause floods which can overtop the dam and cause it to fail. The size and geometry of the reservoir or the dam (side slopes, top width, height, etc.), the homogeneity of the material used in the construction of the dam, overtopping depth, and the presence or absence of tailwater are all elements that influence this type of failure which will be illustrated in the following literature. Overtopping failures of earth dams may be divided into several failure mechanisms based on the material composition and the inner structure of the dam. For cohesive earth dams because of low permeability, no seepage exists on the slopes. Erosion often begins at the earth dam toe during turbulent erosion and moves upstream, undercutting the slope, causing the removal of large chunks of materials. While for non-cohesive earth dams the downstream face of the dam flattens progressively and is often said to rotate around a point near the downstream toe [4][5][6] In the last few decades, the study of failures due to overtopping has gained popularity among researchers. The overtopping failure, in fact, has been widely investigated in coastal and river hydraulics and morpho dynamic. In addition, several laboratory experimental studies have been conducted in this field in order to better understand different involved factors. Also, many numerical types of research have been conducted to investigate the process of overtopping failure as well as the elements that influence this type of failure.

Tabrizi et al. [5] conducted a series of embankment overtopping tests to find the effect of compaction on the failure of a homogenous sand embankment. A plane breach process occurred across the flume width due to the narrow flume width. They measured the downstream hydrographs and embankment surface profile for every case. They concluded that the peak discharge decreased with a high compaction level, while the time to peak increased. Kansoh et al. [6] studied experimentally the failure of compacted homogeneous non-cohesive earthen embankment due to overtopping. They investigated the influence of different shape parameters including the downstream slope, the crest width, and the height of the embankment on the erosion process. The erosion process was initiated by carving a pilot channel into the embankment crest. They evaluated the time of embankment failure for different shape parameters. They concluded that the failure time increases with increasing the downstream slope and the crest width. Zhu et al. [7] investigated experimentally the breaching of five embankments, one constructed with pure sand, and four with different sand-silt–clay mixtures. The erosion pattern was similar across the flume width. They stated that for cohesive soil mixtures the head cut erosion was the most important factor that affected the breach growth, while for non-cohesive soil the breach erosion was affected by shear erosion.

Amaral et al. [8] studied experimentally the failure by overtopping for two embankments built from silt sand material. They studied the effect of the degree of compaction of the embankment and the geometry of the pilot channel carved at the centre of the dam crest. They studied two shapes of pilot channel a rectangular shape and triangular shape. They stated that the breach development is influenced by a higher degree of compaction, however, the pilot channel geometry did not influence the breach’s final form. Bereta et al. [9] studied experimentally the breach formation of five dam models, three of them were homogenous clay soil while two were sandy-clay mixtures. The erosion process was initiated by cutting a pilot channel at the centre of the dam crest. They observed the initiation of erosion, flow shear erosion, sidewall bottom erosion, and distinguished the soil mechanical slope mass failure from the head cut vertically and laterally during these tests. Verma et al. [10] investigated experimentally a two-dimensional erosion phenomenon due to overtopping by using a wooden fuse plug model and five different soils. They concluded that the erosion process was affected mostly by cohesiveness and degree of compaction. For cohesive soils, a head cut erosion was observed, while for non-cohesive soils surface erosion occurred gradually. Also, the dimensions of fuse plug, type of fill material, reservoir capacity, and inflow were found to affect the behaviour of the overall breaching process.

Wu and Qin [11] studied the effect of adding coarse grains to the downstream face of a non-cohesive dam as a result of tailings deposition. The process of overtopping during tailings dam failures is analyzed and its effect on delaying the dam-break process and disaster mitigation are investigated. They found that the tested protective measures decreased the breach area, the maximum breaching flow discharge and flow velocity, and the downstream inundated area. Khankandi et al. [12] studied experimentally the effect of reservoir geometry on dam break flow in case of dry and wet bed conditions. They considered four different reservoir shapes, a long reservoir, a wide, a trapezoidal shaped and one with a 90◦ bend all with identical water volume and horizontal bed. The dam break is simulated by the sudden gate removal using a pneumatic jack. They measured the variation of water level over time with ultrasonic sensors and flow velocity component with an acoustic Doppler velocimeter. Also, the experimental results of water level variation are compared with Ritters solution (1892) [13]. They stated that for dry bed condition the long and 90 bend reservoirs results are close to the analytical solution by ritter also in these two shapes a 1D flow is noticed. However, for wide and trapezoidal reservoirs a 2D effect is significant due to flow contraction at channel entrance.

Rifai et al. [14] conducted a series of experiments to investigate the effect of tailwater depth on the outflow discharge and breach geometry during non-cohesive homogenous fluvial dikes overtopping failure. They cut an initial notch in the crest at 0.8 m from the upstream end of the dike to initiate overtopping. They compared their results to previous experiments under different main channel inflow discharges combined with a free floodplain. They divided the dike breaching process into three stages: gradual start of overtopping flow resulting in slow initiation of dike erosion, deepening and widening breach due to large flow depth and velocity, finally the flow depth starts stabilizing at its minimal level with or without sustained breach expansion. They stated that breach discharge has lower values than in free floodplain tests. Jiang [15] studied the effect of bed slope on breach parameters and peak discharge in non-cohesive embankment failure. An initial triangular breach with a depth and width of 4 cm was pre-set on one side of the dam. He stated that peak discharge increases with the increase of bed slope and then decreases.

Ozmen-cagatay et al. [16] studied experimentally flood wave propagation resulted from a sudden dam break event. For dam-break modelling, they used a mechanism that permitted the rapid removal of a vertical plate with a thickness of 4 mm and made of rigid plastic. They conducted three tests, one with dry bed condition and two tests with tailwater depths equal 0.025 m and 0.1 m respectively. They recorded the free surface profile during initial stages of dam break by using digital image processing. Finally, they compared the experimental results with the with a commercially available VOF-based CFD program solving the Reynolds-averaged Navier –Stokes equations (RANS) with the k– Ɛ turbulence model and the shallow water equations (SWEs). They concluded that Wave breaking was delayed with increasing the tailwater depth to initial reservoir depth ratio. They also stated that the SWE approach is sufficient more to represent dam break flows for wet bed condition. Evangelista [17] investigated experimentally and numerically using a depth-integrated two-phase model, the erosion of sand dike caused by the impact of a dam break wave. The dam break is simulated by a sudden opening of an upstream reservoir gate resulting in the overtopping of a downstream trapezoidal sand dike. The evolution of the water wave caused from the gate opening and dike erosion process are recorded by using a computer-controlled camera. The experimental results demonstrated that the progression of the wave front and dike erosion have a considerable influence on each other during the process. In addition, the dike constructed from fine sands was more resistant to erosion than the one built with coarse sand. They also stated that the numerical model can is capable of accurately predicting wave front position and dike erosion. Also, Di Cristo et al. [18] studied the effect of dam break wave propagation on a sand embankment both experimentally and numerically using a two-phase shallow-water model. The evolution of free surface and of the embankment bottom are recorded and used in numerical model assessment. They stated that the model allows reasonable simulation of the experimental trends of the free surface elevation regardeless of the geofailure operator.

Lots of numerical models have been developed over the past few years to simulate the dam break flooding problem. A one-dimensional model, such as Hec-Ras, DAMBRK and MIKE 11, ect. A two-dimensional model such as iRIC Nay2DH is used in earth embankment breach simulation. Other researchers studied the failure process numerically using (3D) computational fluid dynamics (CFD) models, such as FLOW-3D, and FLUENT. Goharnejad et al. [19] determined the outflow hydrograph which results from the embankment dam break due to overtopping. Hu et al. [20] performed a comparison between Flow-3D and MIKE3 FM numerical models in simulating a dam break event under dry and wet bed conditions with different tailwater depths. Kaurav et al. [21] simulated a planar dam breach process due to overtopping. They conducted a sensitivity analysis to find the effect of dam material, dam height, downstream slope, crest width, and inlet discharge on the erosion process and peak discharge through breach. They concluded that downstream slope has a significant influence on breaching process. Yusof et al. [22] studied the effect of embankment sediment sizes and inflow rates on breaching geometric and hydrodynamic parameters. They stated that the peak outflow hydrograph increases with increasing sediment size and inflow rates while time of failure decreases.

In the present work, the effect of tailwater depth on earth dam failure during overtopping is studied experimentally. The relation between the eroded volume of the dam and the tailwater depth is presented. Also, the percentage of reduction in peak discharge due to tailwater existence is calculated. An assessment of Flow 3D software performance in simulating the erosion process during earth dam failure is introduced. The statistical variable root- mean- square error, RMSE, and the agreement degree index, d, are used in model assessment.

2. Material and methods

The tests are conducted in a straight rectangular flume in the laboratory of Irrigation Engineering and Hydraulics Department, Faculty of Engineering, Alexandria University, Egypt. The flume dimensions are 10 m long, 0.86 m wide, and 0.5 m deep. The front part of the flume is connected to a storage basin 1 m long by 0.86 m wide. The storage basin is connected to a collecting tank for water recirculation during the experiments as shown in Fig. 1Fig. 2. A sharp-crested weir is placed at a distance of 4 m downstream the constructed dam to keep a constant tailwater depth in each experiment and to measure the outflow discharge.

To measure the eroded volume with time a rods technique is used. This technique consists of two parallel wooden plates with 10 cm distance in between and five rows of stainless-steel rods passing vertically through the wooden plates at a spacing of 20 cm distributed across flume width. Each row consists of four rods with 15 cm spacing between them. Also, a graph board is provided to measure the drop in each rod with time as shown in Fig. 3Fig. 4. After dam construction the rods are carefully rested on the dam, with the first line of rods resting in the middle of the dam crest and then a constant distance of 15 cm between rods lines is maintained.

A soil sample is taken and tested in the laboratory of the soil mechanics to find the soil geotechnical parameters. The soil particle size distribution is also determined by sieve analysis as shown in Fig. 5. The soil mean diameter d50,equals 0.38 mm and internal friction angle equals 32.6°.

2.1. Experimental procedures

To investigate the effect of the tailwater depth (do), the tailwater depth is changed four times 5, 15, 20, and 25 cm on the sand dam model. The dam profile is 35 cm height, with crest width = 15 cm, the dam base width is 155 cm, and the upstream and downstream slopes are 2:1 as shown in Fig. 6. The dam dimensions are set as the flume permitted to allow observation of the dam erosion process under the available flume dimensions and conditions. All of the conducted experiments have the same dimensions and configurations.

The optimum water content, Wc, from the standard proctor test is found to be 8 % and the maximum dry unit weight is 19.42 kN/m3. The soil and water are mixed thoroughly to ensure consistency and then placed on three horizontal layers. Each layer is compacted according to ASTM standard with 25 blows by using a rammer (27 cm × 20.5 cm) weighing 4 kg. Special attention is paid to the compaction of the soil to guarantee the repeatability of the tests.

After placing and compacting the three layers, the dam slopes are trimmed carefully to form the trapezoidal shape of the dam. A small triangular pilot channel with 1 cm height and 1:1 side slopes is cut into the dam crest to initiate the erosion process. The position of triangular pilot channel is presented in Fig. 1. Three digital video cameras with a resolution of 1920 × 1080 pixels and a frame rate of 60 fps are placed in three different locations. One camera on one side of the flume to record the progress of the dam profile during erosion. Another to track the water level over the sharp-crested rectangular weir placed at the downstream end of the flume. And the third camera is placed above the flume at the downstream side of the dam and in front of the rods to record the drop of the tip of the rods with time as shown previously in Fig. 1.

Before starting the experiment, the water is pumped into the storage basin by using pump with capacity 360 m3/hr, and then into the upstream section of the flume. The upstream boundary is an inflow condition. The flow discharge provided to the storage basin is kept at a constant rate of 6 L/sec for all experiments, while the downstream boundary is an outflow boundary condition.

Also, the required tailwater depth for each experiment is filled to the desired depth. A dye container valve is opened to color the water upstream of the dam to make it easy to distinguish the dam profile from the water profile. A wooden board is placed just upstream of the dam to prevent water from overtopping the dam until the water level rises to a certain level above the dam crest and then the wooden board is removed slowly to start the experiment.

2.2. Repeatability

To verify the accuracy of the results, each experiment is repeated two times under the same conditions. Fig. 7 shows the relative eroded crest height, Zeroded / Zo, with time for 5 cm tailwater depth. From the Figure, it can be noticed that results for all runs are consistent, and accuracy is achieved.

3. Numerical model

The commercially available numerical model, Flow 3D is used to simulate the dam failure due to overtopping for the cases of 15 cm, 20 cm and 25 cm tailwater depths. For numerical model calibration, experimental results for dam surface evolution are used. The numerical model is calibrated for selection of the optimal turbulence model (RNG, K-e, and k-w) and sediment scour equations (Van Rin, Meyer- peter and Muller, and Nielsen) that produce the best results. In this, the flow field is solved by the RNG turbulence model, and the van Rijn equation is used for the sediment scour model. A geometry file is imported before applying the mesh.

A Mesh sensitivity is analyzed and checked for various cell sizes, and it is found that decreasing the cell size significantly increases the simulation time with insignificant differences in the result. It is noticed that the most important factor influencing cell size selection is the value of the dam’s upstream and downstream slopes. For example, the slopes in the dam model are 2:1, thus the cell size ratio in X and Z directions should be 2:1 as well. The cell size in a mesh block is set to be 0.02 m, 0.025 m, and 0.01 m in X, Y and Z directions respectively.

In the numerical computations, the boundary conditions employed are the walls for sidewalls and the channel bottom. The pressure boundary condition is applied at the top, at the air–water interface, to account for atmospheric pressure on the free surface. The upstream boundary is volume flow rate while the downstream boundary is outflow discharge.

The initial condition is a fluid region, which is used to define fluid areas both upstream and downstream of the dam. To assess the model accuracy, the statistical variable root- mean- square error, RMSE, and the agreement degree index, d, are calculated as(1)RMSE=1N∑i=1N(Pi-Mi)2(2)d=1-∑Mi-Pi2∑Mi-M¯+Pi-P¯2

where N is the number of samples, Pi and Mi are the models and experimental values, P and M are the means of the model and experimental values. The best fit between the experimental and model results would have an RMSE = 0 and degree of agreement, d = 1.

4. Results of experimental work

The results of the total time of failure, t (defined as the time from when the water begins to overtop the dam crest until the erosion reaches a steady state, when no erosion occurs), time of crest width erosion t1, cumulative eroded volume Veroded, and peak discharge Qpeak for each experiment are listed in Table 1. The case of 5 cm tailwater depth is considered as a reference case in this work.

Table 1. Results of experimental work.

Tailwater depth, do (cm)Total time of failure, t (sec)Time of crest width erosion, t1 (sec)cumulative eroded volume, Veroded (m3)Peak discharge, Qpeak (liter/sec)
5255220.2113.12
15165300.1612.19
20140340.1311.29
25110390.0510.84

5. Discussion

5.1. Side erosion

The evolution of the bathymetry of the erosion line recorded by the video camera1. The videos are split into frames (60 frames/sec) by the Free Video to JPG Converter v.5.063 build and then converted into an excel spreadsheet using MATLAB code as shown in Fig. 8.

Fig. 9 shows a sample of numerical model output. Fig. 10Fig. 11Fig. 12 show a dam profile development for different time steps from both experimental and numerical model, for tailwater depths equal 15 cm, 20 cm and 25 cm. Also, the values of RMSE and d for each figure are presented. The comparison shows that the Flow 3D software can simulate the erosion process of non-cohesive earth dam during overtopping with an RMSE value equals 0.023, 0.0218, and 0.0167 and degree of agreement, d, equals 0.95, 0.968, and 0.988 for relative tailwater depths, do/(do)ref, = 3, 4 and 5, respectively. The low values of RMSE and high values of d show that the Flow 3D can effectively simulate the erosion process. From Fig. 10Fig. 11Fig. 12, it can be noticed that the model is not capable of reproducing the head cut, while it can simulate well the degradation of the crest height with a minor difference from experimental work. The reason of this could be due to inability of simulation of all physical conditions which exists in the experimental work, such as channel friction and the grain size distribution of the dam soil which is surely has a great effect on the erosion process and breach development. In the experimental work the grain size distribution is shown in Fig. 5, while the numerical model considers that the soil is uniform and exactly 50 % of the dam particles diameter are equal to the d50 value. Another reason is that the model is not considering the increased resistance of the dam due to the apparent cohesion which happens due to dam saturation [23].

It is clear from both the experimental and numerical results that for a 5 cm tailwater depth, do/(do)ref = 1.0, erosion begins near the dam toe and continues upward on the downstream slope until it reaches the crest. After eroding the crest width, the crest is lowered, resulting in increased flow rates and the speeding up of the erosion process. While for relative tailwater depths, do/(do)ref = 3, 4, and 5 erosion starts at the point of intersection between the downstream slope and tailwater. The existence of tailwater works as an energy dissipater for the falling water which reduces the erosion process and prevents the dam from failure as shown in Fig. 13. It is found that the time of the failure decreases with increasing the tailwater depth because most of the dam height is being submerged with water which decreases the erosion process. The reduction in time of failure from the referenced case is found to be 35.3, 45, and 57 % for relative tailwater depth, do /(do)ref equals 3, 4, and 5, respectively.

The relation between the relative eroded crest height, Zeroded /Zo, with time is drawn as shown in Fig. 14. It is found that the relative eroded crest height decreases with increasing tailwater depth by 10, 41, and 77.6 % for relative tailwater depth, do /(do)ref equals 3, 4, and 5, respectively. The time required for the erosion of the crest width, t1, is calculated for each experiment. The relation between relative tailwater depth and relative time of crest width erosion is shown in Fig. 15. It is found that the time of crest width erosion increases linearly with increasing, do /Zo. The percent of increase is 36.4, 54.5 and 77.3 % for relative tailwater depth, do /(do)ref = 3, 4 and 5, respectively.

Crest height, Zcrest is calculated from the experimental results and the Flow 3D results for relative tailwater depths, do/(do)ref, = 3, 4, and 5. A relation between relative crest height, Zcrest/Zo with time from experimental and numerical results is presented in Fig. 16. From Fig. 16, it is seen that there is a good consistency between the results of numerical model and the experimental results in the case of tracking the erosion of the crest height with time.

5.2. Upstream and downstream water depths

It is noticed that at the beginning of the erosion process, both upstream and downstream water depths increase linearly with time as long as erosion of the crest height did not take place. However, when the crest height starts to lower the upstream water depth decreases with time while the downstream water depth increases. At the end of the experiment, the two depths are nearly equal. A relation between relative downstream and upstream water depths with time is drawn for each experiment as shown in Fig. 17.

5.3. Eroded volume

A MATLAB code is used to calculate the cumulative eroded volume every time interval for each experiment. The total volume of the dam, Vtotal is 0.256 m3. The cumulative eroded volume, Veroded is 0.21, 0.16, 0.13, and 0.05 m3 for tailwater depths, do = 5, 15, 20, and 25 cm, respectively. Fig. 18 presents the relation between cumulative eroded volume, Veroded and time. From Fig. 18, it is observed that the cumulative eroded volume decreases with increasing the tailwater depth. The reduction in cumulative eroded volume is 23, 36.5, and 75 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively. The relative remained volume of the dam equals 0.18, 0.375, 0.492, and 0.8 for tailwater depths = 5, 15, 20, and 25 cm, respectively. Fig. 19 shows a relation between relative tailwater depth and relative cumulative eroded volume from experimental results. From that figure, it is noticed that the eroded volume decreases exponentially with increasing relative tailwater depth.

5.4. The outflow discharge

The inflow discharge provided to the storage tank is maintained constant for all experiments. The water surface elevation, H, over the sharp-crested weir placed at the downstream side is recorded by the video camera 2. For each experiment, the outflow discharge is then calculated by using the sharp-crested rectangular weir equation every 10 sec.

The outflow discharge is found to increase rapidly until it reaches its peak then it decreases until it is constant. For high values of tailwater depths, the peak discharge becomes less than that in the case of small tailwater depth as shown in Fig. 20 which agrees well with the results of Rifai et al. [14] The reduction in peak discharge is 7, 14, and 17.35 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively.

The scenario presented in this article in which the tailwater depth rises due to unexpected heavy rainfall, is investigated to find the effect of rising tailwater depth on earth dam failure. The results revealed that rising tailwater depth positively affects the process of dam failure in terms of preventing the dam from complete failure and reducing the outflow discharge.

6. Conclusions

The effect of tailwater depth on earth dam failure due to overtopping is investigated experimentally in this work. The study focuses on the effect of tailwater depth on side erosion, upstream and downstream water depths, eroded volume, outflow hydrograph, and duration of the failure process. The Flow 3D numerical software is used to simulate the dam failure, and a comparison is made between the experimental and numerical results to find the ability of this software to simulate the erosion process. The following are the results of the investigation:

The existence of tailwater with high depths prevents the dam from completely collapsing thereby turning it into a broad crested weir. The failure time decreases with increasing the tailwater depth and the reduction from the reference case is found to be 35.3, 45, and 57 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively. The difference between the upstream and downstream water depths decreases with time till it became almost negligible at the end of the experiment. The reduction in cumulative eroded volume is 23, 36.5, and 75 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively. The peak discharge decreases by 7, 14, and 17.35 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively. The relative eroded crest height decreases linearly with increasing the tailwater depth by 10, 41, and 77.6 % for relative tailwater depth, do /(do)ref = 3, 4, and 5, respectively. The numerical model can reproduce the erosion process with a minor deviation from the experimental results, particularly in terms of tracking the degradation of the crest height with time.

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.

Reference

[1]

D. McCullough

The Johnstown Flood

Simon and Schuster, NY (1968)

Google Scholar[2]Rose AT. The influence of dam failures on dam safety laws in Pennsylvania. Association of State Dam Safety Officials Annual Conference 2013, Dam Safety 2013. 2013;1:738–56.

Google Scholar[3]

M. Foster, R. Fell, M. Spannagle

The statistics of embankment dam failures and accidents

Can Geotech J, 37 (5) (2000), pp. 1000-1024, 10.1139/t00-030 View PDF

View Record in ScopusGoogle Scholar[4]Pickert, G., Jirka, G., Bieberstein, A., Brauns, J. Soil/water interaction during the breaching process of overtopped embankments. In: Greco, M., Carravetta, A., Morte, R.D. (Eds.), Proceedings of the Conference River-Flow 2004, Balkema.

Google Scholar[5]

A. Asghari Tabrizi, E. Elalfy, M. Elkholy, M.H. Chaudhry, J. Imran

Effects of compaction on embankment breach due to overtopping

J Hydraul Res, 55 (2) (2017), pp. 236-247, 10.1080/00221686.2016.1238014 View PDF

View Record in ScopusGoogle Scholar[6]

R.M. Kansoh, M. Elkholy, G. Abo-Zaid

Effect of Shape Parameters on Failure of Earthen Embankment due to Overtopping

KSCE J Civ Eng, 24 (5) (2020), pp. 1476-1485, 10.1007/s12205-020-1107-x View PDF

View Record in ScopusGoogle Scholar[7]

YongHui Zhu, P.J. Visser, J.K. Vrijling, GuangQian Wang

Experimental investigation on breaching of embankments

Experimental investigation on breaching of embankments, 54 (1) (2011), pp. 148-155 View PDF

CrossRefView Record in ScopusGoogle Scholar[8]Amaral S, Jónatas R, Bento AM, Palma J, Viseu T, Cardoso R, et al. Failure by overtopping of earth dams. Quantification of the discharge hydrograph. Proceedings of the 3rd IAHR Europe Congress: 14-15 April 2014, Portugal. 2014;(1):182–93.

Google Scholar[9]

G. Bereta, P. Hui, H. Kai, L. Guang, P. Kefan, Y.Z. Zhao

Experimental study of cohesive embankment dam breach formation due to overtopping

Periodica Polytechnica Civil Engineering, 64 (1) (2020), pp. 198-211, 10.3311/PPci.14565 View PDF

View Record in ScopusGoogle Scholar[10]

D.K. Verma, B. Setia, V.K. Arora

Experimental study of breaching of an earthen dam using a fuse plug model

Int J Eng Trans A, 30 (4) (2017), pp. 479-485, 10.5829/idosi.ije.2017.30.04a.04 View PDF

View Record in ScopusGoogle Scholar[11]Wu T, Qin J. Experimental Study of a Tailings Impoundment Dam Failure Due to Overtopping. Mine Water and the Environment [Internet]. 2018;37(2):272–80. Available from: doi: 10.1007/s10230-018-0529-x.

Google Scholar[12]

A. Feizi Khankandi, A. Tahershamsi, S. Soares-Frazo

Experimental investigation of reservoir geometry effect on dam-break flow

J Hydraul Res, 50 (4) (2012), pp. 376-387 View PDF

CrossRefView Record in ScopusGoogle Scholar[13]

A. Ritter

Die Fortpflanzung der Wasserwellen (The propagation of water waves)

Zeitschrift Verein Deutscher Ingenieure, 36 (33) (1892), pp. 947-954

[in German]

View Record in ScopusGoogle Scholar[14]

I. Rifai, K. El Kadi Abderrezzak, S. Erpicum, P. Archambeau, D. Violeau, M. Pirotton, et al.

Floodplain Backwater Effect on Overtopping Induced Fluvial Dike Failure

Water Resour Res, 54 (11) (2018), pp. 9060-9073 View PDF

This article is free to access.

CrossRefView Record in ScopusGoogle Scholar[15]

X. Jiang

Laboratory Experiments on Breaching Characteristics of Natural Dams on Sloping Beds

Advances in Civil Engineering, 2019 (2019), pp. 1-14

View Record in ScopusGoogle Scholar[16]

H. Ozmen-Cagatay, S. Kocaman

Dam-break flows during initial stage using SWE and RANS approaches

J Hydraul Res, 48 (5) (2010), pp. 603-611 View PDF

CrossRefView Record in ScopusGoogle Scholar[17]

S. Evangelista

Experiments and numerical simulations of dike erosion due to a wave impact

Water (Switzerland), 7 (10) (2015), pp. 5831-5848 View PDF

CrossRefView Record in ScopusGoogle Scholar[18]

C. Di Cristo, S. Evangelista, M. Greco, M. Iervolino, A. Leopardi, A. Vacca

Dam-break waves over an erodible embankment: experiments and simulations

J Hydraul Res, 56 (2) (2018), pp. 196-210 View PDF

CrossRefView Record in ScopusGoogle Scholar[19]Goharnejad H, Sm M, Zn M, Sadeghi L, Abadi K. Numerical Modeling and Evaluation of Embankment Dam Break Phenomenon (Case Study : Taleghan Dam) ISSN : 2319-9873. 2016;5(3):104–11.

Google Scholar[20]Hu H, Zhang J, Li T. Dam-Break Flows : Comparison between Flow-3D , MIKE 3 FM , and Analytical Solutions with Experimental Data. 2018;1–24. doi: 10.3390/app8122456.

Google Scholar[21]

R. Kaurav, P.K. Mohapatra, D. Ph

Studying the Peak Discharge through a Planar Dam Breach, 145 (6) (2019), pp. 1-8 View PDF

CrossRef[22]

Z.M. Yusof, Z.A.L. Shirling, A.K.A. Wahab, Z. Ismail, S. Amerudin

A hydrodynamic model of an embankment breaching due to overtopping flow using FLOW-3D

IOP Conference Series: Earth and Environmental Science, 920 (1) (2021)

Google Scholar[23]

G. Pickert, V. Weitbrecht, A. Bieberstein

Breaching of overtopped river embankments controlled by apparent cohesion

J Hydraul Res, 49 (2) (Apr. 2011), pp. 143-156, 10.1080/00221686.2011.552468 View PDF

View Record in ScopusGoogle Scholar

Cited by (0)

My name is Shaimaa Ibrahim Mohamed Aman and I am a teaching assistant in Irrigation and Hydraulics department, Faculty of Engineering, Alexandria University. I graduated from the Faculty of Engineering, Alexandria University in 2013. I had my MSc in Irrigation and Hydraulic Engineering in 2017. My research interests lie in the area of earth dam Failures.

Peer review under responsibility of Ain Shams University.

© 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Ain Shams University.

Fig. 8. Variation of water surface profile (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.

Numerical study of the dam-break waves and Favre waves down sloped wet rigid-bed at laboratory scale

WenjunLiuaBoWangaYakunGuobaState Key Laboratory of Hydraulics and Mountain River Engineering, College of Water Resource and Hydropower, Sichuan University, Chengdu 610065, ChinabFaculty of Engineering & Informatics, University of Bradford, BD7 1DP, UK

Highlights

경사진 습윤층에서 댐파괴유동과 FFavre 파를 수치적으로 조사하였다.
수직 대 수평 속도의 비율이 먼저 정량화됩니다.
유동 상태는 유상 경사가 큰 후기 단계에서 크게 변경됩니다.
Favre 파도는 수직 속도와 수직 가속도에 큰 영향을 미칩니다.
베드 전단응력의 변화는 베드 기울기와 꼬리물의 영향을 받습니다.

Abstract

The bed slope and the tailwater depth are two important ones among the factors that affect the propagation of the dam-break flood and Favre waves. Most previous studies have only focused on the macroscopic characteristics of the dam-break flows or Favre waves under the condition of horizontal bed, rather than the internal movement characteristics in sloped channel. The present study applies two numerical models, namely, large eddy simulation (LES) and shallow water equations (SWEs) models embedded in the CFD software package FLOW-3D to analyze the internal movement characteristics of the dam-break flows and Favre waves, such as water level, the velocity distribution, the fluid particles acceleration and the bed shear stress, under the different bed slopes and water depth ratios. The results under the conditions considered in this study show that there is a flow state transition in the flow evolution for the steep bed slope even in water depth ratio α = 0.1 (α is the ratio of the tailwater depth to the reservoir water depth). The flow state transition shows that the wavefront changes from a breaking state to undular. Such flow transition is not observed for the horizontal slope and mild bed slope. The existence of the Favre waves leads to a significant increase of the vertical velocity and the vertical acceleration. In this situation, the SWEs model has poor prediction. Analysis reveals that the variation of the maximum bed shear stress is affected by both the bed slope and tailwater depth. Under the same bed slope (e.g., S0 = 0.02), the maximum bed shear stress position develops downstream of the dam when α = 0.1, while it develops towards the end of the reservoir when α = 0.7. For the same water depth ratio (e.g., α = 0.7), the maximum bed shear stress position always locates within the reservoir at S0 = 0.02, while it appears in the downstream of the dam for S0 = 0 and 0.003 after the flow evolves for a while. The comparison between the numerical simulation and experimental measurements shows that the LES model can predict the internal movement characteristics with satisfactory accuracy. This study improves the understanding of the effect of both the bed slope and the tailwater depth on the internal movement characteristics of the dam-break flows and Favre waves, which also provides a valuable reference for determining the flood embankment height and designing the channel bed anti-scouring facility.

Fig. 1. Sketch of related variables involved in shallow water model.
Fig. 1. Sketch of related variables involved in shallow water model.
Fig. 2. Flume model in numerical simulation.
Fig. 2. Flume model in numerical simulation.
Fig. 3. Grid sensitivity analysis (a) water surface profile; (b) velocity profile.
Fig. 3. Grid sensitivity analysis (a) water surface profile; (b) velocity profile.
Fig. 4. Sketch of experimental set-up for validating the velocity profile.
Fig. 4. Sketch of experimental set-up for validating the velocity profile.
Fig. 5. Sketch of experimental set-up for validating the bed shear stress.
Fig. 5. Sketch of experimental set-up for validating the bed shear stress.
Fig. 6. Model validation results (a) variation of the velocity profile; (b) error value of the velocity profile; (c) variation of the bed shear stress; (d) error value of the bed shear stress.
Fig. 6. Model validation results (a) variation of the velocity profile; (b) error value of the velocity profile; (c) variation of the bed shear stress; (d) error value of the bed shear stress.
Fig. 7. Schematic diagram of regional division.
Fig. 7. Schematic diagram of regional division.
Fig. 8. Variation of water surface profile (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 8. Variation of water surface profile (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 8. (continued).
Fig. 8. (continued).
Fig. 8. (continued).
Fig. 8. (continued).
Fig. 8. (continued).
Fig. 8. (continued).
Fig. 9. Froude number for α = 0.1 (a) variation with time; (b) variation with wavefront position.
Fig. 9. Froude number for α = 0.1 (a) variation with time; (b) variation with wavefront position.
Fig. 10. Characteristics of velocity distribution (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 10. Characteristics of velocity distribution (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 11. Average proportion of the vertical velocity (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 11. Average proportion of the vertical velocity (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 12. Bed shear stress distribution (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 12. Bed shear stress distribution (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 12. (continued).
Fig. 12. (continued).
Fig. 13. Variation of the maximum bed shear stress position with time (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 13. Variation of the maximum bed shear stress position with time (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 14. Time when the maximum bed shear stress appears at different positions (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 14. Time when the maximum bed shear stress appears at different positions (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 15. Movement characteristics of the fluid particles (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 15. Movement characteristics of the fluid particles (a) α = 0.1; (b) α = 0.3; (c) α = 0.5; (d) α = 0.7.
Fig. 15. (continued).
Fig. 15. (continued).

Keywords

Dam-break flow, Bed slope, Wet bed, Velocity profile, Bed shear stress, Large eddy simulation

References

Barnes, M.P., Baldock, T.E. 2006. Bed shear stress measurements in dam break and swash
flows. Proceedings of International Conference on Civil and Environmental
Engineering. Hiroshima University, Japan, 28–29 September.
Biscarini, C., Francesco, S.D., Manciola, P., 2010. CFD modelling approach for dam break
flow studies. Hydrol. Earth Syst. Sc. 14, 705–718. https://doi.org/10.5194/hess-14-
705-2010.
Fig. 15. (continued).
W. Liu et al.
Journal of Hydrology 602 (2021) 126752
19
Bristeau, M.-O., Goutal, N., Sainte-Marie, J., 2011. Numerical simulations of a nonhydrostatic shallow water model. Comput. Fluids. 47 (1), 51–64. https://doi.org/
10.1016/j.compfluid.2011.02.013.
Bung, D.B., Hildebrandt, A., Oertel, M., Schlenkhoff, A., Schlurmann, T. 2008. Bore
propagation over a submerged horizontal plate by physical and numerical
simulation. Proc. 31st Intl.Conf. Coastal Eng., Hamburg, Germany, 3542–3553.
Cantero-Chinchilla, F.N., Castro-Orgaz, O., Dey, S., Ayuso, J.L., 2016. Nonhydrostatic
dam break flows. I: physical equations and numerical schemes. J. Hydraul. Eng. 142
(12), 04016068. https://doi.org/10.1061/(ASCE)HY.1943-7900.0001205.
Castro-Orgaz, O., Chanson, H., 2020. Undular and broken surges in dam-break flows: A
review of wave breaking strategies in a boussinesq-type framework. Environ. Fluid
Mech. 154 https://doi.org/10.1007/s10652-020-09749-3.
Chang, T.-J., Kao, H.-M., Chang, K.-H., Hsu, M.-H., 2011. Numerical simulation of
shallow-water dam break flows in open channels using smoothed particle
hydrodynamics. J. Hydrol. 408 (1-2), 78–90. https://doi.org/10.1016/j.
jhydrol.2011.07.023.
Chen, H., Xu, W., Deng, J., Xue, Y., Li, J., 2009. Experimental investigation of pressure
load exerted on a downstream dam by dam-break flow. J. Hydraul. Eng. 140,
199–207. https://doi.org/10.1061/(ASCE)HY.1943-7900.0000743.
Favre H. 1935. Etude th´eorique et exp´erimentale des ondes de translation dans les
canaux d´ecouverts. Dunod, Paris. (in French).
Flow Science Inc. 2016. Flow-3D User’s Manuals. Santa Fe NM.
Fraccarollo, L., Toro, E.F., 1995. Experimental and numerical assessment of the shallow
water model for two-dimensional dam-break type problems. J. Hydraul. Res. 33 (6),
843–864. https://doi.org/10.1080/00221689509498555.
Guo, Y., Wu, X., Pan, C., Zhang, J., 2012. Numerical simulation of the tidal flow and
suspended sediment transport in the qiantang estuary. J Waterw. Port Coastal. 138
(3), 192–202. https://doi.org/10.1061/(ASCE)WW.1943-5460.0000118.
Guo, Y., Zhang, Z., Shi, B., 2014. Numerical simulation of gravity current descending a
slope into a linearly stratified environment. J. Hydraulic Eng. 140 (12), 04014061.
https://doi.org/10.1061/(ASCE)HY.1943-7900.0000936.
Khosronejad, A., Kang, S., Flora, K., 2019. Fully coupled free-surface flow and sediment
transport modelling of flash floods in a desert stream in the mojave desert, california.
Hydrol. Process 33 (21), 2772–2791. https://doi.org/10.1002/hyp.v33.2110.1002/
hyp.13527.
Khosronejad, A., Arabi, M.G., Angelidis, D., Bagherizadeh, E., Flora, K., Farhadzadeh, A.,
2020a. A comparative study of rigid-lid and level-set methods for LES of openchannel flows: morphodynamics. Environ. Fluid Mech. 20 (1), 145–164. https://doi.
org/10.1007/s10652-019-09703-y.
Khosronejad, A., Flora, K., Zhang, Z.X., Kang, S., 2020b. Large-eddy simulation of flash
flood propagation and sediment transport in a dry-bed desert stream. Int. J.
Sediment Res. 35 (6), 576–586. https://doi.org/10.1016/j.ijsrc.2020.02.002.
Khoshkonesh, A., Nsom, B., Gohari, S., Banejad, H., 2019. A comprehensive study of dam
break over the dry and wet beds. Ocean Eng. 188, 106279.1–106279.18. https://doi.
org/10.1016/j.oceaneng.2019.106279.
Kocaman, S., Ozmen-Cagatay, H., 2012. The effect of lateral channel contraction on dam
break flows: laboratory experiment. J. Hydrol. 432–433, 145–153. https://doi.org/
10.1016/j.jhydrol.2012.02.035.
Kocaman, S., Ozmen-Cagatay, H., 2015. Investigation of dam-break induced shock waves
impact on a vertical wall. J. Hydrol. 525, 1–12. https://doi.org/10.1016/j.
jhydrol.2015.03.040.
LaRocque, L.A., Imran, J., Chaudhry, M.H., 2013a. Experimental and numerical
investigations of two-dimensional dam-break flows. J. Hydraul. Eng. 139 (6),
569–579. https://doi.org/10.1061/(ASCE)HY.1943-7900.0000705.
Larocque, L.A., Imran, J., Chaudhry, M.H., 2013b. 3D numerical simulation of partial
breach dam-break flow using the LES and k-ε turbulence models. J. Hydraul. Res. 51,
145–157. https://doi.org/10.1080/00221686.2012.734862.
Lauber, G., Hager, W.H., 1998a. Experiments to dam break wave: Horizontal channel.
J. Hydraul. Res. 36 (3), 291–307. https://doi.org/10.1080/00221689809498620.
Lauber, G., Hager, W.H., 1998b. Experiments to dam break wave: Sloping channel.
J. Hydraul. Res. 36 (5), 761–773. https://doi.org/10.1080/00221689809498601.
Leal, J.G., Ferreira, R.M., Cardoso, A.H., 2006. Dam-break wave-front celerity.
J. Hydraul. Eng. 132 (1), 69–76. https://doi.org/10.1061/(ASCE)0733-9429(2006)
132:1(69).
Liu, W., Wang, B., Guo, Y., Zhang, J., Chen, Y., 2020. Experimental investigation on the
effects of bed slope and tailwater on dam-break flows. J. Hydrol. 590, 125256.
https://doi.org/10.1016/j.jhydrol.2020.125256.
Marche, C., Beauchemin P. EL Kayloubi, A. 1995. Etude num´erique et exp´erimentale des
ondes secondaires de Favre cons´ecutives a la rupture d’un harrage. Can. J. Civil Eng.
22, 793–801, (in French). https://doi.org/10.1139/l95-089.
Marra, D., Earl, T., Ancey, C. 2011. Experimental investigations of dam break flows down
an inclined channel. Proceedings of the 34th World Congress of the International
Association for Hydro-Environment Research and Engineering: 33rd Hydrology and
Water Resources Symposium and 10th Conference on Hydraulics in Water
Engineering, Brisbane, Australia.
Marsooli, R., Wu, W., 2014. 3-D finite-volume model of dam-break flow over uneven
beds based on vof method. Adv. Water Resour. 70, 104–117. https://doi.org/
10.1016/j.advwatres.2014.04.020.
Miller, S., Chaudhry, M.H., 1989. Dam-break flows in curved channel. J. Hydraul. Eng.
115 (11), 1465–1478. https://doi.org/10.1061/(ASCE)0733-9429(1989)115:11
(1465).
Mohapatra, P.K., Chaudhry, M.H., 2004. Numerical solution of Boussinesq equations to
simulate dam-break flows. J. Hydraul. Eng. 130 (2), 156–159. https://doi.org/
10.1061/(ASCE)0733-9429(2004)130:2(156).
Oertel, M., Bung, D.B., 2012. Initial stage of two-dimensional dam-break waves:
laboratory versus VOF. J. Hydraul. Res. 50 (1), 89–97. https://doi.org/10.1080/
00221686.2011.639981.
Ozmen-Cagatay, H., Kocaman, S., 2012. Investigation of dam-break flow over abruptly
contracting channel with trapezoidal-shaped lateral obstacles. J. Fluids Eng. 134,
081204 https://doi.org/10.1115/1.4007154.
Ozmen-Cagatay, H., Kocaman, S., Guzel, H., 2014. Investigation of dam-break flood
waves in a dry channel with a hump. J. Hydro-environ. Res. 8 (3), 304–315. https://
doi.org/10.1016/j.jher.2014.01.005.
Park, I.R., Kim, K.S., Kim, J., Van, S.H., 2012. Numerical investigation of the effects of
turbulence intensity on dam-break flows. Ocean Eng. 42, 176–187. https://doi.org/
10.1016/j.oceaneng.2012.01.005.
Peregrine, D.H., 1966. Calculations of the development of an undular bore. J. Fluid
Mech. 25 (2), 321–330. https://doi.org/10.1017/S0022112066001678.
Savic, L.j., Holly, F.M., 1993. Dam break flood waves computed by modified Godunov
method. J. Hydraul. Res. 31 (2), 187–204. https://doi.org/10.1080/
00221689309498844.
Shigematsu, T., Liu, P., Oda, K., 2004. Numerical modeling of the initial stages of dambreak waves. J. Hydraul. Res. 42 (2), 183–195. https://doi.org/10.1080/
00221686.2004.9628303.
Smagorinsky, J., 1963. General circulation experiments with the primitive equations.
Part I: the basic experiment. Mon. Weather Rev. 91, 99–164. https://doi.org/
10.1126/science.27.693.594.
Soares-Frazao, S., Zech, Y., 2002. Undular bores and secondary waves – Experiments and
hybrid finite-volume modeling. J. Hydraul. Res. 40, 33–43. https://doi.org/
10.1080/00221680209499871.
Stansby, P.K., Chegini, A., Barnes, T.C.D., 1998. The initial stages of dam-break flow.
J. Fluid Mech. 370, 203–220. https://doi.org/10.1017/022112098001918.
Treske, A., 1994. Undular bores (favre-waves) in open channels – experimental studies.
J. Hydraul. Res. 32 (3), 355–370. https://doi.org/10.1080/00221689409498738.
Wang, B., Chen, Y., Wu, C., Dong, J., Ma, X., Song, J., 2016. A semi-analytical approach
for predicting peak discharge of floods caused by embankment dam failures. Hydrol.
Process 30 (20), 3682–3691. https://doi.org/10.1002/hyp.v30.2010.1002/
hyp.10896.
Wang, B., Chen, Y., Wu, C., Peng, Y., Ma, X., Song, J., 2017. Analytical solution of dambreak flood wave propagation in a dry sloped channel with an irregular-shaped
cross-section. J. Hydro-environ. Res. 14, 93–104. https://doi.org/10.1016/j.
jher.2016.11.003.
Wang, B., Chen, Y., Wu, C., Peng, Y., Song, J., Liu, W., Liu, X., 2018. Empirical and semianalytical models for predicting peak outflows caused by embankment dam failures.
J. Hydrol. 562, 692–702. https://doi.org/10.1016/j.jhydrol.2018.05.049.
Wang, B., Zhang, J., Chen, Y., Peng, Y., Liu, X., Liu, W., 2019. Comparison of measured
dam-break flood waves in triangular and rectangular channels. J. Hydrol. 575,
690–703. https://doi.org/10.1016/j.jhydrol.2019.05.081.
Wang, B., Liu, W., Zhang, J., Chen, Y., Wu, C., Peng, Y., Wu, Z., Liu, X., Yang, S., 2020a.
Enhancement of semi-theoretical models for predicting peak discharges in breached
embankment dams. Environ. Fluid Mech. 20 (4), 885–904. https://doi.org/10.1007/
s10652-019-09730-9.
Wang, B., Chen, Y., Peng, Y., Zhang, J., Guo, Y., 2020b. Analytical solution of shallow
water equations for ideal dam-break flood along a wet bed slope. J. Hydraul. Eng.
146 (2), 06019020. https://doi.org/10.1061/(ASCE)HY.1943-7900.0001683.
Wang, B., Liu, W., Wang, W., Zhang, J., Chen, Y., Peng, Y., Liu, X., Yang, S., 2020c.
Experimental and numerical investigations of similarity for dam-break flows on wet
bed. J. Hydrol. 583, 124598. https://doi.org/10.1016/j.jhydrol.2020.124598.
Wang, B., Liu, X., Zhang, J., Guo, Y., Chen, Y., Peng, Y., Liu, W., Yang, S., Zhang, F.,
2020d. Analytical and experimental investigations of dam-break flows in triangular
channels with wet-bed conditions. J. Hydraul. Eng. 146 (10), 04020070. https://doi.
org/10.1061/(ASCE)HY.1943-7900.0001808.
Wu, W., Wang, S., 2007. One-dimensional modeling of dam-break flow over movable
beds. J. Hydraul. Eng. 133 (1), 48–58. https://doi.org/10.1061/(ASCE)0733-9429
(2007)133:1(48).
Xia, J., Lin, B., Falconer, R.A., Wang, G., 2010. Modelling dam-break flows over mobile
beds using a 2d coupled approach. Adv. Water Resour. 33 (2), 171–183. https://doi.
org/10.1016/j.advwatres.2009.11.004.
Yang, S., Yang, W., Qin, S., Li, Q., Yang, B., 2018a. Numerical study on characteristics of
dam-break wave. Ocean Eng. 159, 358–371. https://doi.org/10.1016/j.
oceaneng.2018.04.011.
Yang, S., Yang, W., Qin, S., Li, Q., 2018b. Comparative study on calculation methods of
dam-break wave. J. Hydraul. Res. 57 (5), 702–714. https://doi.org/10.1080/
00221686.2018.1494057.

Fig. 6. Experiment of waves passing through a single block of porous medium.

Generalization of a three-layer model for wave attenuation in n-block submerged porous breakwater

NadhiraKarimaaIkhaMagdalenaabIndrianaMarcelaaMohammadFaridbaFaculty of Mathematics and Natural Sciences, Bandung Institute of Technology, 40132, IndonesiabCenter for Coastal and Marine Development, Bandung Institute of Technology, Indonesia

Highlights

•A new three-layer model for n-block submerged porous breakwaters is developed.

•New analytical approach in finding the wave transmission coefficient is presented.

•A finite volume method successfully simulates the wave attenuation process.

•Porous media blocks characteristics and configuration can optimize wave reduction.

Abstract

높은 파도 진폭은 해안선에 위험한 영향을 미치고 해안 복원력을 약화시킬 수 있습니다. 그러나 다중 다공성 매체는 해양 생태계의 환경 친화적인 해안 보호 역할을 할 수 있습니다.

이 논문에서 우리는 n개의 잠긴 다공성 미디어 블록이 있는 영역에서 파동 진폭 감소를 계산하기 위해 3층 깊이 통합 방정식을 사용합니다. 수학적 모델은 파동 전달 계수를 얻기 위해 여러 행렬 방정식을 포함하는 변수 분리 방법을 사용하여 해석적으로 해결됩니다.

이 계수는 진폭 감소의 크기에 대한 정보를 제공합니다. 또한 모델을 수치적으로 풀기 위해 지그재그 유한 체적 방법이 적용됩니다.

수치 시뮬레이션을 통해 다공성 매질 블록의 구성과 특성이 투과파 진폭을 줄이는 데 중요하다는 결론을 내렸습니다.

High wave amplitudes may cause dangerous effects on the shoreline and weaken coastal resilience. However, multiple porous media can act as environmental friendly coastal protectors of the marine ecosystem. In this paper, we use three-layer depth-integrated equations to calculate wave amplitude reduction in a domain with n submerged porous media blocks. The mathematical model is solved analytically using the separation of variables method involving several matrix equations to obtain the wave transmission coefficient. This coefficient provides information about the magnitude of amplitude reduction. Additionally, a staggered finite volume method is applied to solve the model numerically. By conducting numerical simulations, we conclude that porous media blocks’ configuration and characteristics are crucial in reducing transmitted wave amplitude.

Keywords

Three-layer equations, Submerged porous media, Wave transmission coefficient, Finite volume method

Fig. 1. Sketch of the problem configuration.
Fig. 1. Sketch of the problem configuration.
Fig. 6. Experiment of waves passing through a single block of porous medium.
Fig. 6. Experiment of waves passing through a single block of porous medium.

References

[1]M. Beck, G. Lange, Managing Coasts with Natural Solutions: Guidelines for Measuring and Valuing the Coastal Protection Services of Mangroves and Coral Reefs.

Google Scholar[2]

Y. Zhao, Y. Liu, H. Li, A. Chang

Oblique wave motion over multiple submerged porous bars near a vertical wall

J. Ocean Univ. China, 16 (2017), pp. 568-574, 10.1007/s11802-017-3333-5 View PDF

View Record in ScopusGoogle Scholar[3]C. K. Sollitt, R. H. Cross, Wave transmission through permeable breakwaters, Coast. Eng..

Google Scholar[4]J.-F. Lee, L.-F. Tu, C.-C. Liu, Nonlinear wave evolution above rectangular submerged structures, J. Mar. Sci. Technol. 22. doi:10.6119/JMST-013-0503-3.

Google Scholar[5]

Y.T. Wu, C.L. Yeh, S.-C. Hsiao

Three-dimensional numerical simulation on the interaction of solitary waves and porous breakwaters

Coast. Eng., 85 (2014), pp. 12-29

ArticleDownload PDFView Record in ScopusGoogle Scholar[6]

Y. feng Xu, X. he Xia, J. hua Wang, J. jian Chen

Numerical analysis on cnoidal wave induced response of porous seabed with definite thickness

J. Shanghai Jiao Tong Univ. (Sci.), 18 (2013), pp. 650-654, 10.1007/s12204-013-1446-6 View PDF

Google Scholar[7]

D.M. Pérez-Romero, M. Ortega-Sánchez, A. Moñino, M.A. Losada

Characteristic friction coefficient and scale effects in oscillatory porous flow

Coast. Eng., 56 (9) (2009), pp. 931-939, 10.1016/j.coastaleng.2009.05.002

ArticleDownload PDFView Record in ScopusGoogle Scholar[8]

A. Torres-Freyermuth, M. Brocchini, S. Corvaro, J.C. Pintado-Patiño

Wave attenuation over porous seabeds: a numerical study

Ocean Model., 117 (2017), pp. 28-40, 10.1016/j.ocemod.2017.07.004

ArticleDownload PDFView Record in ScopusGoogle Scholar[9]F. Hajivalie, S. M. Mahmoudof, Experimental study of energy dissipation at rectangular submerged breakwater, Proceedings of the 8th International Conference on Fluid Mechanics.

Google Scholar[10]G. T. Klonaris, A. S. Metallinos, C. D. Memos, K. A. Galani, Experimental and numerical investigation of bed morphology in the lee of porous submerged breakwaters, Coast. Eng. 155.

Google Scholar[11]

A. Kubowicz-Grajewska

Experimental investigation into wave interaction with a rubble-mound submerged breakwater (case study)

J. Mar. Sci. Technol., 22 (2) (2017), pp. 313-326 View PDF

CrossRefView Record in ScopusGoogle Scholar[12]

S.M. Mahmoudof, F. Hajivalie

Experimental study of hydraulic response of smooth submerged breakwaters to irregular waves

Oceanologia, 63 (4) (2021), pp. 448-462

ArticleDownload PDFView Record in ScopusGoogle Scholar[13]

C. Tsai, H. Chen, F. Lee

Wave transformation over submerged permeable breakwater on porous bottom

Ocean Eng., 33 (2006), pp. 1623-1643, 10.1016/j.oceaneng.2005.09.006

ArticleDownload PDFView Record in ScopusGoogle Scholar[14]

S. Rojanakamthorn, M. Isobe, A. Watanabe

A mathematical model of wave transformation over a submerged breakwater

Coastal Engineering in Japan, 31 (1989), pp. 209-234, 10.1080/05785634.1989.11924515 View PDF

View Record in ScopusGoogle Scholar[15]

Q. Lin, Q.r. Meng, D.q. Lu

Waves propagating over a two-layer porous barrier on a seabed

J. Hydrodyn., 30 (3) (2018), pp. 453-462 View PDF

CrossRefView Record in ScopusGoogle Scholar[16]X. Yu, A. T. Chwang, Wave motion through porous structures, J. Eng. Mech. 120. doi:10.1061/(ASCE)0733-9399(1994)120:5(989).

Google Scholar[17]

K.G. Vijay, V. Venkateswarlu, D. Karmakar

Scattering of gravity waves by multiple submerged rubble-mound breakwaters

Arabian J. Sci. Eng., 45 (10) (2020), pp. 8529-8550 View PDF

CrossRefView Record in ScopusGoogle Scholar[18]

I. Magdalena, G. Jonathan

Water waves resonance and its interaction with submerged breakwater

Results in Engineering, 13 (2022), Article 100343, 10.1016/j.rineng.2022.100343

ArticleDownload PDFView Record in ScopusGoogle Scholar[19]

I. Magdalena, K. Firdaus, D. Jayadi

Analytical and numerical studies for wave generated by submarine landslide

Alex. Eng. J., 61 (9) (2022), pp. 7303-7313, 10.1016/j.aej.2021.12.069

ArticleDownload PDFView Record in ScopusGoogle Scholar[20]

L. Arpaia, M. Ricchiuto, A.G. Filippini, R. Pedreros

An efficient covariant frame for the spherical shallow water equations: well balanced dg approximation and application to tsunami and storm surge

Ocean Model., 169 (2022), Article 101915, 10.1016/j.ocemod.2021.101915

ArticleDownload PDFView Record in ScopusGoogle Scholar[21]

M. Briani, G. Puppo, M. Ribot

Angle dependence in coupling conditions for shallow water equations at channel junctions

Comput. Math. Appl., 108 (2022), pp. 49-65, 10.1016/j.camwa.2021.12.021

ArticleDownload PDFView Record in ScopusGoogle Scholar[22]

I. Magdalena, G.R. Andadari, D.E. Reeve

An integrated study of wave attenuation by vegetation

Wave Motion, 110 (2022), Article 102878, 10.1016/j.wavemoti.2021.102878

ArticleDownload PDFView Record in ScopusGoogle Scholar[23]

I. Magdalena, R. La’lang, R. Mendoza

Quantification of wave attenuation in mangroves in manila bay using nonlinear shallow water equations

Results in Applied Mathematics, 12 (2021), Article 100191, 10.1016/j.rinam.2021.100191

ArticleDownload PDFView Record in ScopusGoogle Scholar[24]

K.T. Mandli

A numerical method for the two layer shallow water equations with dry states

Ocean Model., 72 (2013), pp. 80-91, 10.1016/j.ocemod.2013.08.001

ArticleDownload PDFView Record in ScopusGoogle Scholar[25]

M. Farhan, Z. Omar, F. Mebarek-Oudina, J. Raza, Z. Shah, R.V. Choudhari, O.D. Makinde

Implementation of the one-step one-hybrid block method on the nonlinear equation of a circular sector oscillator

Comput. Math. Model., 31 (2020), pp. 116-132, 10.1007/s10598-020-09480-0 View PDF

View Record in ScopusGoogle Scholar[26]

R. Djebali, F. Mebarek-Oudina, C. Rajashekhar

Similarity solution analysis of dynamic and thermal boundary layers: further formulation along a vertical flat plate

Phys. Scripta, 96 (8) (2021), Article 085206, 10.1088/1402-4896/abfe31 View PDF

View Record in ScopusGoogle Scholar[27]

M. Alkasassbeh, O. Zurni, F. Mebarek-Oudina, J. Raza

Heat transfer study of convective fin with temperature,Äêdependent internal heat generation by hybrid block method

Heat Tran. Asian Res., 48 (2019), pp. 1225-1244, 10.1002/htj.21428 View PDF

View Record in ScopusGoogle Scholar[28]

I. Magdalena, M.F. Eka Pebriansyah

Numerical treatment of finite difference method for solving dam break model on a wet-dry bed with an obstacle

Results in Engineering, 14 (2022), Article 100382, 10.1016/j.rineng.2022.100382

ArticleDownload PDFView Record in ScopusGoogle Scholar[29]

M. Uddin, S. Rasel, J.K. Adewole, K.S. Al Kalbani

Finite element simulation on the convective double diffusive water-based copper oxide nanofluid flow in a square cavity having vertical wavy surfaces in presence of hydro-magnetic field

Results in Engineering, 13 (2022), Article 100364, 10.1016/j.rineng.2022.100364

ArticleDownload PDFView Record in ScopusGoogle Scholar[30]

E.H.H. Al-Qadami, A.S. Abdurrasheed, Z. Mustaffa, K.W. Yusof, M. Malek, A.A. Ghani

Numerical modelling of flow characteristics over sharp crested triangular hump

Results in Engineering, 4 (2019), Article 100052, 10.1016/j.rineng.2019.100052

ArticleDownload PDFView Record in ScopusGoogle Scholar[31]

I. Magdalena, V. Kusnowo, M.I. Azis

Widowati, 1d-2d numerical model for wave attenuation by mangroves as a porous structure

Computation, 9 (6) (2021), pp. 1-21

Google Scholar[32]

I. Magdalena, M.F. Atras, L. Sembiring, M.A. Nugroho, R.S.B. Labay, M.P. Roque

Wave transmission by rectangular submerged breakwaters

Computation, 8 (2) (2020), pp. 1-18 View PDF

View Record in ScopusGoogle Scholar[33]

I. Magdalena, S.R. Pudjaprasetya

Numerical modeling for gravity waves over submerged porous media

Australian Journal of Basic and Applied Sciences, 9 (28) (2015), pp. 124-130

View Record in ScopusGoogle Scholar[34]

I. Magdalena, A. Hariz, M. Farid, M.S.B. Kusuma

Numerical studies using staggered finite volume for dam break flow with an obstacle through different geometries

Results in Applied Mathematics, 12 (2021), Article 100193, 10.1016/j.rinam.2021.100193

ArticleDownload PDFView Record in ScopusGoogle Scholar[35]

R. Walters, E. Hanert, J. Pietrzak, D. Le Roux

Comparison of unstructured, staggered grid methods for the shallow water equations

Ocean Model., 28 (1) (2009), pp. 106-117, 10.1016/j.ocemod.2008.12.004

the Sixth International Workshop on Unstructured Mesh Numerical Modelling of Coastal, Shelf and Ocean Flows

ArticleDownload PDFView Record in ScopusGoogle Scholar[36]

F. Mebarek-Oudina

Numerical modeling of the hydrodynamic stability in vertical annulus with heat source of different lengths, Engineering Science and Technology

Int. J., 20 (4) (2017), pp. 1324-1333, 10.1016/j.jestch.2017.08.003

ArticleDownload PDFView Record in ScopusGoogle Scholar[37]

S. Pudjaprasetya, I. Magdalena

Numerical modeling for gravity waves over submerged porous media

Australian Journal of Basic and Applied Sciences, 9 (2015), pp. 124-130

Google Scholar

Fig. 9 Test facility a plan view, b the bottom elevation of the reservoir to the channel (Bell et al. 1992)

2-D Dam-Break Flow Modeling Based on Weighted Average Flux Method

Iranian Journal of Science and Technology, Transactions of Civil Engineering volume 46, pages1515–1525 (2022)Cite this article

Abstract

천해 방정식을 기반으로 하는 2차원 흐름 모델은 댐 붕괴 흐름을 모델링하기 위해 개발되었습니다. 공간 이산화는 유한 체적 셀 중심 유형 방법에 의해 얻어집니다.

수치 시스템은 명시적인 방식으로 해결됩니다. 플럭스 모델링은 시간과 공간 모두에서 2차 정확도로 TVD WAF 방식으로 배포되었습니다. 로컬 리만 문제는 셀 인터페이스에서 HLLC 방법으로 해결됩니다. 수치 모델은 모델 결과와 해석 솔루션을 비교하여 검증합니다.

그런 다음 수치 모델의 결과는 90° 및 180° 편차 각도를 갖는 수로 및 삼각형 바텀 씰 위의 직선 수로에서 사용 가능한 실험 데이터와 비교됩니다. 결과는 댐 파괴파를 예측하는 현재 모델의 합리적인 성능을 확인합니다.

A two-dimensional flow model based on shallow water equations is developed for modeling dam-break flows. The spatial discretization is obtained by the finite volume cell centered type method. The numerical system is solved in explicit way. The flux modeling has been deployed by TVD WAF scheme with a second-order accuracy in both time and space. The local Riemann problem is solved by the HLLC method in the interface of the cells. The numerical model is verified by comparison of model results and analytical solutions. Then the results of numerical model are compared with available experimental data of dam-break waves in a channel with 90° and 180° deviation angle and in a straight channel over a triangular bottom sill. The results confirm the reasonable performance of the present model in predicting dam-break waves.

This is a preview of subscription content, access via your institution.

Keywords

  • Finite volume
  • Shallow water equations
  • Dam-break
  • HLLC
  • TVD
  • WAF
Fig. 2 Generic control volume and notations
Fig. 2 Generic control volume and notations
Fig. 1 The generated grid for a channel with a 180° bend
Fig. 1 The generated grid for a channel with a 180° bend
Fig. 4 a Water surface profle and b velocity profle of dam-break problem with left dry bed
Fig. 4 a Water surface profle and b velocity profle of dam-break problem with left dry bed
Fig. 5 a Water surface profle and b velocity profle of appearance dry region
Fig. 5 a Water surface profle and b velocity profle of appearance dry region
Fig. 6 Comparison of the present model results and exact solution for transcritical fow over a bump with a shock
Fig. 6 Comparison of the present model results and exact solution for transcritical fow over a bump with a shock
Fig. 7 Geometry of the reservoir and L-shaped channel: plan view (Soares-Frazao et al. 2019)
Fig. 7 Geometry of the reservoir and L-shaped channel: plan view (Soares-Frazao et al. 2019)
Fig. 9 Test facility a plan view, b the bottom elevation of the reservoir to the channel (Bell et al. 1992)
Fig. 9 Test facility a plan view, b the bottom elevation of the reservoir to the channel (Bell et al. 1992)

References

  • Ata R, Pavan S, Khelladi S, Toro EF (2013) A Weighted Average Flux (WAF) scheme applied to shallow water equations for real-life applications. Adv Water Resour 62:155–172Article Google Scholar 
  • Aureli F, Maranzoni A, Mignosa P, Ziveri C (2008) A weighted surface-depth gradient method for the numerical integration of the 2D shallow water equations with topography. Adv Water Resour 31(7):962–974Article Google Scholar 
  • Bell SW, Elliot RC, Hanif Chaudhry M (1992) Experimental results of two-dimensional dam-break flows. J Hydraul Res 30(2):225–252Article Google Scholar 
  • Benkhaldoun F, Sari S, Seaid M (2015) Projection finite volume method for shallow water flows. Math Comput Simul 118:87–101MathSciNet Article Google Scholar 
  • Chung TJ (2010) Computational fluid dynamics. Cambridge University PressBook Google Scholar 
  • Einfeldt B, Munz CD, Roe PL, Sjögreen B (1991) On Godunov-type methods near low densities. J Comput Phys 92(2):273–295MathSciNet Article Google Scholar 
  • Erduran KS, Kutija V, Hewett CJM (2002) Performance of finite volume solutions to the shallow water equations with shock-capturing schemes. Int J Numer Meth Fluids 40(10):1237–1273MathSciNet Article Google Scholar 
  • Gallardo JM, Schneider KA, Castro MJ (2019) On a class of two-dimensional incomplete Riemann solvers. J Comput Phys 386:541–567MathSciNet Article Google Scholar 
  • Fraccarollo L, Toro EF (1995) Experimental and numerical assessment of the shallow water model for two-dimensional dam-break type problems. J Hydraul Res 33(6):843–864Article Google Scholar 
  • Gupta SC (2019) Numerical grid generation and parallel computations. J Aerosp Eng Technol 5(1):11–22Google Scholar 
  • Hosseinzadeh-Tabrizi A, Ghaeini-Hessaroeyeh M (2015) Coupled dam-break flow and bed load modelling using HLLC-WAF scheme. Water Sci Technol 72(7):1155–1167Article Google Scholar 
  • Kocaman S, Ozmen-Cagatay H (2015) Investigation of dam-break induced shock waves impact on a vertical wall. J Hydrol 525:1–12Article Google Scholar 
  • Liang Q, Borthwick AGL, Stelling G (2004) Simulation of dam-and dyke-break hydrodynamics on dynamically adaptive quadtree grids. Int J Numer Meth Fluids 46(2):127–162Article Google Scholar 
  • Liang D, Lin B, Falconer RA (2007) Simulation of rapidly varying flow using an efficient TVD–MacCormack scheme. Int J Numer Meth Fluids 53(5):811–826Article Google Scholar 
  • Munoz DH, Constantinescu G (2020) 3-D dam break flow simulations in simplified and complex domains. Adv Water Resour 103510
  • Muscat L, Puigt G, Montagnac M, Brenner P (2019) A coupled implicit-explicit time integration method for compressible unsteady flows. J Comput Phys 398:183MathSciNet Article Google Scholar 
  • Ozmen-Cagatay H, Kocaman S (2010) Dam-break flows during initial stage using SWE and RANS approaches. J Hydraul Res 48(5):603–611Article Google Scholar 
  • Ozmen-Cagatay H, Kocaman S, Guzel H (2014) Investigation of dam-break flood waves in a dry channel with a hump. J Hydro Environ Res 8(3):304–315Article Google Scholar 
  • Pongsanguansin T, Maleewong M, Mekchay K (2015) Consistent weighted average flux of well-balanced TVD-RK discontinuous galerkin method for shallow water flows. Modelling and simulation in Engineering, pp. 591282
  • Pongsanguansin T, Maleewong M, Mekchay K (2016) Shallow-water simulations by a well-balanced WAF finite volume method: a case study to the great flood in 2011. Thailand Comput Geosci 20(6):1269–1285MathSciNet Article Google Scholar 
  • Soares-Frazão S (2007) Experiments of dam-break wave over a triangular bottom sill. J Hydraul Res 45(sup1):19–26MathSciNet Article Google Scholar 
  • Soares Frazao S, Abbeels S, Messens M (2019) Dam-break flow in a channel with a 90° bend and mobile bed: experimental and numerical simulations. In: 38th IAHR world congress
  • Toro EF (1992) Riemann problems and the WAF method for solving the two-dimensional shallow water equations. Philos Trans Roy Soc Lond Ser A Phys Eng Sci 338(1649):43–68MathSciNet MATH Google Scholar 
  • Toro EF, Spruce M, Speares W (1994) Restoration of the contact surface in the HLL-Riemann solver. Shock Waves 4(1):25–34Article Google Scholar 
  • Toro EF (2001) Shock-capturing methods for free-surface shallow flows, vol 868. Wiley, New YorkMATH Google Scholar 
  • Toro EF, Garcia-Navarro P (2007) Godunov-type methods for free-surface shallow flows: a review. J Hydraul Res 45(6):736–751Article Google Scholar 
  • Toro, E.F., 2013. Riemann solvers and numerical methods for fluid dynamics: a practical introduction. Springer Science & Business Media.
  • Valiani A, Caleffi V, Zanni A (2002) Case study: Malpasset dam-break simulation using a two-dimensional finite volume method. J Hydraul Eng 128(5):460–472Article Google Scholar 
  • Xia J, Lin B, Falconer RA, Wang G (2010) Modelling dam-break flows over mobile beds using a 2-D coupled approach. Adv Water Resour 33(2):171–183Article Google Scholar 
  • Zia A, Banihashemi MA (2008) Simple efficient algorithm (SEA) for shallow flows with shock wave on dry and irregular beds. Int J Numer Meth Fluids 56(11):2021–2043MathSciNet Article Google Scholar 
  • Zoppou C, Roberts S (2003) Explicit schemes for dam-break simulations. J Hydraul Eng 129(1):11–34Article Google Scholar 
  • Zhou JG, Causon DM, Ingram DM, Mingham CG (2002) Numerical solutions of the shallow water equations with discontinuous bed topography. Int J Numer Meth Fluids 38(8):769–788Article Google Scholar 
Fig. 5. The predicted shapes of initial breach (a) Rectangular (b) V-notch. Fig. 6. Dam breaching stages.

Investigating the peak outflow through a spatial embankment dam breach

공간적 제방댐 붕괴를 통한 최대 유출량 조사

Mahmoud T.GhonimMagdy H.MowafyMohamed N.SalemAshrafJatwaryFaculty of Engineering, Zagazig University, Zagazig 44519, Egypt

Abstract

Investigating the breach outflow hydrograph is an essential task to conduct mitigation plans and flood warnings. In the present study, the spatial dam breach is simulated by using a three-dimensional computational fluid dynamics model, FLOW-3D. The model parameters were adjusted by making a comparison with a previous experimental model. The different parameters (initial breach shape, dimensions, location, and dam slopes) are studied to investigate their effects on dam breaching. The results indicate that these parameters have a significant impact. The maximum erosion rate and peak outflow for the rectangular shape are higher than those for the V-notch by 8.85% and 5%, respectively. Increasing breach width or decreasing depth by 5% leads to increasing maximum erosion rate by 11% and 15%, respectively. Increasing the downstream slope angle by 4° leads to an increase in both peak outflow and maximum erosion rate by 2.0% and 6.0%, respectively.

유출 유출 수문곡선을 조사하는 것은 완화 계획 및 홍수 경보를 수행하는 데 필수적인 작업입니다. 본 연구에서는 3차원 전산유체역학 모델인 FLOW-3D를 사용하여 공간 댐 붕괴를 시뮬레이션합니다. 이전 실험 모델과 비교하여 모델 매개변수를 조정했습니다.

다양한 매개변수(초기 붕괴 형태, 치수, 위치 및 댐 경사)가 댐 붕괴에 미치는 영향을 조사하기 위해 연구됩니다. 결과는 이러한 매개변수가 상당한 영향을 미친다는 것을 나타냅니다. 직사각형 형태의 최대 침식율과 최대 유출량은 V-notch보다 각각 8.85%, 5% 높게 나타났습니다.

위반 폭을 늘리거나 깊이를 5% 줄이면 최대 침식률이 각각 11% 및 15% 증가합니다. 하류 경사각을 4° 증가시키면 최대 유출량과 최대 침식률이 각각 2.0% 및 6.0% 증가합니다.

Keywords

Spatial dam breach; FLOW-3D; Overtopping erosion; Computational fluid dynamics (CFD)

1. Introduction

There are many purposes for dam construction, such as protection from flood disasters, water storage, and power generationEmbankment failures may have a catastrophic impact on lives and infrastructure in the downstream regions. One of the most common causes of embankment dam failure is overtopping. Once the overtopping of the dam begins, the breach formation will start in the dam body then end with the dam failure. This failure occurs within a very short time, which threatens to be very dangerous. Therefore, understanding and modeling the embankment breaching processes is essential for conducting mitigation plans, flood warnings, and forecasting flood damage.

The analysis of the dam breaching process is implemented by different techniques: comparative methods, empirical models with dimensional and dimensionless solutions, physical-based models, and parametric models. These models were described in detail [1]Parametric modeling is commonly used to simulate breach growth as a time-dependent linear process and calculate outflow discharge from the breach using hydraulics principles [2]. Alhasan et al. [3] presented a simple one-dimensional mathematical model and a computer code to simulate the dam breaching process. These models were validated by small dams breaching during the floods in 2002 in the Czech Republic. Fread [4] developed an erosion model (BREACH) based on hydraulics principles, sediment transport, and soil mechanics to estimate breach size, time of formation, and outflow discharge. Říha et al. [5] investigated the dam break process for a cascade of small dams using a simple parametric model for piping and overtopping erosion, as well as a 2D shallow-water flow model for the flood in downstream areas. Goodarzi et al. [6] implemented mathematical and statistical methods to assess the effect of inflows and wind speeds on the dam’s overtopping failure.

Dam breaching studies can be divided into two main modes of erosion. The first mode is called “planar dam breach” where the flow overtops the whole dam width. While the second mode is called “spatial dam breach” where the flow overtops through the initial pilot channel (i.e., a channel created in the dam body). Therefore, the erosion will be in both vertical and horizontal directions [7].

The erosion process through the embankment dams occurs due to the shear stress applied by water flows. The dam breaching evolution can be divided into three stages [8][9], but Y. Yang et al. [10] divided the breach development into five stages: Stage I, the seepage erosion; Stage II, the initial breach formation; Stage III, the head erosion; Stage IV, the breach expansion; and Stage V, the re-equilibrium of the river channel through the breach. Many experimental tests have been carried out on non-cohesive embankment dams with an initial breach to examine the effect of upstream inflow discharges on the longitudinal profile evolution and the time to inflection point [11].

Zhang et al. [12] studied the effect of changing downstream slope angle, sediment grain size, and dam crest length on erosion rates. They noticed that increasing dam crest length and decreasing downstream slope angle lead to decreasing sediment transport rate. While the increase in sediment grain size leads to an increased sediment transport rate at the initial stages. Höeg et al. [13] presented a series of field tests to investigate the stability of embankment dams made of various materials. Overtopping and piping were among the failure tests carried out for the dams composed of homogeneous rock-fill, clay, or gravel with a height of up to 6.0 m. Hakimzadeh et al. [14] constructed 40 homogeneous cohesive and non-cohesive embankment dams to study the effect of changing sediment diameter and dam height on the breaching process. They also used genetic programming (GP) to estimate the breach outflow. Refaiy et al. [15] studied different scenarios for the downstream drain geometry, such as length, height, and angle, to minimize the effect of piping phenomena and therefore increase dam safety.

Zhu et al. [16] examined the effect of headcut erosion on dam breach growth, especially in the case of cohesive dams. They found that the breach growth in non-cohesive embankments is slower than cohesive embankments due to the little effect of headcut. Schmocker and Hager [7] proposed a relationship for estimating peak outflow from the dam breach process.(1)QpQin-1=1.7exp-20hc23d5013H0

where: Qp = peak outflow discharge.

Qin = inflow discharge.

hc = critical flow depth.

d50 = mean sediment diameter.

Ho = initial dam height.

Yu et al. [17] carried out an experimental study for homogeneous non-cohesive embankment dams in a 180° bending rectangular flume to determine the effect of overtopping flows on breaching formation. They found that the main factors influencing breach formation are water level, river discharge, and embankment material diameter.

Wu et al. [18] carried out a series of experiments to investigate the effect of breaching geometry on both non-cohesive and cohesive embankment dams in a U-bend flume due to overtopping flows. In the case of non-cohesive embankments, the non-symmetrical lateral expansion was noticed during the breach formation. This expansion was described by a coefficient ranging from 2.7 to 3.3.

The numerical models of the dam breach can be categorized according to different parameters, such as flow dimensions (1D, 2D, or 3D), flow governing equations, and solution methods. The 1D models are mainly used to predict the outflow hydrograph from the dam breach. Saberi et al. [19] applied the 1D Saint-Venant equation, which is solved by the finite difference method to investigate the outflow hydrograph during dam overtopping failure. Because of the ability to study dam profile evolution and breach formation, 2D models are more applicable than 1D models. Guan et al. [20] and Wu et al. [21] employed both 2D shallow water equations (SWEs) and sediment erosion equations, which are solved by the finite volume method to study the effect of the dam’s geometry parameters on outflow hydrograph and dam profile evolution. Wang et al. [22] also proposed a second-order hybrid-type of total variation diminishing (TVD) finite-difference to estimate the breach outflow by solving the 2D (SWEs). The accuracy of (SWEs) for both vertical flow contraction and surface roughness has been assessed [23]. They noted that the accuracy of (SWEs) is acceptable for milder slopes, but in the case of steeper slopes, modelers should be more careful. Generally, the accuracy of 2D models is still low, especially with velocity distribution over the flow depth, lateral momentum exchange, density-driven flows, and bottom friction [24]. Therefore, 3D models are preferred. Larocque et al. [25] and Yang et al. [26] started to use three-dimensional (3D) models that depend on the Reynolds-averaged Navier-Stokes (RANS) equations.

Previous experimental studies concluded that there is no clear relationship between the peak outflow from the dam breach and the initial breach characteristics. Some of these studies depend on the sharp-crested weir fixed at the end of the flume to determine the peak outflow from the breach, which leads to a decrease in the accuracy of outflow calculations at the microscale. The main goals of this study are to carry out a numerical simulation for a spatial dam breach due to overtopping flows by using (FLOW-3D) software to find an empirical equation for the peak outflow discharge from the breach and determine the worst-case that leads to accelerating the dam breaching process.

2. Numerical simulation

The current study for spatial dam breach is simulated by using (FLOW-3D) software [27], which is a powerful computational fluid dynamics (CFD) program.

2.1. Geometric presentations

A stereolithographic (STL) file is prepared for each change in the initial breach geometry and dimensions. The CAD program is useful for creating solid objects and converting them to STL format, as shown in Fig. 1.

2.2. Governing equations

The governing equations for water flow are three-dimensional Reynolds Averaged Navier-Stokes equations (RANS).

The continuity equation:(2)∂ui∂xi=0

The momentum equation:(3)∂ui∂t+1VFuj∂ui∂xj=1ρ∂∂xj-pδij+ν∂ui∂xj+∂uj∂xi-ρu`iu`j¯

where u is time-averaged velocity,ν is kinematic viscosity, VF is fractional volume open to flow, p is averaged pressure and -u`iu`j¯ are components of Reynold’s stress. The Volume of Fluid (VOF) technique is used to simulate the free surface profile. Hirt et al. [28] presented the VOF algorithm, which employs the function (F) to express the occupancy of each grid cell with fluid. The value of (F) varies from zero to unity. Zero value refers to no fluid in the grid cell, while the unity value refers to the grid cell being fully occupied with fluid. The free surface is formed in the grid cells having (F) values between zero and unity.(4)∂F∂t+1VF∂∂xFAxu+∂∂yFAyv+∂∂zFAzw=0

where (u, v, w) are the velocity components in (x, y, z) coordinates, respectively, and (AxAyAz) are the area fractions.

2.3. Boundary and initial conditions

To improve the accuracy of the results, the boundary conditions should be carefully determined. In this study, two mesh blocks are used to minimize the time consumed in the simulation. The boundary conditions for mesh block 1 are as follows: The inlet and sides boundaries are defined as a wall boundary condition (wall boundary condition is usually used for bound fluid by solid regions. In the case of viscous flows, no-slip means that the tangential velocity is equal to the wall velocity and the normal velocity is zero), the outlet is defined as a symmetry boundary condition (symmetry boundary condition is usually used to reduce computational effort during CFD simulation. This condition allows the flow to be transferred from one mesh block to another. No inputs are required for this boundary condition except that its location should be defined accurately), the bottom boundary is defined as a uniform flow rate boundary condition, and the top boundary is defined as a specific pressure boundary condition with assigned atmospheric pressure. The boundary conditions for mesh block 2 are as follows: The inlet is defined as a symmetry boundary condition, the outlet is defined as a free flow boundary condition, the bottom and sides boundaries are defined as a wall boundary condition, and the top boundary is defined as a specific pressure boundary condition with assigned atmospheric pressure as shown in Fig. 2. The initial conditions required to be set for the fluid (i.e., water) inside of the domain include configuration, temperature, velocities, and pressure distribution. The configuration of water depends on the dimensions and shape of the dam reservoir. While the other conditions have been assigned as follows: temperature is normal water temperature (25 °c) and pressure distribution is hydrostatic with no initial velocity.

2.4. Numerical method

FLOW-3D uses the finite volume method (FVM) to solve the governing equation (Reynolds-averaged Navier-Stokes) over the computational domain. A finite-volume method is an Eulerian approach for representing and evaluating partial differential equations in algebraic equations form [29]. At discrete points on the mesh geometry, values are determined. Finite volume expresses a small volume surrounding each node point on a mesh. In this method, the divergence theorem is used to convert volume integrals with a divergence term to surface integrals. After that, these terms are evaluated as fluxes at each finite volume’s surfaces.

2.5. Turbulent models

Turbulence is the chaotic, unstable motion of fluids that occurs when there are insufficient stabilizing viscous forces. In FLOW-3D, there are six turbulence models available: the Prandtl mixing length model, the one-equation turbulent energy model, the two-equation (k – ε) model, the Renormalization-Group (RNG) model, the two-equation (k – ω) models, and a large eddy simulation (LES) model. For simulating flow motion, the RNG model is adopted to simulate the motion behavior better than the k – ε and k – ω.

models [30]. The RNG model consists of two main equations for the turbulent kinetic energy KT and its dissipation.εT(5)∂kT∂t+1VFuAx∂kT∂x+vAy∂kT∂y+wAz∂kT∂z=PT+GT+DiffKT-εT(6)∂εT∂t+1VFuAx∂εT∂x+vAy∂εT∂y+wAz∂εT∂z=C1.εTKTPT+c3.GT+Diffε-c2εT2kT

where KT is the turbulent kinetic energy, PT is the turbulent kinetic energy production, GT is the buoyancy turbulence energy, εT is the turbulent energy dissipation rate, DiffKT and Diffε are terms of diffusion, c1, c2 and c3 are dimensionless parameters, in which c1 and c3 have a constant value of 1.42 and 0.2, respectively, c2 is computed from the turbulent kinetic energy (KT) and turbulent production (PT) terms.

2.6. Sediment scour model

The sediment scour model available in FLOW-3D can calculate all the sediment transport processes including Entrainment transport, Bedload transport, Suspended transport, and Deposition. The erosion process starts once the water flows remove the grains from the packed bed and carry them into suspension. It happens when the applied shear stress by water flows exceeds critical shear stress. This process is represented by entrainment transport in the numerical model. After entrained, the grains carried by water flow are represented by suspended load transport. After that, some suspended grains resort to settling because of the combined effect of gravity, buoyancy, and friction. This process is described through a deposition. Finally, the grains sliding motions are represented by bedload transport in the model. For the entrainment process, the shear stress applied by the fluid motion on the packed bed surface is calculated using the standard wall function as shown in Eq.7.(7)ks,i=Cs,i∗d50

where ks,i is the Nikuradse roughness and Cs,i is a user-defined coefficient. The critical bed shear stress is defined by a dimensionless parameter called the critical shields number as expressed in Eq.8.(8)θcr,i=τcr,i‖g‖diρi-ρf

where θcr,i is the critical shields number, τcr,i is the critical bed shear stress, g is the absolute value of gravity acceleration, di is the diameter of the sediment grain, ρi is the density of the sediment species (i) and ρf is the density of the fluid. The value of the critical shields number is determined according to the Soulsby-Whitehouse equation.(9)θcr,i=0.31+1.2d∗,i+0.0551-exp-0.02d∗,i

where d∗,i is the dimensionless diameter of the sediment, given by Eq.10.(10)d∗,i=diρfρi-ρf‖g‖μf213

where μf is the fluid dynamic viscosity. For the sloping bed interface, the value of the critical shields number is modified according to Eq.11.(11)θ`cr,i=θcr,icosψsinβ+cos2βtan2φi-sin2ψsin2βtanφi

where θ`cr,i is the modified critical shields number, φi is the angle of repose for the sediment, β is the angle of bed slope and ψ is the angle between the flow and the upslope direction. The effects of the rolling, hopping, and sliding motions of grains along the packed bed surface are taken by the bedload transport process. The volumetric bedload transport rate (qb,i) per width of the bed is expressed in Eq.12.(12)qb,i=Φi‖g‖ρi-ρfρfdi312

where Φi is the dimensionless bedload transport rate is calculated by using Meyer Peter and Müller equation.(13)Φi=βMPM,iθi-θ`cr,i1.5cb,i

where βMPM,i is the Meyer Peter and Müller user-defined coefficient and cb,i is the volume fraction of species i in the bed material. The suspended load transport is calculated as shown in Eq.14.(14)∂Cs,i∂t+∇∙Cs,ius,i=∇∙∇DCs,i

where Cs,i is the suspended sediment mass concentration, D is the diffusivity, and us,i is the grain velocity of species i. Entrainment and deposition are two opposing processes that take place at the same time. The lifting and settling velocities for both entrainment and deposition processes are calculated according to Eq.15 and Eq.16, respectively.(15)ulifting,i=αid∗,i0.3θi-θ`cr,igdiρiρf-1(16)usettling,i=υfdi10.362+1.049d∗,i3-10.36

where αi is the entrainment coefficient of species i and υf is the kinematic viscosity of the fluid.

2.7. Grid type

Using simple rectangular orthogonal elements in planes and hexahedral in volumes in the (FLOW-3D) program makes the mesh generation process easier, decreases the required memory, and improves numerical accuracy. Two mesh blocks were used in a joined form with a size ratio of 2:1. The first mesh block is coarser, which contains the reservoir water, and the second mesh block is finer, which contains the dam. For achieving accuracy and efficiency in results, the mesh size is determined by using a grid convergence test. The optimum uniform cell size for the first mesh block is 0.012 m and for the second mesh block is 0.006 m.

2.8. Time step

The maximum time step size is determined by using a Courant number, which controls the distance that the flow will travel during the simulation time step. In this study, the Courant number was taken equal to 0.25 to prevent the flow from traveling through more than one cell in the time step. Based on the Courant number, a maximum time step value of 0.00075 s was determined.

2.9. Numerical model validation

The numerical model accuracy was achieved by comparing the numerical model results with previous experimental results. The experimental study of Schmocker and Hager [7] was based on 31 tests with changes in six parameters (d50, Ho, Bo, Lk, XD, and Qin). All experimental tests were conducted in a straight open glass-sided flume. The horizontal flume has a rectangular cross-section with a width of 0.4 m and a height of 0.7 m. The flume was provided with a flow straightener and an intake with a length of 0.66 m. All tested dams were inserted at various distances (XD) from the intake. Test No.1 from this experimental program was chosen to validate the numerical model. The different parameters used in test No.1 are as follows:

(1) uniform sediment with a mean diameter (d50 = 0.31 mm), (2) Ho = 0.2 m, (3) Bo = 0.2 m, (4) Lk = 0.1 m,

(5) XD = 1.0 m, (6) Qin = 6.0 lit/s, (7) Su and Sd = 2:1, (8) mass density (ρs = 2650 kg/m3(9) Homogenous and non-cohesive embankment dam. As shown in Fig. 2, the simulation is contained within a rectangular grid with dimensions: 3.56 m in the x-direction (where 0.66 m is used as inlet, 0.9 m as dam base width, and 1.0 m as outlet), in y-direction 0.2 m (dam length), and in the z-direction 0.3 m, which represents the dam height (0.2 m) with a free distance (0.1 m) above the dam. There are two main reasons that this experimental program is preferred for the validation process. The first reason is that this program deals with homogenous, non-cohesive soil, which is available in FLOW-3D. The second reason is that this program deals with small-scale models which saves time for numerical simulation. Finally, some important assumptions were considered during the validation process. The flow is assumed to be incompressible, viscous, turbulent, and three-dimensional.

By comparing dam profiles at different time instants for the experimental test with the current numerical model, it appears that the numerical model gives good agreement as shown in Fig. 3 and Fig. 4, with an average error percentage of 9% between the experimental results and the numerical model.

3. Analysis and discussions

The current model is used to study the effects of different parameters such as (initial breach shapes, dimensions, locations, upstream and downstream dam slopes) on the peak outflow discharge, QP, time of peak outflow, tP, and rate of erosion, E.

This study consists of a group of scenarios. The first scenario is changing the shapes of the initial breach according to Singh [1], the most predicted shapes are rectangular and V-notch as shown in Fig. 5. The second scenario is changing the initial breach dimensions (i.e., width and depth). While the third scenario is changing the location of the initial breach. Eventually, the last scenario is changing the upstream and downstream dam slopes.

All scenarios of this study were carried out under the same conditions such as inflow discharge value (Qin=1.0lit/s), dimensions of the tested dam, where dam height (Ho=0.20m), crest width.

(Lk=0.1m), dam length (Bo=0.20m), and homogenous & non-cohesive soil with a mean diameter (d50=0.31mm).

3.1. Dam breaching process evolution

The dam breaching process is a very complex process due to the quick changes in hydrodynamic conditions during dam failure. The dam breaching process starts once water flows reach the downstream face of the dam. During the initial stage of dam breaching, the erosion process is relatively quiet due to low velocities of flow. As water flows continuously, erosion rates increase, especially in two main zones: the crest and the downstream face. As soon as the dam crest is totally eroded, the water levels in the dam reservoir decrease rapidly, accompanied by excessive erosion in the dam body. The erosion process continues until the water levels in the dam reservoir equal the remaining height of the dam.

According to Zhou et al. [11], the breaching process consists of three main stages. The first stage starts with beginning overtopping flow, then ends when the erosion point directed upstream and reached the inflection point at the inflection time (ti). The second stage starts from the end of the stage1 until the occurrence of peak outflow discharge at the peak outflow time (tP). The third stage starts from the end of the stage2 until the value of outflow discharge becomes the same as the value of inflow discharge at the final time (tf). The outflow discharge from the dam breach increases rapidly during stage1 and stage2 because of the large dam storage capacity (i.e., the dam reservoir is totally full of water) and excessive erosion. While at stage3, the outflow values start to decrease slowly because most of the dam’s storage capacity was run out. The end of stage3 indicates that the dam storage capacity was totally run out, so the outflow equalized with the inflow discharge as shown in Fig. 6 and Fig. 7.

3.2. The effect of initial breach shape

To identify the effect of the initial breach shape on the evolution of the dam breaching process. Three tests were carried out with different cross-section areas for each shape. The initial breach is created at the center of the dam crest. Each test had an ID to make the process of arranging data easier. The rectangular shape had an ID (Rec5h & 5b), which means that its depth and width are equal to 5% of the dam height, and the V-notch shape had an ID (V-noch5h & 1:1) which means that its depth is equal to 5% of the dam height and its side slope is equal to 1:1. The comparison between rectangular and V-notch shapes is done by calculating the ratio between maximum dam height at different times (ZMax) to the initial dam height (Ho), rate of erosion, and hydrograph of outflow discharge for each test. The rectangular shape achieves maximum erosion rate and minimum inflection time, in addition to a rapid decrease in the dam reservoir levels. Therefore, the dam breaching is faster in the case of a rectangular shape than in a V-notch shape, which has the same cross-section area as shown in Fig. 8.

Also, by comparing the hydrograph for each test, the peak outflow discharge value in the case of a rectangular shape is higher than the V-notch shape by 5% and the time of peak outflow for the rectangular shape is shorter than the V-notch shape by 9% as shown in Fig. 9.

3.3. The effect of initial breach dimensions

The results of the comparison between the different initial breach shapes indicate that the worst initial breach shape is rectangular, so the second scenario from this study concentrated on studying the effect of a change in the initial rectangular breach dimensions. Groups of tests were carried out with different depths and widths for the rectangular initial breach. The first group had a depth of 5% from the dam height and with three different widths of 5,10, and 15% from the dam height, the second group had a depth of 10% with three different widths of 5,10, and 15%, the third group had a depth of 15% with three different widths of 5,10, and 15% and the final group had a width of 15% with three different heights of 5, 10, and 15% for a rectangular breach shape. The comparison was made as in the previous section to determine the worst case that leads to the quick dam failure as shown in Fig. 10.

The results show that the (Rec 5 h&15b) test achieves a maximum erosion rate for a shorter period of time and a minimum ratio for (Zmax / Ho) as shown in Fig. 10, which leads to accelerating the dam failure process. The dam breaching process is faster with the minimum initial breach depth and maximum initial breach width. In the case of a minimum initial breach depth, the retained head of water in the dam reservoir is high and the crest width at the bottom of the initial breach (L`K) is small, so the erosion point reaches the inflection point rapidly. While in the case of the maximum initial breach width, the erosion perimeter is large.

3.4. The effect of initial breach location

The results of the comparison between the different initial rectangular breach dimensions indicate that the worst initial breach dimension is (Rec 5 h&15b), so the third scenario from this study concentrated on studying the effect of a change in the initial breach location. Three locations were checked to determine the worst case for the dam failure process. The first location is at the center of the dam crest, which was named “Center”, the second location is at mid-distance between the dam center and dam edge, which was named “Mid”, and the third location is at the dam edge, which was named “Edge” as shown in Fig. 11. According to this scenario, the results indicate that the time of peak outflow discharge (tP) is the same in the three cases, but the maximum value of the peak outflow discharge occurs at the center location. The difference in the peak outflow values between the three cases is relatively small as shown in Fig. 12.

The rates of erosion were also studied for the three cases. The results show that the maximum erosion rate occurs at the center location as shown in Fig. 13. By making a comparison between the three cases for the dam storage volume. The results show that the center location had the minimum values for the dam storage volume, which means that a large amount of water has passed to the downstream area as shown in Fig. 14. According to these results, the center location leads to increased erosion rate and accelerated dam failure process compared with the two other cases. Because the erosion occurs on both sides, but in the case of edge location, the erosion occurs on one side.

3.5. The effect of upstream and downstream dam slopes

The results of the comparison between the different initial rectangular breach locations indicate that the worst initial breach location is the center location, so the fourth scenario from this study concentrated on studying the effect of a change in the upstream (Su) and downstream (Sd) dam slopes. Three slopes were checked individually for both upstream and downstream slopes to determine the worst case for the dam failure process. The first slope value is (2H:1V), the second slope value is (2.5H:1V), and the third slope value is (3H:1V). According to this scenario, the results show that the decreasing downstream slope angle leads to increasing time of peak outflow discharge (tP) and decreasing value of peak outflow discharge. The difference in the peak outflow values between the three cases for the downstream slope is 2%, as shown in Fig. 15, but changing the upstream slope has a negligible impact on the peak outflow discharge and its time as shown in Fig. 16.

The rates of erosion were also studied in the three cases for both upstream and downstream slopes. The results show that the maximum erosion rate increases by 6.0% with an increasing downstream slope angle by 4°, as shown in Fig. 17. The results also indicate that the erosion rates aren’t affected by increasing or decreasing the upstream slope angle, as shown in Fig. 18. According to these results, increasing the downstream slope angle leads to increased erosion rate and accelerated dam failure process compared with the upstream slope angle. Because of increasing shear stress applied by water flows in case of increasing downstream slope.

According to all previous scenarios, the dimensionless peak outflow discharge QPQin is presented for a fixed dam height (Ho) and inflow discharge (Qin). Fig. 19 illustrates the relationship between QP∗=QPQin and.

Lr=ho2/3∗bo2/3Ho. The deduced relationship achieves R2=0.96.(17)QP∗=2.2807exp-2.804∗Lr

4. Conclusions

A spatial dam breaching process was simulated by using FLOW-3D Software. The validation process was performed by making a comparison between the simulated results of dam profiles and the dam profiles obtained by Schmocker and Hager [7] in their experimental study. And also, the peak outflow value recorded an error percentage of 12% between the numerical model and the experimental study. This model was used to study the effect of initial breach shape, dimensions, location, and dam slopes on peak outflow discharge, time of peak outflow, and the erosion process. By using the parameters obtained from the validation process, the results of this study can be summarized in eight points as follows.1.

The rectangular initial breach shape leads to an accelerating dam failure process compared with the V-notch.2.

The value of peak outflow discharge in the case of a rectangular initial breach is higher than the V-notch shape by 5%.3.

The time of peak outflow discharge for a rectangular initial breach is shorter than the V-notch shape by 9%.4.

The minimum depth and maximum width for the initial breach achieve maximum erosion rates (increasing breach width, b0, or decreasing breach depth, h0, by 5% from the dam height leads to an increase in the maximum rate of erosion by 11% and 15%, respectively), so the dam failure is rapid.5.

The center location of the initial breach leads to an accelerating dam failure compared with the edge location.6.

The initial breach location has a negligible effect on the peak outflow discharge value and its time.7.

Increasing the downstream slope angle by 4° leads to an increase in both peak outflow discharge and maximum rate of erosion by 2.0% and 6.0%, respectively.8.

The upstream slope has a negligible effect on the dam breaching process.

References

[1]V. SinghDam breach modeling technologySpringer Science & Business Media (1996)Google Scholar[2]Wahl TL. Prediction of embankment dam breach parameters: a literature review and needs assessment. 1998.Google Scholar[3]Z. Alhasan, J. Jandora, J. ŘíhaStudy of dam-break due to overtopping of four small dams in the Czech RepublicActa Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 63 (3) (2015), pp. 717-729 View PDFCrossRefView Record in ScopusGoogle Scholar[4]D. FreadBREACH, an erosion model for earthen dam failures: Hydrologic Research LaboratoryNOAA, National Weather Service (1988)Google Scholar[5]J. Říha, S. Kotaška, L. PetrulaDam Break Modeling in a Cascade of Small Earthen Dams: Case Study of the Čižina River in the Czech RepublicWater, 12 (8) (2020), p. 2309, 10.3390/w12082309 View PDFView Record in ScopusGoogle Scholar[6]E. Goodarzi, L. Teang Shui, M. ZiaeiDam overtopping risk using probabilistic concepts–Case study: The Meijaran DamIran Ain Shams Eng J, 4 (2) (2013), pp. 185-197ArticleDownload PDFView Record in ScopusGoogle Scholar[7]L. Schmocker, W.H. HagerPlane dike-breach due to overtopping: effects of sediment, dike height and dischargeJ Hydraul Res, 50 (6) (2012), pp. 576-586 View PDFCrossRefView Record in ScopusGoogle Scholar[8]J.S. Walder, R.M. Iverson, J.W. Godt, M. Logan, S.A. SolovitzControls on the breach geometry and flood hydrograph during overtopping of noncohesive earthen damsWater Resour Res, 51 (8) (2015), pp. 6701-6724View Record in ScopusGoogle Scholar[9]H. Wei, M. Yu, D. Wang, Y. LiOvertopping breaching of river levees constructed with cohesive sedimentsNat Hazards Earth Syst Sci, 16 (7) (2016), pp. 1541-1551 View PDFCrossRefView Record in ScopusGoogle Scholar[10]Y. Yang, S.-Y. Cao, K.-J. Yang, W.-P. LiYang K-j, Li W-p. Experimental study of breach process of landslide dams by overtopping and its initiation mechanismsJ Hydrodynamics, 27 (6) (2015), pp. 872-883ArticleDownload PDFCrossRefView Record in ScopusGoogle Scholar[11]G.G.D. Zhou, M. Zhou, M.S. Shrestha, D. Song, C.E. Choi, K.F.E. Cui, et al.Experimental investigation on the longitudinal evolution of landslide dam breaching and outburst floodsGeomorphology, 334 (2019), pp. 29-43ArticleDownload PDFView Record in ScopusGoogle Scholar[12]J. Zhang, Z.-x. Guo, S.-y. CaoYang F-g. Experimental study on scour and erosion of blocked damWater Sci Eng, 5 (2012), pp. 219-229ArticleDownload PDFView Record in ScopusGoogle Scholar[13]K. Höeg, A. Løvoll, K. VaskinnStability and breaching of embankment dams: Field tests on 6 m high damsInt J Hydropower Dams, 11 (2004), pp. 88-92View Record in ScopusGoogle Scholar[14]H. Hakimzadeh, V. Nourani, A.B. AminiGenetic programming simulation of dam breach hydrograph and peak outflow dischargeJ Hydrol Eng, 19 (4) (2014), pp. 757-768View Record in ScopusGoogle Scholar[15]A.R. Refaiy, N.M. AboulAtta, N.Y. Saad, D.A. El-MollaModeling the effect of downstream drain geometry on seepage through earth damsAin Shams Eng J, 12 (3) (2021), pp. 2511-2531ArticleDownload PDFView Record in ScopusGoogle Scholar[16]Y. Zhu, P.J. Visser, J.K. Vrijling, G. WangExperimental investigation on breaching of embankmentsScience China Technological Sci, 54 (1) (2011), pp. 148-155 View PDFCrossRefView Record in ScopusGoogle Scholar[17]M.-H. Yu, H.-Y. Wei, Y.-J. Liang, Y. ZhaoInvestigation of non-cohesive levee breach by overtopping flowJ Hydrodyn, 25 (4) (2013), pp. 572-579ArticleDownload PDFCrossRefView Record in ScopusGoogle Scholar[18]S. Wu, M. Yu, H. Wei, Y. Liang, J. ZengNon-symmetrical levee breaching processes in a channel bend due to overtoppingInt J Sedim Res, 33 (2) (2018), pp. 208-215ArticleDownload PDFView Record in ScopusGoogle Scholar[19]O. Saberi, G. ZenzNumerical investigation on 1D and 2D embankment dams failure due to overtopping flowInt J Hydraulic Engineering, 5 (2016), pp. 9-18View Record in ScopusGoogle Scholar[20]M. Guan, N.G. Wright, P.A. Sleigh2D Process-Based Morphodynamic Model for Flooding by Noncohesive Dyke BreachJ Hydraul Eng, 140 (7) (2014), p. 04014022, 10.1061/(ASCE)HY.1943-7900.0000861 View PDFView Record in ScopusGoogle Scholar[21]W. Wu, R. Marsooli, Z. HeDepth-Averaged Two-Dimensional Model of Unsteady Flow and Sediment Transport due to Noncohesive Embankment Break/BreachingJ Hydraul Eng, 138 (6) (2012), pp. 503-516View Record in ScopusGoogle Scholar[22]Z. Wang, D.S. BowlesThree-dimensional non-cohesive earthen dam breach model. Part 1: Theory and methodologyAdv Water Resour, 29 (10) (2006), pp. 1528-1545ArticleDownload PDFView Record in ScopusGoogle Scholar[23]Říha J, Duchan D, Zachoval Z, Erpicum S, Archambeau P, Pirotton M, et al. Performance of a shallow-water model for simulating flow over trapezoidal broad-crested weirs. J Hydrology Hydromechanics. 2019;67:322-8.Google Scholar[24]C.B. VreugdenhilNumerical methods for shallow-water flowSpringer Science & Business Media (1994)Google Scholar[25]L.A. Larocque, J. Imran, M.H. Chaudhry3D numerical simulation of partial breach dam-break flow using the LES and k–∊ turbulence modelsJ Hydraul Res, 51 (2) (2013), pp. 145-157 View PDFCrossRefView Record in ScopusGoogle Scholar[26]C. Yang, B. Lin, C. Jiang, Y. LiuPredicting near-field dam-break flow and impact force using a 3D modelJ Hydraul Res, 48 (6) (2010), pp. 784-792 View PDFCrossRefView Record in ScopusGoogle Scholar[27]FLOW-3D. Version 11.1.1 Flow Science, Inc., Santa Fe, NM. https://wwwflow3dcom.Google Scholar[28]C.W. Hirt, B.D. NicholsVolume of fluid (VOF) method for the dynamics of free boundariesJ Comput Phys, 39 (1) (1981), pp. 201-225ArticleDownload PDFGoogle Scholar[29]S.V. PatankarNumerical heat transfer and fluid flow, Hemisphere PublCorp, New York, 58 (1980), p. 288View Record in ScopusGoogle Scholar[30]M. Alemi, R. MaiaNumerical simulation of the flow and local scour process around single and complex bridge piersInt J Civil Eng, 16 (5) (2018), pp. 475-487 View PDFCrossRefView Record in ScopusGoogle Scholar

A 3-D numerical simulation of the characteristics of open channel flows with submerged rigid vegetation

A 3-D numerical simulation of the characteristics of open channel flows with submerged rigid vegetation

수중 강성 식생이 있는 개방 수로 흐름의 특성에 대한 3차원 수치 시뮬레이션

Journal of Hydrodynamics (2021)Cite this article

Abstract

이 논문은 FLOW-3D를 적용하여 다양한 흐름 배출 및 식생 시나리오가 유속(종방향, 횡방향 및 수직 속도 포함)에 미치는 영향을 조사합니다.

실험적 측정을 통한 검증 후 식생직경, 식생높이, 유출량에 대한 민감도 분석을 수행하였습니다. 종방향 속도의 경우 흐름 구조에 대한 가장 큰 영향은 배출보다는 식생 직경에서 비롯됩니다.

그러나 식생 높이는 수직 분포의 변곡점을 결정합니다. 식생 지역, 즉 상류와 하류의 두 위치에서 횡단 속도를 비교하면 수심을 따라 대칭 패턴이 식별됩니다. 식생 지역의 횡단 및 수직 유체 순환 패턴을 포함하여 흐름 또는 식생 시나리오에 관계없이 수직 속도에서도 동일한 패턴이 관찰됩니다.

또한 식생 직경이 클수록 이러한 패턴이 더 분명해집니다. 상부 순환은 식생 캐노피 근처에서 발생합니다. 식생 지역의 가로 세로 방향 순환에 관한 이러한 발견은 수중 식생을 통한 3차원 흐름 구조를 밝혀줍니다.

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

References

  1. [1]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.Article Google Scholar 
  2. [2]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.Article Google Scholar 
  3. [3]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.Article Google Scholar 
  4. [4]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.MathSciNet Article Google Scholar 
  5. [5]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.Article Google Scholar 
  6. [6]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.Article Google Scholar 
  7. [7]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.MathSciNet Article Google Scholar 
  8. [8]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.Article Google Scholar 
  9. [9]Panigrahi K., Khatua K. K. Prediction of velocity distribution in straight channel with rigid vegetation [J]. Aquatic Procedia, 2015, 4: 819–825.Article Google Scholar 
  10. [10]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.Article Google Scholar 
  11. [11]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.Article Google Scholar 
  12. [12]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.Article Google Scholar 
  13. [13]Wu F. S. Characteristics of flow resistance in open channels with non-submerged rigid vegetation [J]. Journal of Hydrodynamics, 2008, 20(2): 239–245.Article Google Scholar 
  14. [14]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.Article Google Scholar 
  15. [15]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.Article Google Scholar 
  16. [16]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.MathSciNet Article Google Scholar 
  17. [17]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.Article Google Scholar 
  18. [18]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.Article Google Scholar 
  19. [19]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.Article Google Scholar 
  20. [20]Aydin M. C., Emiroglu M. E. Determination of capacity of labyrinth side weir by CFD [J]. Flow Measurement and Instrumentation, 2013, 29: 1–8.Article Google Scholar 
  21. [21]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 
  22. [22]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.Article Google Scholar 
  23. [23]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.Article Google Scholar 
  24. [24]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.Article Google Scholar 
  25. [25]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.Article Google Scholar 
  26. [26]Fathi-Maghadam M., Kouwen N. Nonrigid, nonsubmerged, vegetative roughness on floodplains [J]. Journal of Hydraulic Engineering, ASCE, 1997, 123(1): 51–57.Article Google Scholar 
  27. [27]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.MathSciNet Article Google Scholar 
  28. [28]Ghisalberti M., Nepf H. Mass transport in vegetated shear flows [J]. Environmental Fluid Mechanics, 2005, 5(6): 527–551.
Fig. 5 Comparison of experimental SEM image and CtFD simulated melt pool with beam diameters of(a)700 μm,(b)1000 μm, and(c)1300 μm and an absorption rate of 0.3. Electron beam power and scan speed are 900 W and 100 mm s-1, respectively

추가 생산용 전자빔 조사에 의한 316L 스테인리스 용융 · 응고 거동

Melting and Solidification Behavior of 316L Steel Induced by Electron-Beam Irradiation for Additive Manufacturing

付加製造用電子ビーム照射による 316L ステンレス鋼の溶融・凝固挙動

奥 川 将 行*・宮 田 雄一朗*・王     雷*・能 勢 和 史*
小 泉 雄一郎*・中 野 貴 由*
Masayuki OKUGAWA, Yuichiro MIYATA, Lei WANG, Kazufumi NOSE,
Yuichiro KOIZUMI and Takayoshi NAKANO

Abstract

적층 제조(AM) 기술은 복잡한 형상의 3D 부품을 쉽게 만들고 미세 구조 제어를 통해 재료 특성을 크게 제어할 수 있기 때문에 많은 관심을 받았습니다. PBF(Powderbed fusion) 방식의 AM 공정에서는 금속 분말을 레이저나 전자빔으로 녹이고 응고시키는 과정을 반복하여 3D 부품을 제작합니다.

일반적으로 응고 미세구조는 Hunt[Mater. 과학. 영어 65, 75(1984)]. 그러나 CET 이론이 일반 316L 스테인리스강에서도 높은 G와 R로 인해 PBF형 AM 공정에 적용될 수 있을지는 불확실하다.

본 연구에서는 미세구조와 응고 조건 간의 관계를 밝히기 위해 전자빔 조사에 의해 유도된 316L 강의 응고 미세구조를 분석하고 CtFD(Computational Thermal-Fluid Dynamics) 방법을 사용하여 고체/액체 계면에서의 응고 조건을 평가했습니다.

CET 이론과 반대로 높은 G 조건에서 등축 결정립이 종종 형성되는 것으로 밝혀졌다. CtFD 시뮬레이션은 약 400 mm s-1의 속도까지 유체 흐름이 있음을 보여 주며 수상 돌기의 파편 및 이동의 영향으로 등축 결정립이 형성됨을 시사했습니다.

Additive manufacturing(AM)technologies have attracted much attention because it enables us to build 3D parts with complicated geometry easily and control material properties significantly via the control of microstructures. In the powderbed fusion(PBF)type AM process, 3D parts are fabricated by repeating a process of melting and solidifying metal powders by laser or electron beams. In general, the solidification microstructures can be predicted from solidification conditions defined by the combination of temperature gradient G and solidification rate R on the basis of columnar-equiaxed transition(CET)theory proposed by Hunt [Mater. Sci. Eng. 65, 75(1984)]. However, it is unclear whether the CET theory can be applied to the PBF type AM process because of the high G and R, even for general 316L stainless steel. In this study, to reveal relationships between microstructures and solidification conditions, we have analyzed solidification microstructures of 316L steel induced by electronbeam irradiation and evaluated solidification conditions at the solid/liquid interface using a computational thermal-fluid dynamics (CtFD)method. It was found that equiaxed grains were often formed under high G conditions contrary to the CET theory. CtFD simulation revealed that there is a fluid flow up to a velocity of about 400 mm s-1, and suggested that equiaxed grains are formed owing to the effect of fragmentations and migrations of dendrites.

Keywords

Additive Manufacturing, 316L Stainless Steel, Powder Bed Fusion, Electron Beam Melting, Computational Thermal
Fluid Dynamics Simulation

Fig. 1 Width, height, and height differences calculated from laser microscope analysis of melt tracks formed by scanning electron beam. Fig. 2(a)Scanning electron microscope(SEM)image and(b) corresponding electron back-scattering diffraction(EBSD) IPF-map taken from the electron-beam irradiated region in P900-V100 sample. Fig. 3 Average grain size and their aspect ratio calculated from EBSD IPF-map taken from the electron-beam irradiated region.
Fig. 1 Width, height, and height differences calculated from laser microscope analysis of melt tracks formed by scanning electron beam. Fig. 2(a)Scanning electron microscope(SEM)image and(b) corresponding electron back-scattering diffraction(EBSD) IPF-map taken from the electron-beam irradiated region in P900-V100 sample. Fig. 3 Average grain size and their aspect ratio calculated from EBSD IPF-map taken from the electron-beam irradiated region.
Fig. 4 Comparison of experimental SEM image and computational thermal fluid dynamics(CtFD)simulated melt pool with a beam diameter of 700 μm and absorption rates of(a)0.3,(b)0.5, and (c)0.7. Electron beam power and scan speed are 900 W and 100 mm s-1, respectively.
Fig. 4 Comparison of experimental SEM image and computational thermal fluid dynamics(CtFD)simulated melt pool with a beam diameter of 700 μm and absorption rates of(a)0.3,(b)0.5, and (c)0.7. Electron beam power and scan speed are 900 W and 100 mm s-1, respectively.
Fig. 5 Comparison of experimental SEM image and CtFD simulated melt pool with beam diameters of(a)700 μm,(b)1000 μm, and(c)1300 μm and an absorption rate of 0.3. Electron beam power and scan speed are 900 W and 100 mm s-1, respectively
Fig. 5 Comparison of experimental SEM image and CtFD simulated melt pool with beam diameters of(a)700 μm,(b)1000 μm, and(c)1300 μm and an absorption rate of 0.3. Electron beam power and scan speed are 900 W and 100 mm s-1, respectively
Fig. 6 Depth of melt tracks calculated from experimental SEM image and CtFD simulation results.
Fig. 6 Depth of melt tracks calculated from experimental SEM image and CtFD simulation results.
Fig. 7 G-R plots of 316L steel colored by(a)aspect ratio of crystalline grains and(b)fluid velocity.
Fig. 7 G-R plots of 316L steel colored by(a)aspect ratio of crystalline grains and(b)fluid velocity.
Fig. 8 Comparison of solidification microstructure(EBSD IPF-map)of melt region formed by scanning electron beam and corresponding snap shot of CtFD simulation colored by fluid velocity
Fig. 8 Comparison of solidification microstructure(EBSD IPF-map)of melt region formed by scanning electron beam and corresponding snap shot of CtFD simulation colored by fluid velocity

References

1) M.C. Sow, T. De Terris, O. Castelnau, Z. Hamouche, F. Coste, R.
Fabbro and P. Peyre: “Influence of beam diameter on Laser Powder

Bed Fusion(L-PBF)process”, Addit. Manuf. 36(2020), 101532.
2) J.C. Simmons, X. Chen, A. Azizi, M.A. Daeumer, P.Y. Zavalij, G.
Zhou and S.N. Schiffres: “Influence of processing and microstructure
on the local and bulk thermal conductivity of selective laser melted
316L stainless steel”, Addit. Manuf. 32(2020), 100996.
3) S. Dryepondt, P. Nandwana, P. Fernandez-Zelaia and F. List:
“Microstructure and High Temperature Tensile properties of 316L
Fabricated by Laser Powder-Bed Fusion”, Addit. Manuf. 37(2020),
101723.
4) S.H. Sun, T. Ishimoto, K. Hagihara, Y. Tsutsumi, T. Hanawa and T.
Nakano: “Excellent mechanical and corrosion properties of austenitic
stainless steel with a unique crystallographic lamellar microstructure
via selective laser melting”, Scr. Mater. 159(2019), 89-93.
5) T. Ishimoto, S. Wu, Y. Ito, S.H. Sun, H. Amano and T. Nakano:
“Crystallographic orientation control of 316L austenitic stainless
steel via selective laser melting”, ISIJ Int. 60(2020), 1758-1764.
6) T. Ishimoto, K. Hagihara, K. Hisamoto, S.H. Sun and T. Nakano:
“Crystallographic texture control of beta-type Ti-15Mo-5Zr3Al alloy by selective laser melting for the development of novel
implants with a biocompatible low Young’s modulus”, Scr. Mater.
132(2017), 34-38.
7) X. Ding, Y. Koizumi, D. Wei and A. Chiba: “Effect of process
parameters on melt pool geometry and microstructure development
for electron beam melting of IN718: A systematic single bead
analysis study”, Addit. Manuf. 26(2019), 215-226.
8) K. Karayagiz, L. Johnson, R. Seede, V. Attari, B. Zhang, X. Huang,
S. Ghosh, T. Duong, I. Karaman, A. Elwany and R. Arróyave: “Finite
interface dissipation phase field modeling of Ni-Nb under additive
manufacturing conditions”, Acta Mater. 185(2020), 320-339.
9) M.M. Kirka, Y. Lee, D.A. Greeley, A. Okello, M.J. Goin, M.T.
Pearce and R.R. Dehoff: “Strategy for Texture Management in
Metals Additive Manufacturing”, JOM, 69(2017), 523-531.
10) S.S. Babu, N. Raghavan, J. Raplee, S.J. Foster, C. Frederick, M.
Haines, R. Dinwiddie, M.K. Kirka, A. Plotkowski, Y. Lee and
R.R. Dehoff: “Additive Manufacturing of Nickel Superalloys:
Opportunities for Innovation and Challenges Related to
Qualification”, Metall. Mater. Trans. A. 49(2018), 3764-3780.
11) M.R. Gotterbarm, A.M. Rausch and C. Körner: “Fabrication of
Single Crystals through a μ-Helix Grain Selection Process during
Electron Beam Metal Additive Manufacturing”, Metals, 10(2020),
313.
12) J.D.D. Hunt: “Steady state columnar and equiaxed growth of
dendrites and eutectic”, Mater. Sci. Eng. 65(1984), 75-83.
13) S. Bontha, N.W. Klingbeil, P.A. Kobryn and H.L. Fraser: “Effects of
process variables and size-scale on solidification microstructure in
beam-based fabrication of bulky 3D structures”, Mater. Sci. Eng. A.
513-514(2009), 311-318.
14) J. Gockel and J. Beuth: “Understanding Ti-6Al-4V microstructure
control in additive manufacturing via process maps”, 24th Int. SFF
Symp. – An Addit. Manuf. Conf. SFF 2013.(2013), 666-674.
15) B. Schoinochoritis, D. Chantzis and K. Salonitis: “Simulation of
metallic powder bed additive manufacturing processes with the finite
element method: A critical review”, Proc. of Instit. Mech. Eng., Part
B: J. Eng. Manuf. 231(2017), 96-117.
16)小泉雄一郎: “計算機シミュレーションを用いたAdditive
Manufacturing プロセス最適化予測”, スマートプロセス学会誌,
8-4(2019), 132-138.
17) Y. Zhao, Y. Koizumi, K. Aoyagi, D. Wei, K. Yamanaka and A. Chiba:
“Molten pool behavior and effect of fluid flow on solidification
conditions in selective electron beam melting(SEBM)of a
biomedical Co-Cr-Mo alloy”, Addit. Manuf. 26(2019), 202-214.
18) C. Tang, J.L. Tan and C.H. Wong: “A numerical investigation on
the physical mechanisms of single track defects in selective laser
melting”, Int. J. Heat Mass Transf. 126(2018), 957-968.
19) Technical data for Iron, [Online]. Available: http://periodictable.com/
Elements/026/data.html. [Accessed: 8-Feb-2021].
20) N. Raghavan, R. Dehoff, S. Pannala, S. Simunovic, M. Kirka, J.
Turner, N. Carl-son and S.S. Babu: “Numerical modeling of heattransfer and the influence of process parameters on tailoring the grain
morphology of IN718 in electron beam additive manufacturing”,
Acta Mater. 112(2016), 303-314.
21) S. Morita, Y. Miki and K. Toishi: “Introduction of Dendrite
Fragmentation to Microstructure Calculation by Cellular Automaton
Method”, Tetsu-to-Hagane. 104(2018), 559-566.
22) H. Esaka and M. Tamura: “Model Experiment Using Succinonitrile
on the Formation of Equiaxed Grains caused by Forced Convection”,
Tetsu-to-Hagane. 86(2000), 252-258.

Numerical study of the dam-break waves and Favre waves down sloped wet rigid-bed at laboratory scale

Numerical study of the dam-break waves and Favre waves down sloped wet rigid-bed at laboratory scale

WenjunLiua  BoWangb  YakunGuoc

a State Key Laboratory of Hydraulics and Mountain River Engineering, College of Water Resource and Hydropower, Sichuan University, Chengdu, 610065, China
State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University, College Of Water Resource and Hydropower, Chengdu, 610065, China
faculty of Engineering & Informatics, University of Bradford, BD7 1DP, UK

Abstract

The bed slope and the tailwater depth are two important ones among the factors that affect the propagation of the dam-break flood and Favre waves. Most previous studies have only focused on the macroscopic characteristics of the dam-break flows or Favre waves under the condition of horizontal bed, rather than the internal movement characteristics in sloped channel. The present study applies two numerical models, namely, large eddy simulation (LES) and shallow water equations (SWEs) models embedded in the CFD software package FLOW-3D to analyze the internal movement characteristics of the dam-break flows and Favre waves, such as water level, the velocity distribution, the fluid particles acceleration and the bed shear stress, under the different bed slopes and water depth ratios. The results under the conditions considered in this study show that there is a flow state transition in the flow evolution for the steep bed slope even in water depth ratio α = 0.1 (α is the ratio of the tailwater depth to the reservoir water depth). The flow state transition shows that the wavefront changes from a breaking state to undular. Such flow transition is not observed for the horizontal slope and mild bed slope. The existence of the Favre waves leads to a significant increase of the vertical velocity and the vertical acceleration. In this situation, the SWEs model has poor prediction. Analysis reveals that the variation of the maximum bed shear stress is affected by both the bed slope and tailwater depth. Under the same bed slope (e.g., S0 = 0.02), the maximum bed shear stress position develops downstream of the dam when α = 0.1, while it develops towards the end of the reservoir when α = 0.7. For the same water depth ratio (e.g., α = 0.7), the maximum bed shear stress position always locates within the reservoir at S0 = 0.02, while it appears in the downstream of the dam for S0 = 0 and 0.003 after the flow evolves for a while. The comparison between the numerical simulation and experimental measurements shows that the LES model can predict the internal movement characteristics with satisfactory accuracy. This study improves the understanding of the effect of both the bed slope and the tailwater depth on the internal movement characteristics of the dam-break flows and Favre waves, which also provides a valuable reference for determining the flood embankment height and designing the channel bed anti-scouring facility.

댐붕괴 홍수와 파브르 파도의 전파에 영향을 미치는 요인 중 하상경사와 후미수심은 두 가지 중요한 요소이다. 대부분의 선행 연구들은 경사 수로에서의 내부 이동 특성보다는 수평층 조건에서 댐파괴류나 Favre파동의 거시적 특성에만 초점을 맞추었다.

본 연구에서는 CFD 소프트웨어 패키지 FLOW-3D에 내장된 LES(Large Eddy Simulation) 및 SWE(Shallow Water Equation) 모델의 두 가지 수치 모델을 적용하여 댐-파괴 흐름 및 Favre 파도의 내부 이동 특성을 분석합니다.

수위, 속도 분포, 유체 입자 가속도 및 층 전단 응력, 다양한 층 경사 및 수심 비율로. 본 연구에서 고려한 조건하의 결과는 수심비 α = 0.1(α는 저수지 수심에 대한 tailwater 깊이의 비율)에서도 급경사면에 대한 유동상태 전이가 있음을 보여준다. 유동 상태 전이는 파면이 파단 상태에서 비정형으로 변하는 것을 보여줍니다.

수평 경사와 완만한 바닥 경사에서는 이러한 흐름 전이가 관찰되지 않습니다. Favre 파의 존재는 수직 속도와 수직 가속도의 상당한 증가로 이어집니다. 이 상황에서 SWE 모델은 예측이 좋지 않습니다.

분석에 따르면 최대 바닥 전단 응력의 변화는 바닥 경사와 꼬리 수심 모두에 영향을 받습니다. 동일한 바닥 경사(예: S0 = 0.02)에서 최대 바닥 전단 응력 위치는 α = 0.1일 때 댐의 하류에서 발생하고 α = 0.7일 때 저수지의 끝쪽으로 발생합니다.

동일한 수심비(예: α = 0.7)에 대해 최대 바닥 전단 응력 위치는 항상 S0 = 0.02에서 저수지 내에 위치하는 반면, S0 = 0 및 0.003에 대해 흐름이 진화한 후 댐 하류에 나타납니다. 수치적 시뮬레이션과 실험적 측정을 비교한 결과 LES 모델이 내부 움직임 특성을 만족스러운 정확도로 예측할 수 있음을 알 수 있습니다.

본 연구는 댐 파절류 및 Favre파의 내부 이동 특성에 대한 하상 경사 및 후미 수심의 영향에 대한 이해를 향상 시키며, 이는 또한 제방 높이를 결정하고 수로 저반위 설계를 위한 귀중한 참고자료를 제공한다.

Keywords

Figure Numerical study of the dam-break waves and Favre waves down sloped wet rigid-bed at laboratory scale
Figure Numerical study of the dam-break waves and Favre waves down sloped wet rigid-bed at laboratory scale

Dam-break flow, Bed slope, Wet bed, Velocity profile, Bed shear stress, Large eddy simulation
댐파괴유동, 하상경사, 습상, 유속분포, 하상전단응력, 대와류 시뮬레이션

Dam-Break Flows: Comparison between Flow-3D, MIKE 3 FM, and Analytical Solutions with Experimental Data

Dam-Break Flows: Comparison between Flow-3D, MIKE 3 FM, and Analytical Solutions with Experimental Data

by Hui Hu,Jianfeng Zhang andTao Li *
State Key Laboratory Base of Eco-Hydraulic Engineering in Arid Area, School of Water Resources and Hydropower, Xi’an University of Technology, Xi’an 710048, China
*Author to whom correspondence should be addressed.
Appl. Sci.20188(12), 2456; https://doi.org/10.3390/app8122456Received: 14 October 2018 /
Revised: 20 November 2018 / Accepted: 29 November 2018 / Published: 2 December 2018

Abstract

The objective of this study was to evaluate the applicability of a flow model with different numbers of spatial dimensions in a hydraulic features solution, with parameters such a free surface profile, water depth variations, and averaged velocity evolution in a dam-break under dry and wet bed conditions with different tailwater depths. Two similar three-dimensional (3D) hydrodynamic models (Flow-3D and MIKE 3 FM) were studied in a dam-break simulation by performing a comparison with published experimental data and the one-dimensional (1D) analytical solution. The results indicate that the Flow-3D model better captures the free surface profile of wavefronts for dry and wet beds than other methods. The MIKE 3 FM model also replicated the free surface profiles well, but it underestimated them during the initial stage under wet-bed conditions. However, it provided a better approach to the measurements over time. Measured and simulated water depth variations and velocity variations demonstrate that both of the 3D models predict the dam-break flow with a reasonable estimation and a root mean square error (RMSE) lower than 0.04, while the MIKE 3 FM had a small memory footprint and the computational time of this model was 24 times faster than that of the Flow-3D. Therefore, the MIKE 3 FM model is recommended for computations involving real-life dam-break problems in large domains, leaving the Flow-3D model for fine calculations in which knowledge of the 3D flow structure is required. The 1D analytical solution was only effective for the dam-break wave propagations along the initially dry bed, and its applicability was fairly limited. 

Keywords: dam breakFlow-3DMIKE 3 FM1D Ritter’s analytical solution

이 연구의 목적은 자유 표면 프로파일, 수심 변화 및 건식 및 댐 파괴에서 평균 속도 변화와 같은 매개 변수를 사용하여 유압 기능 솔루션에서 서로 다른 수의 공간 치수를 가진 유동 모델의 적용 가능성을 평가하는 것이었습니다.

테일 워터 깊이가 다른 습식베드 조건. 2 개의 유사한 3 차원 (3D) 유체 역학 모델 (Flow-3D 및 MIKE 3 FM)이 게시된 실험 데이터와 1 차원 (1D) 분석 솔루션과의 비교를 수행하여 댐 브레이크 시뮬레이션에서 연구되었습니다.

결과는 FLOW-3D 모델이 다른 방법보다 건식 및 습식 베드에 대한 파면의 자유 표면 프로파일을 더 잘 포착함을 나타냅니다. MIKE 3 FM 모델도 자유 표면 프로파일을 잘 복제했지만, 습식 조건에서 초기 단계에서 과소 평가했습니다. 그러나 시간이 지남에 따라 측정에 더 나은 접근 방식을 제공했습니다.

측정 및 시뮬레이션 된 수심 변화와 속도 변화는 두 3D 모델 모두 합리적인 추정치와 0.04보다 낮은 RMSE (root mean square error)로 댐 브레이크 흐름을 예측하는 반면 MIKE 3 FM은 메모리 공간이 적고 이 모델의 계산 시간은 Flow-3D보다 24 배 더 빠릅니다.

따라서 MIKE 3 FM 모델은 대규모 도메인의 실제 댐 브레이크 문제와 관련된 계산에 권장되며 3D 흐름 구조에 대한 지식이 필요한 미세 계산을 위해 Flow-3D 모델을 남겨 둡니다. 1D 분석 솔루션은 초기 건조 층을 따라 전파되는 댐 파괴에만 효과적이었으며 그 적용 가능성은 상당히 제한적이었습니다.

1. Introduction

저수지에 저장된 물의 통제되지 않은 방류[1]로 인해 댐 붕괴와 그로 인해 하류에서 발생할 수 있는 잠재적 홍수로 인해 큰 자연 위험이 발생한다. 이러한 영향을 최대한 완화하기 위해서는 홍수[2]로 인한 위험을 관리하고 감소시키기 위해 홍수의 시간적 및 공간적 진화를 모두 포착하여 댐 붕괴 파동의 움직임을 예측하고 댐 붕괴 파동의 전파 과정 효과를 다운스트림[3]으로 예측하는 것이 중요하다. 

그러나 이러한 수량을 예측하는 것은 어려운 일이며, 댐 붕괴 홍수의 움직임을 정확하게 시뮬레이션하고 유동장에 대한 유용한 정보를 제공하기 위한 적절한 모델을 선택하는 것은 그러므로 필수적인 단계[4]이다.

적절한 수학적 및 수치적 모델의 선택은 댐 붕괴 홍수 분석에서 매우 중요한 것으로 나타났다.분석적 해결책에서 행해진 댐 붕괴 흐름에 대한 연구는 100여 년 전에 시작되었다. 

리터[5]는 먼저 건조한 침대 위에 1D de 생베넌트 방정식의 초기 분석 솔루션을 도출했고, 드레슬러[6,7]와 휘담[8]은 마찰저항의 영향을 받은 파동학을 연구했으며, 스토커[9]는 젖은 침대를 위한 1D 댐 붕괴 문제에 리터의 솔루션을 확장했다. 

마샬과 멩데즈[10]는 고두노프가 가스 역학의 오일러 방정식을 위해 개발한 방법론[11]을 적용하여 젖은 침대 조건에서 리만 문제를 해결하기 위한 일반적인 절차를 고안했다. Toro [12]는 습식 및 건식 침대 조건을 모두 해결하기 위해 완전한 1D 정밀 리만 용해제를 실시했다. 

Chanson [13]은 특성 방법을 사용하여 갑작스러운 댐 붕괴로 인한 홍수에 대한 간단한 분석 솔루션을 연구했다. 그러나 이러한 분석 솔루션은 특히 댐 붕괴 초기 단계에서 젖은 침대의 정확한 결과를 도출하지 못했다[14,15].과거 연구의 발전은 이른바 댐 붕괴 홍수 문제 해결을 위한 여러 수치 모델[16]을 제공했으며, 헥-라스, DAMBRK, MIK 11 등과 같은 1차원 모델을 댐 붕괴 홍수를 모델링하는 데 사용하였다.

[17 2차원(2D) 깊이 평균 방정식도 댐 붕괴 흐름 문제를 시뮬레이션하는 데 널리 사용되어 왔으며[18,19,20,21,22] 그 결과 얕은 물 방정식(SWE)이 유체 흐름을 나타내는 데 적합하다는 것을 알 수 있다. 그러나, 경우에 따라 2D 수치해결기가 제공하는 해결책이 특히 근거리 분야에서 실험과 일관되지 않을 수 있다[23,24]. 더욱이, 1차원 및 2차원 모델은 3차원 현상에 대한 일부 세부사항을 포착하는 데 한계가 있다.

[25]. RANS(Reynolds-averageed Navier-Stok크스 방정식)에 기초한 여러 3차원(3D) 모델이 얕은 물 모델의 일부 단점을 극복하기 위해 적용되었으며, 댐 붕괴 초기 단계에서의 복잡한 흐름의 실제 동작을 이해하기 위해 사용되었다 [26,27,28]장애물이나 바닥 실에 대한 파장의 충격으로 인한 튜디 댐 붕괴 흐름 [19,29] 및 근거리 영역의 난류 댐 붕괴 흐름 거동 [4] 최근 상용화된 수치 모델 중 잘 알려진 유체 방식(VOF) 기반 CFD 모델링 소프트웨어 FLOW-3D는 컴퓨터 기술의 진보에 따른 계산력 증가로 인해 불안정한 자유 표면 흐름을 분석하는 데 널리 사용되고 있다. 

이 소프트웨어는 유한 차이 근사치를 사용하여 RANS 방정식에 대한 수치 해결책을 계산하며, 자유 표면을 추적하기 위해 VOF를 사용한다 [30,31]; 댐 붕괴 흐름을 모델링하는 데 성공적으로 사용되었다 [32,33].그러나, 2D 얕은 물 모델을 사용하여 포착할 수 없는 공간과 시간에 걸친 댐 붕괴 흐름의 특정한 유압적 특성이 있다. 

실생활 현장 척도 시뮬레이션을 위한 완전한 3D Navier-Stokes 방정식의 적용은 더 높은 계산 비용[34]을 가지고 있으며, 원하는 결과는 얕은 물 모델[35]보다 더 정확한 결과를 산출하지 못할 수 있다. 따라서, 본 논문은 3D 모델의 기능과 그 계산 효율을 평가하기 위해 댐 붕괴 흐름 시뮬레이션을 위한 단순화된 3D 모델-MIKE 3 FM을 시도한다. 

MIK 3 모델은 자연 용수 분지의 여러 유체 역학 시뮬레이션 조사에 적용되었다. 보치 외 연구진이 사용해 왔다. [36], 니콜라오스 및 게오르기오스 [37], 고얄과 라토드[38] 등 현장 연구에서 유체역학 시뮬레이션을 위한 것이다. 이러한 저자들의 상당한 연구에도 불구하고, MIK 3 FM을 이용한 댐 붕괴의 모델링에 관한 연구는 거의 없었다. 

또한 댐 붕괴 홍수 전파 문제를 해결하기 위한 3D 얕은 물과 완전한 3D RANS 모델의 성능을 비교한 연구도 아직 보고되지 않았다. 이 공백을 메우기 위해 현재 연구의 주요 목표는 댐 붕괴 흐름을 시뮬레이션하기 위한 단순화된 3D SWE, 상세 RANS 모델 및 분석 솔루션을 평가하여 댐 붕괴 문제에 대한 정확도와 적용 가능성을 평가하는 것이다.실제 댐 붕괴 문제를 해결하기 위해 유체역학 시뮬레이션을 시도하기 전에 수치 모델을 검증할 필요가 있다. 

일련의 실험 벤치마크를 사용하여 수치 모델을 확인하는 것은 용인된 관행이다. 현장 데이터 확보가 어려워 최근 몇 년 동안 제한된 측정 데이터를 취득했다. 

본 논문은 Ozmen-Cagatay와 Kocaman[30] 및 Khankandi 외 연구진이 제안한 두 가지 테스트 사례에 의해 제안된 검증에서 인용한 것이다. [39] 오즈멘-카가테이와 코카만[30]이 수행한 첫 번째 실험에서, 다른 미숫물 수위에 걸쳐 초기 단계 동안 댐 붕괴 홍수파가 발생했으며, 자유 지표면 프로파일의 측정치를 제공했다. Ozmen-Cagatay와 Kocaman[30]은 초기 단계에서 Flow-3D 소프트웨어가 포함된 2D SWE와 3D RANS의 숫자 솔루션에 의해 계산된 자유 표면 프로필만 비교했다. 

Khankandi 등이 고안한 두 번째 실험 동안. [39], 이 실험의 측정은 홍수 전파를 시뮬레이션하고 측정된 데이터를 제공하는 것을 목적으로 하는 수치 모델을 검증하기 위해 사용되었으며, 말기 동안의 자유 표면 프로필, 수위의 시간 진화 및 속도 변화를 포함한다. Khankandi 등의 연구. [39] 주로 실험 조사에 초점을 맞추었으며, 초기 단계에서는 리터의 솔루션과의 수위만을 언급하고 있다.

경계 조건(상류 및 하류 모두 무한 채널 길이를 갖는 1D 분석 솔루션에서는 실험 결과를 리터와 비교하는 것이 타당하지 않기 때문이다(건조 be)d) 또는 스토커(웨트 베드) 솔루션은 벽의 반사가 깊이 프로파일에 영향을 미쳤을 때, 그리고 참조 [39]의 실험에 대한 수치 시뮬레이션과의 추가 비교가 불량할 때. 이 논문은 이러한 문제를 직접 겨냥하여 전체 댐 붕괴 과정에서의 자유 표면 프로필, 수심 변화 및 속도 변화에 대한 완전한 비교 연구를 제시한다. 

여기서 댐 붕괴파의 수치 시뮬레이션은 초기에 건조하고 습한 직사각형 채널을 가진 유한 저장소의 순간 댐 붕괴에 대해 두 개의 3D 모델을 사용하여 개발된다.본 논문은 다음과 같이 정리되어 있다. 두 모델에 대한 통치 방정식은 숫자 체계를 설명하기 전에 먼저 도입된다. 

일반적인 단순화된 시험 사례는 3D 수치 모델과 1D 분석 솔루션을 사용하여 시뮬레이션했다. 모델 결과와 이들이 실험실 실험과 비교하는 방법이 논의되고, 서로 다른 수심비에서 시간에 따른 유압 요소의 변동에 대한 시뮬레이션 결과가 결론을 도출하기 전에 제시된다.

2. Materials and Methods

2.1. Data

첫째, 수평 건조 및 습식 침상에 대한 초기 댐 붕괴 단계 동안의 자유 표면 프로필 측정은 Ozmen-Cagatay와 Kocaman에 의해 수행되었다[30]. 이 시험 동안, 매끄럽고 직사각형의 수평 채널은 그림 1에서 표시한 대로 너비 0.30m, 높이 0.30m, 길이 8.9m이었다. 

채널은 채널 입구에서 4.65m 떨어진 수직 플레이트(담) 즉, 저장소의 길이 L0=4.65mL0에 의해 분리되었다., 및 다운스트림 채널 L1=4.25 mL1. m저수지는 댐의 좌측에 위치하고 처음에는 침수된 것으로 간주되었다; 저수지의 초기 상류 수심 h0 0.25m로 일정했다.

오른쪽의 초기 수심 h1h1 건식침대의 경우 0m, 습식침대의 경우 0.025m, 0.1m이므로 수심비 α=h1/h0α으로 세 가지 상황이 있었다. 0, 0.1, 0.4의 습식침대 조건은 플룸 끝에 낮은 보를 사용함으로써 만들어졌다. 물 표면 프로필은 3개의 고속 디지털 카메라(50프레임/s)를 사용하여 초기에 관찰되었으며, 계측 측정의 정확도는 참고문헌 [30]에서 입증되었다. In the following section, the corresponding numerical results refer to positions x = −1 m (P1), −0.5 m (P2), −0.2 m (P3), +0.2 m (P4), +0.5 m (P5), +1 m (P6), +2 m (P7), and +2.85 m (P8), where the origin of the coordinate system x = 0 is at the dam site. 3수심비 ααα 0, 0.1, 0.4의 경우 x,yx의 경우 좌표는 h0.으로 정규화된다.

<중략> ……

Figure 1. Schematic view of the experimental conditions by Ozmen-Cagatay and Kocaman [30]: (a) α = 0; (b) α = 0.1; and (c) α = 0.4.
Figure 1. Schematic view of the experimental conditions by Ozmen-Cagatay and Kocaman [30]: (a) α = 0; (b) α = 0.1; and (c) α = 0.4.

Figure 2. Schematic view of the experimental conditions by Khankandi et al. [39]: (a) α = 0 and (b) α = 0.2.
Figure 2. Schematic view of the experimental conditions by Khankandi et al. [39]: (a) α = 0 and (b) α = 0.2.
Figure 3. Typical profiles of the dam-break flow regimes for Stoker’s analytical solution [9]: Wet-bed downstream
Figure 3. Typical profiles of the dam-break flow regimes for Stoker’s analytical solution [9]: Wet-bed downstream
Figure 4. Sensitivity analysis of the numerical simulation using Flow-3D for the different mesh sizes of the experiments in Reference [30].
Figure 4. Sensitivity analysis of the numerical simulation using Flow-3D for the different mesh sizes of the experiments in Reference [30].
Figure 5. Sensitivity analysis of the numerical simulation using MIKE 3 FM for the different mesh sizes of the experiments in Reference [30].
Figure 5. Sensitivity analysis of the numerical simulation using MIKE 3 FM for the different mesh sizes of the experiments in Reference [30].
Figure 6. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for dry-bed (α=0). The experimental data are from Reference [30].
Figure 6. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for dry-bed (α=0). The experimental data are from Reference [30].
Figure 7. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for a wet-bed (α = 0.1). The experimental data are from Reference [30].
Figure 7. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for a wet-bed (α = 0.1). The experimental data are from Reference [30].
Figure 8. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for the wet-bed (α = 0.4). The experimental data are from Reference [30].
Figure 8. Comparison between observed and simulated free surface profiles at dimensionless times T = t(g/h0)1/2 and for the wet-bed (α = 0.4). The experimental data are from Reference [30].
Figure 9. Experimental and numerical comparison of free surface profiles h/h0(x/h0) during late stages at various dimensionless times T after the failure in the dry-bed by Khankandi et al. [39].
Figure 9. Experimental and numerical comparison of free surface profiles h/h0(x/h0) during late stages at various dimensionless times T after the failure in the dry-bed by Khankandi et al. [39].

Table 2. RMSE values for the free surface profiles observed by Khankandi et al. [39].

Table 2. RMSE values for the free surface profiles observed by Khankandi et al. [39].
Table 2. RMSE values for the free surface profiles observed by Khankandi et al. [39].
Figure 10. Measured and computed water level hydrograph at various positions for dry-bed by Khankandi et al. [39]: (a) G1 (−0.5 m); (b) G2 (−0.1 m); (c) G3 (0.1 m); (d) G4 (0.8 m); (e) G6 (1.2 m); (f) G8 (5.5 m).
Figure 10. Measured and computed water level hydrograph at various positions for dry-bed by Khankandi et al. [39]: (a) G1 (−0.5 m); (b) G2 (−0.1 m); (c) G3 (0.1 m); (d) G4 (0.8 m); (e) G6 (1.2 m); (f) G8 (5.5 m).
Figure 11. Measured and computed water level hydrographs at various positions for the wet-bed by Khankandi et al. [39]: (a) G1 (−0.5 m); (b) G2 (−0.1 m); (c) G4 (0.8 m); and (d) G5 (1.0 m).
Figure 11. Measured and computed water level hydrographs at various positions for the wet-bed by Khankandi et al. [39]: (a) G1 (−0.5 m); (b) G2 (−0.1 m); (c) G4 (0.8 m); and (d) G5 (1.0 m).

Table 3. RMSE values for the water depth variations observed by Khankandi et al. [39] at the late stage.

Table 3. RMSE values for the water depth variations observed by Khankandi et al. [39] at the late stage.
Table 3. RMSE values for the water depth variations observed by Khankandi et al. [39] at the late stage.
Figure 13. Comparison of simulated velocity profiles at various locations upstream and downstream of the dam at t = 0.8 s, 2 s, and 5 s for water depth ratios α = 0.1 by Ozmen-Cagatay and Kocaman [30]: (a) P1(−1 m); (b) P3 (+0.2 m); (c) P5 (+1 m); and (d) P6 (+2 m).
Figure 13. Comparison of simulated velocity profiles at various locations upstream and downstream of the dam at t = 0.8 s, 2 s, and 5 s for water depth ratios α = 0.1 by Ozmen-Cagatay and Kocaman [30]: (a) P1(−1 m); (b) P3 (+0.2 m); (c) P5 (+1 m); and (d) P6 (+2 m).
Table 5. The required computational time for the two models to address dam break flows in all cases
Table 5. The required computational time for the two models to address dam break flows in all cases

References

  1. Gallegos, H.A.; Schubert, J.E.; Sanders, B.F. Two-dimensional high-resolution modeling of urban dam-break flooding: A case study of Baldwin Hills, California. Adv. Water Resour. 200932, 1323–1335. [Google Scholar] [CrossRef]
  2. Kim, K.S. A Mesh-Free Particle Method for Simulation of Mobile-Bed Behavior Induced by Dam Break. Appl. Sci. 20188, 1070. [Google Scholar] [CrossRef]
  3. Robb, D.M.; Vasquez, J.A. Numerical simulation of dam-break flows using depth-averaged hydrodynamic and three-dimensional CFD models. In Proceedings of the Canadian Society for Civil Engineering Hydrotechnical Conference, Québec, QC, Canada, 21–24 July 2015. [Google Scholar]
  4. LaRocque, L.A.; Imran, J.; Chaudhry, M.H. 3D numerical simulation of partial breach dam-break flow using the LES and k-ε. J. Hydraul. Res. 201351, 145–157. [Google Scholar] [CrossRef]
  5. Ritter, A. Die Fortpflanzung der Wasserwellen (The propagation of water waves). Z. Ver. Dtsch. Ing. 189236, 947–954. [Google Scholar]
  6. Dressler, R.F. Hydraulic resistance effect upon the dam-break functions. J. Res. Nat. Bur. Stand. 195249, 217–225. [Google Scholar] [CrossRef]
  7. Dressler, R.F. Comparison of theories and experiments for the hydraulic dam-break wave. Int. Assoc. Sci. Hydrol. 195438, 319–328. [Google Scholar]
  8. Whitham, G.B. The effects of hydraulic resistance in the dam-break problem. Proc. R. Soc. Lond. 1955227A, 399–407. [Google Scholar] [CrossRef]
  9. Stoker, J.J. Water Waves: The Mathematical Theory with Applications; Wiley and Sons: New York, NY, USA, 1957; ISBN 0-471-57034-6. [Google Scholar]
  10. Marshall, G.; Méndez, R. Computational Aspects of the Random Choice Method for Shallow Water Equations. J. Comput. Phys. 198139, 1–21. [Google Scholar] [CrossRef]
  11. Godunov, S.K. Finite Difference Methods for the Computation of Discontinuous Solutions of the Equations of Fluid Dynamics. Math. Sb. 195947, 271–306. [Google Scholar]
  12. Toro, E.F. Shock-Capturing Methods for Free-Surface Shallow Flows; Wiley and Sons Ltd.: New York, NY, USA, 2001. [Google Scholar]
  13. Chanson, H. Application of the method of characteristics to the dam break wave problem. J. Hydraul. Res. 200947, 41–49. [Google Scholar] [CrossRef][Green Version]
  14. Cagatay, H.; Kocaman, S. Experimental Study of Tail Water Level Effects on Dam-Break Flood Wave Propagation; 2008 Kubaba Congress Department and Travel Services: Ankara, Turkey, 2008; pp. 635–644. [Google Scholar]
  15. Stansby, P.K.; Chegini, A.; Barnes, T.C.D. The initial stages of dam-break flow. J. Fluid Mech. 1998374, 407–424. [Google Scholar] [CrossRef]
  16. Soares-Frazao, S.; Zech, Y. Dam Break in Channels with 90° Bend. J. Hydraul. Eng. 2002128, 956–968. [Google Scholar] [CrossRef]
  17. Zolghadr, M.; Hashemi, M.R.; Zomorodian, S.M.A. Assessment of MIKE21 model in dam and dike-break simulation. IJST-Trans. Mech. Eng. 201135, 247–262. [Google Scholar]
  18. Bukreev, V.I.; Gusev, A.V. Initial stage of the generation of dam-break waves. Dokl. Phys. 200550, 200–203. [Google Scholar] [CrossRef]
  19. Soares-Frazao, S.; Noel, B.; Zech, Y. Experiments of dam-break flow in the presence of obstacles. Proc. River Flow 20042, 911–918. [Google Scholar]
  20. Aureli, F.; Maranzoni, A.; Mignosa, P.; Ziveri, C. Dambreak flows: Acquisition of experimental data through an imaging technique and 2D numerical modelling. J. Hydraul. Eng. 2008134, 1089–1101. [Google Scholar] [CrossRef]
  21. Rehman, K.; Cho, Y.S. Bed Evolution under Rapidly Varying Flows by a New Method for Wave Speed Estimation. Water 20168, 212. [Google Scholar] [CrossRef]
  22. Wu, G.F.; Yang, Z.H.; Zhang, K.F.; Dong, P.; Lin, Y.T. A Non-Equilibrium Sediment Transport Model for Dam Break Flow over Moveable Bed Based on Non-Uniform Rectangular Mesh. Water 201810, 616. [Google Scholar] [CrossRef]
  23. Ferrari, A.; Fraccarollo, L.; Dumbser, M.; Toro, E.F.; Armanini, A. Three-dimensional flow evolution after a dam break. J. Fluid Mech. 2010663, 456–477. [Google Scholar] [CrossRef]
  24. Liang, D. Evaluating shallow water assumptions in dam-break flows. Proc. Inst. Civ. Eng. Water Manag. 2010163, 227–237. [Google Scholar] [CrossRef]
  25. Biscarini, C.; Francesco, S.D.; Manciola, P. CFD modelling approach for dam break flow studies. Hydrol. Earth Syst. Sci. 201014, 705–718. [Google Scholar] [CrossRef][Green Version]
  26. Oertel, M.; Bung, D.B. Initial stage of two-dimensional dam-break waves: Laboratory versus VOF. J. Hydraul. Res. 201250, 89–97. [Google Scholar] [CrossRef]
  27. Quecedo, M.; Pastor, M.; Herreros, M.I.; Merodo, J.A.F.; Zhang, Q. Comparison of two mathematical models for solving the dam break problem using the FEM method. Comput. Method Appl. Mech. Eng. 2005194, 3984–4005. [Google Scholar] [CrossRef]
  28. Shigematsu, T.; Liu, P.L.F.; Oda, K. Numerical modeling of the initial stages of dam-break waves. J. Hydraul. Res. 200442, 183–195. [Google Scholar] [CrossRef]
  29. Soares-Frazao, S. Experiments of dam-break wave over a triangular bottom sill. J. Hydraul. Res. 200745, 19–26. [Google Scholar] [CrossRef]
  30. Ozmen-Cagatay, H.; Kocaman, S. Dam-break flows during initial stage using SWE and RANS approaches. J. Hydraul. Res. 201048, 603–611. [Google Scholar] [CrossRef]
  31. Vasquez, J.; Roncal, J. Testing River2D and FLOW-3D for Sudden Dam-Break Flow Simulations. In Proceedings of the Canadian Dam Association’s 2009 Annual Conference: Protecting People, Property and the Environment, Whistler, BC, Canada, 3–8 October 2009. [Google Scholar]
  32. Ozmen-Cagatay, H.; Kocaman, S. Dam-break flow in the presence of obstacle: Experiment and CFD simulation. Eng. Appl. Comput. Fluid 20115, 541–552. [Google Scholar] [CrossRef]
  33. Ozmen-Cagatay, H.; Kocaman, S.; Guzel, H. Investigation of dam-break flood waves in a dry channel with a hump. J. Hydro-Environ. Res. 20148, 304–315. [Google Scholar] [CrossRef]
  34. Gu, S.L.; Zheng, S.P.; Ren, L.Q.; Xie, H.W.; Huang, Y.F.; Wei, J.H.; Shao, S.D. SWE-SPHysics Simulation of Dam Break Flows at South-Gate Gorges Reservoir. Water 20179, 387. [Google Scholar] [CrossRef]
  35. Evangelista, S. Experiments and Numerical Simulations of Dike Erosion due to a Wave Impact. Water 20157, 5831–5848. [Google Scholar] [CrossRef][Green Version]
  36. Bocci, M.; Chiarlo, R.; De Nat, L.; Fanelli, A.; Petersen, O.; Sorensen, J.T.; Friss-Christensen, A. Modelling of impacts from a long sea outfall outside of the Venice Lagoon (Italy). In Proceedings of the MWWD—IEMES 2006 Conference, Antalya, Turkey, 6–10 November 2006; MWWD Organization: Antalya, Turkey, 2006. [Google Scholar]
  37. Nikolaos, T.F.; Georgios, M.H. Three-dimensional numerical simulation of wind-induced barotropic circulation in the Gulf of Patras. Ocean Eng. 201037, 355–364. [Google Scholar]
  38. Goyal, R.; Rathod, P. Hydrodynamic Modelling for Salinity of Singapore Strait and Johor Strait using MIKE 3FM. In Proceedings of the 2011 2nd International Conference on Environmental Science and Development, Singapore, 26–28 February 2011. [Google Scholar]
  39. Khankandi, A.F.; Tahershamsi, A.; Soares-Frazão, S. Experimental investigation of reservoir geometry effect on dam-break flow. J. Hydraul. Res. 201250, 376–387. [Google Scholar] [CrossRef]
  40. Flow Science Inc. FLOW-3D User’s Manuals; Flow Science Inc.: Santa Fe, NM, USA, 2007. [Google Scholar]
  41. Danish Hydraulic Institute (DHI). MIKE 3 Flow Model FM. Hydrodynamic Module-User Guide; DHI: Horsholm, Denmark, 2014. [Google Scholar]
  42. Pilotti, M.; Tomirotti, M.; Valerio, G. Simplified Method for the Characterization of the Hydrograph following a Sudden Partial Dam Break. J. Hydraul. Eng. 2010136, 693–704. [Google Scholar] [CrossRef]
  43. Hooshyaripor, F.; Tahershamsi, A.; Razi, S. Dam break flood wave under different reservoir’s capacities and lengths. Sādhanā 201742, 1557–1569. [Google Scholar] [CrossRef]
  44. Kocaman, S.; Ozmen-Cagatay, H. Investigation of dam-break induced shock waves impact on a vertical Wall. J. Hydrol. 2015525, 1–12. [Google Scholar] [CrossRef]
  45. Liu, H.; Liu, H.J.; Guo, L.H.; Lu, S.X. Experimental Study on the Dam-Break Hydrographs at the Gate Location. J. Ocean Univ. China 201716, 697–702. [Google Scholar] [CrossRef]
  46. Marra, D.; Earl, T.; Ancey, C. Experimental Investigations of Dam Break Flows down an Inclined Channel. In Proceedings of the 34th World Congress of the International Association for Hydro- Environment Research and Engineering: 33rd Hydrology and Water Resources Symposium and 10th Conference on Hydraulics in Water Engineering, Brisbane, Australia, 26 June–1 July 2011. [Google Scholar]
  47. Wang, J.; Liang, D.F.; Zhang, J.X.; Xiao, Y. Comparison between shallow water and Boussinesq models for predicting cascading dam-break flows. Nat. Hazards 201683, 327–343. [Google Scholar] [CrossRef]
  48. Yang, C.; Lin, B.L.; Jiang, C.B.; Liu, Y. Predicting near-field dam-break flow and impact force using a 3D model. J. Hydraul. Res. 201048, 784–792. [Google Scholar] [CrossRef]
Figure 6. Maximum inundation field in simulations with (a) no barrier on the seawall (red line), (b) a 1 m barrier across the entire sea wall, and (c) a 1.7 m barrier partially installed on the seawall.

Storm surge inundation simulations comparing three-dimensional with two-dimensional models based on Typhoon Maemi over Masan Bay of South Korea

Jae-Seol Shim†, Jinah Kim†, Dong-Chul Kim‡, Kiyoung Heo†, Kideok Do†, Sun-Jung Park ‡
† Coastal Disaster Research Center,
Korea Institute of Ocean Science &
Technology, 426-744, Ansan, Gyeonggi,
Korea
jsshim@kiost.ac
jakim@kiost.ac
kyheo21@kiost.ac
kddo@kiost.ac
‡ Technology R&D Institute
Hyein E&C Co., Ltd., Seoul 157-861,
Korea
skkkdc@chol.com
Nayana_sj@nate.com

ABSTRACT

Shim, J., Kim, J., Kim, D., Heo, K., Do, K., Park, S., 2013. Storm surge inundation simulations comparing threedimensional with two-dimensional models based on Typhoon Maemi over Masan Bay of South Korea. In:
Conley, D.C., Masselink, G., Russell, P.E. and O’Hare, T.J. (eds.), Proceedings 12th International Coastal Symposium
(Plymouth, England), Journal of Coastal Research, Special Issue No. 65, pp. 392-397, ISSN 0749-0208.
Severe storm surge inundation was caused by the typhoon Maemi in Masan Bay, South Korea in September 2003. To
investigate the differences in the storm surge inundation simulated by three-dimensional (3D) and two-dimensional
models, we used the ADvanced CIRCulation model (ADCIRC) and 3D computational fluid dynamics (CFD) model
(FLOW3D). The simulation results were compared to the flood plain map of Masan Bay following the typhoon Maemi.
To improve the accuracy of FLOW3D, we used a high-resolution digital surface model with a few tens of centimeterresolution, produced by aerial LIDAR survey. Comparison of the results between ADCRIC and FLOW3D simulations shows that the inclusion of detailed information on buildings and topography has an impact, delaying seawater propagation and resulting in a reduced inundation depth and flooding area. Furthermore, we simulated the effect of the installation of a storm surge barrier on the storm surge inundation. The barrier acted to decrease the water volume of the inundation and delayed the arrival time of the storm surge, implying that the storm surge barrier provides more time for residents’ evacuation.

Keywords: Typhoon Maemi, digital surface elevation model, Reynolds-Averaged NavierStokes equations.

2003 년 9 월 대한민국 마산만 태풍 매미에 의해 심한 폭풍 해일 침수가 발생했습니다. 3 차원 (3D) 및 2 차원 모델로 시뮬레이션 한 폭풍 해일 침수의 차이를 조사하기 위해 ADvanced CIRCulation 모델 ( ADCIRC) 및 3D 전산 유체 역학 (CFD) 모델 (FLOW3D).

시뮬레이션 결과는 태풍 매미 이후 마산만 범람원 지도와 비교되었다. FLOW-3D의 정확도를 높이기 위해 우리는 항공 LIDAR 측량으로 생성된 수십 센티미터 해상도의 고해상도 디지털 표면 모델을 사용했습니다.

ADCRIC과 FLOW3D 시뮬레이션의 결과를 비교하면 건물과 지형에 대한 자세한 정보를 포함하면 해수 전파가 지연되고 침수 깊이와 침수 면적이 감소하는 것으로 나타났습니다.

또한, 폭풍 해일 침수에 대한 폭풍 해일 장벽 설치의 효과를 시뮬레이션했습니다. 이 장벽은 침수 물량을 줄이고 폭풍 해일 도착 시간을 지연시키는 역할을 하여 폭풍 해일 장벽이 주민들의 대피에 더 많은 시간을 제공한다는 것을 의미합니다.

INTRODUCTION

2003 년 9 월 12 일 태풍 매미로 인한 강한 폭풍 해일이 남해안을 강타했습니다. 마산 만 일대는 심한 폭풍우 침수로 인해 최악의 피해를 입었고 광범위한 홍수를 겪었습니다. 따라서 마산 만에 예방 체계를 구축하기 위해 폭풍 해일에 의한 침수에 대한 수치 예측을 시도하는 선행 연구가 수행되었다 (Park et al. 2011).

그러나 일반적인 2 차원 (2D) 또는 3 차원 (3D) 수압 가정을 사용할 때 지형의 해상도는 복잡한 해안 구조를 표현하기에 충분하지 않습니다. 따라서 우리는 마산 만의 고해상도 지형도를 통해 전산 유체 역학 (CFD)의 침수 시뮬레이션을 제시한다.

태풍 매미는 2003 년 9 월 12 일 12시 (UTC)에 한반도에 상륙하여 남동부 해안을 따라 추적했습니다 (그림 1). 2003 년 9 월 13 일 6시 (UTC)에 동 일본해로 이동하여 온대 저기압이되었습니다.

풍속과 기압면에서 한국을 강타한 가장 강력한 태풍 중 하나입니다. 특히 마산 만에 접해있는 마산시는 폭풍 해일 홍수로 최악의 피해를 입어 32 명이 사망하고 심각한 해안 피해를 입었다. 태풍이 지나가는 동안 중앙 기압은 950hPa, 진행 속도는 45kmh-1로 마산항의 조 위계를 통해 최대 약 2.3m의 서지 높이를 기록했다.

마산 만에 접한 주거 및 상업 지역은 홍수가 심했고 지하 시설은 폭풍 해일로 침수로 어려움을 겪었습니다 (Yasuda et al. 2005). 이 논문에서는 3D CFD 모델 (FLOW 3D)과 2D ADvanced CIRCulation 모델 (ADCIRC)을 사용하여 기록 된 마산 만에서 가장 큰 폭풍 해일 중 하나에 의해 생성 된 해안 침수를 시뮬레이션했습니다.

건물의 높이와 공간 정보를 포함하는 디지털 표면 모델 (DSM)은 LiDAR (Airborne Light Detection and Ranging)에 의해 만들어졌으며, 폭풍 해일 침수 모델, 즉 3D CFD 모델 (FLOW 3D)의 입력 데이터로 사용되었습니다. ). 또한 ADCIRC의 시뮬레이션 결과는 FLOW3D의 경계 조건으로 사용됩니다.

본 연구의 목적은 극심한 침수 높이와 해안 육지로의 범람을 포함하여 마산 만에서 태풍 매미로 인한 폭풍 해일 침수를 재현하는 것이다.

<중략>………………

Figure 1. The best track and the central pressures of the typhoon Maemi from the Joint Typhoon Warning Center (JTWC). Open circles indicate the locations of the typhoon in 3 h intervals. Filled circles represent locations of the cited stations; A, B, C and D indicate Jeju, Yeosu, Tongyoung, and Masan, respectively.
Figure 1. The best track and the central pressures of the typhoon Maemi from the Joint Typhoon Warning Center (JTWC). Open circles indicate the locations of the typhoon in 3 h intervals. Filled circles represent locations of the cited stations; A, B, C and D indicate Jeju, Yeosu, Tongyoung, and Masan, respectively.
Figure 2. Model domain with FEM mesh for Typhoon Maemi.
Figure 2. Model domain with FEM mesh for Typhoon Maemi.
Figure 3. Validation of surge height for the four major tidal stations on the south coast of the Korea.
Figure 3. Validation of surge height for the four major tidal stations on the south coast of the Korea.
Figure 4. Inundation depth results from (a) ADCIRC, (b) FLOW3D, and (c) inundation field surveying hazard map following typhoon Maemi.
Figure 4. Inundation depth results from (a) ADCIRC, (b) FLOW3D, and (c) inundation field surveying hazard map following typhoon Maemi.
Figure 5. Inundation depth results computed by Flow3D at each time period following arrival of storm surge wave at harbor mouth.
Figure 5. Inundation depth results computed by Flow3D at each time period following arrival of storm surge wave at harbor mouth.
Figure 6. Maximum inundation field in simulations with (a) no barrier on the seawall (red line), (b) a 1 m barrier across the entire sea wall, and (c) a 1.7 m barrier partially installed on the seawall.
Figure 6. Maximum inundation field in simulations with (a) no barrier on the seawall (red line), (b) a 1 m barrier across the entire sea wall, and (c) a 1.7 m barrier partially installed on the seawall.

LITERATURE CITED

Bunya S, Kubatko EJ, Westerink JJ, Dawson C.,2010. A wetting and drying treatment for the Runge–Kutta discontinuous Galerkin solution to the shallow water equations. Computer Methods in Applied Mechanics and Engineering, Oceanography and Coastal Research, 198, 1548-1562.
Chan, J.C.L. & Shi, J.,1996. Long term trends and interannual variability in tropical cyclone activity over the western North Pacific. Geophysical Research Letters 23, 2765-2767.
Choi, B.H., Kim, D.C., Pelinovsky, E. and Woo, S.B., 2007. Threedimensional simulation of tsunami run-up around conical island. Coastal Engineering, 54, 618-629.
Choi, B.H., Pelinovsky, E., Kim, D.C., Didenkulova, I. and Woo, S.B., Two- and three-dimensional computation of solitary wave runup on non-plane beach. Nonlinear Processes in Geophysics, 15, 489-502.
Choi B.H., Pelinovsky E., Kim D.C., Lee H.J., Min B.I. and Kim K.H., Three-dimensional simulation of 1983 central East (Japan) Sea earthquake tsunami at the Imwon Port (Korea). Ocean Engineering, 35, 1545-1559.
Choi, B.H., Eum, H.M., Kim, H.S., Jeong, W.M. & Shim, J.S., 2004. Wave-tide-surge coupled simulation for typhoon Maemi, Workshop on waves and storm surges around Korean peninsula, 121-144.
Choi, K.S., & Kim, B.J., 2007. Climatological characteristics of tropical cyclone making landfall over the Korean Peninsula. Journal of the Korean Meteorological Society 43, 97-109.
Clark, J.D. & Chu, P., 2002. Interannual variation of tropical cyclone activity over the central North Pacific. Journal of the Meteorological Society of Japan, 80, 403-418.
Davies, A.M. & Flather, R.A., 1978. Application of numerical models of the North West European continental shelf and the North Sea to the computation of the storm surges of November to December 1973.
Deutsche Hydrographische Zeitschrift Ergänzungsheft Reihe A, 14, 72. Flow Science, 2010. FLOW-3D User’s Manual. Fujita, T., 1952. Pressure distribution in a typhoon. Geophysical Magazine 23.
Garratt, J.R., 1977. Review of drag coefficients over oceans and continents. Monthly Weather Review, 105, 915-929.
Gary Padgett, 2004. Gary Padgett September 2003 Tropical Weather Summary. Typhoon 2000.
Goda Y., Kishira Y. and Kamiyama Y., 1975. Laboratory investigation on the overtopping rate of seawalls by irregular waves, Report of Port and Harbour Research Inst.,14(4), 3-44.
Heaps, N.S., 1965. Storm surges on a continental shelf. Philos. Trans. R. Soc. London, Ser. 257, 351-383.
Hirt, C.W. and Nichols, B.D., 1981. Volume of fluid (VOF) method for the dynamics of free boundaries. Journal of Computational Physics, 39, 201-225.
Holland, G.J., 1980. An Analytic Model of the Wind and Pressure Profiles in Hurricanes. Monthly Weather Review, 108, 1212-1218.
Independent Levee Investigation Team, 2006. Investigation of the Performance of the New Orleans Flood Protection Systems in Hurricane Katrina on August 29, 2005
Klotzbach, P. J. , 2006. Trends in global tropical cyclone activity over the past twenty years (1986-2005). Geophysical Research Letters, 33.
Large, W.G. & Pond, S., 1981. Open ocean momentum flux measurements in moderate to strong winds. Journal of Physical Oceanography, 11, 324-336.
Landsea, C.W., Nicholls, N., Gray, W.M. & Avila, L.A., 1996. Downward trends in the frequency of intense Atlantic hurricanes during the past five decades. Geophysical Research Letters, 23, 1697-1700.
Lighthill, J., Holland, G., Gray, W., Landsea, C., Creig, G., Evans, J., Kurikara, Y. and Guard, C., 1994. Global climate change and tropical cyclones. Bulletin of the American Meteorological Society, 75, 2147- 2157.
Luettich, R.A. & Westerink, J.J., 2004. Formulation and Numerical Implementation of the 2D/3D ADCIRC finite element model version 44.XX.
Matsumoto, K., Takanezawa, T. & Ooe, M., 2000. Ocean tide models developed by assimilating TOPEX/POSEIDON altimeter data into hydrodynamical model: A global model and a regional model around Japan, Journal of Oceanography, 56(5) 567-581.
Mitsuyasu, H. and Kusaba, T., 1984. Drag Coefficient over Water Surface Under the Action of Strong Wind. Natural Disaster Science, 6, 43-50.
Mitsuyasu, H., F. Tasai, T. Suhara, S. Mizuno, M. Ohkusu, T. Honda and K. Rikiishi, 1980. Observation of the power spectrum of ocean waves using a cloverleaf buoy. Journal of Physical Oceanography, 10, 286- 296.
Multiple Lines of Defense Assessment Team, 2007. Comprehensive Recommendations Supporting the Use of the Multiple Lines of Defense Strategy to Sustain Coastal Louisiana.
Myers, V.A. and Malkin, W., 1961. Some Properties of Hurricane Wind Fields as Deduced from Trajectories. U.S. Weather Bureau, National Hurricane Research Project, Report 49.
Saito, K., T. Fujita, Y. Yamada, J. Ishida, Y. Kumagai, K. Aranami, S. Ohmori, R. Nagasawa, S. Kumagai, C. Muroi, T. Kato, H. Eito and Y. Yamazaki, 2006. The operational JMA Nonhydrostatic Mesoscale Model. Monthly Weather Review, 134, 1266-1298.
Shibaki H., Nakai K., Suzuyama K. and Watanabe A., 2004. Multi-level storm surge model incorporating density stratification and wave-setup. Proc. of 29th Int. Conf. on Coastal Eng., ASCE, 1539-1551.JSCE (1999). Hydraulic formulas, page 245 (in Japanese).
Shibaki, H., Suzuyama, K., Kim, J.I., & Sun, L., 2007. Numerical simulation of storm surge inundation induced by overflow, overtopping and dike breach. Asian and Pacific Coasts 2007, Nanjing, China.
Smagorinsky J., 1963. General circulation experiments with the primitive equations: I. The basic experiment. Monthly Weather Review, 91, 99- 164.
Smith, S.D. & Banke, E.G., 1975. Variation of the sea surface drag coefficient with wind speed. Quarterly Journal of the Royal Meteorological Society, 101, 665-673.
Versteeg, H.K., Malalasekera, W., 1995.An introduction to computational fluid dynamics. The Finite Volume Method. Prentice Hall, 257p.
Wang Xinian, Yin Qingjiang, Zhang Baoming, 1991. Research and Applications of a Forecasting Model of Typhoon Surges in China Seas. Advances In Water Science.
Wu, J., 1982. Wind-Stress Coefficients over Sea Surface from Breeze to Hurricane. Journal of Geophysical Research, 87, 9704-9706.
Yeh, H., Liu, P., Synolakis, C., 1996. Long-wave Runup Models. World Scientific.
Yakhot, V. and Orszag, S.A., 1986. Renormalization group analysis of turbulence, I. Basic theory. Journal of Scientific Computing, 1, 1-51.
Yakhot, V. and Smith, L.M., 1992. The renormalization group, the expansion and derivation of turbulence models, Journal of Scientific Computing, 7, 35-61
Yasuda, T., T. Hiraishi, H. Kawai, K. Nagase, S.W. Kang, and W.M. Jeong, 2005. Field survey and computation analysis of storm surge disaster in Masan due to Typhoon Maemi, Proceedings of Asian and Pacific Coasts 2005, Jeju, Korea.

Fig. 4 Current lines in the horizontal level in: a 0.70 and b 14 cm from the streambed in tandem pies

3D numerical simulation of flow field around twin piles

트윈 말뚝 주위의 유동장 3D 수치 시뮬레이션

Amini, A; Parto, AA
Amini, A (reprint author), AREEO, Kurdistan Agr & Nat Resources Res & Educ Ctr, Sanandaj, Iran.
, 2017; 65 (6): 1243

Abstract

이 연구에서는, 파일 그룹 주위의 흐름 패턴과 국소적 스크루 메커니즘을 식별하기 위해, 플로우 필드를 FLOW-3D 소프트웨어를 사용해 시뮬레이션했다. 편평한 침대 채널에 나란히 배열되어 있는 한 쌍의 말뚝이 조사되었다. Navier-Stokes 방정식을 확립하기 위해 RNGk-epsilon 난류 모델을 사용하였고 실험 데이터를 사용하여 결과를 검증하였다. FLOW-3D 기능의 경우, 소프트웨어가 파일 그룹 간의 예상 상호작용을 적절히 시뮬레이션할 수 있는 것으로 확인되었다. 플로우 필드 시뮬레이션 결과는 레이놀즈 숫자와 말뚝 간격이 vortices 형성에 가장 큰 영향을 미치는 변수라는 것을 보여주었다. 탠덤 더미 주변의 흐름과 웨이크 바이크 주변의 하향 흐름은 측면 배치와 단일 더미에 비해 더 강렬하고 복잡했다.

In this study to identify the flow pattern and local scour mechanism around pile groups, the flow field was simulated using FLOW-3D software. A pair of pile on a flat-bed channel with side by side and tandem arrangements was investigated. To establish Navier–Stokes equations, the RNGk-e turbulence model was used and the results were verified using experimental data. In case of FLOW-3D capability, it was found that the software was able to properly simulate the expected interaction between the pile groups. The results of flow field simulation showed that Reynolds number and the pile spacing are the most influential variables in forming vortices. The flow around tandem pile and the downward flow around wake vortices were more intense and complicate in comparison with side by side arrangements and single pile.

Keywords : Bridge, Sediment, Flow pattern, Pile group, Local scour

Fig. 1 General view of the channel and measured points a side by side b tandem arrangement
Fig. 1 General view of the channel and measured points a side by side b tandem arrangement
Fig. 2 Meshing around the two side by side piles: a plan and b side view
Fig. 2 Meshing around the two side by side piles: a plan and b side view
Fig. 3 Meshing around the two tandem piles: a plan and b side view
Fig. 3 Meshing around the two tandem piles: a plan and b side view
Fig. 4 Current lines in the horizontal level in: a 0.70 and b 14 cm from the streambed in tandem pies
Fig. 4 Current lines in the horizontal level in: a 0.70 and b 14 cm from the streambed in tandem pies
Fig. 5 Current lines in the horizontal level in: a 0.70 cm, and b 14 cm from the streambed in side by side piles
Fig. 5 Current lines in the horizontal level in: a 0.70 cm, and b 14 cm from the streambed in side by side piles
Fig. 6 Comparing iso-velocity line in longitudinal direction (u): a observed in 0.7 cm; b observed in 14 cm; c simulated in 0.7 cm and d simulated in 14 cm
Fig. 6 Comparing iso-velocity line in longitudinal direction (u): a observed in 0.7 cm; b observed in 14 cm; c simulated in 0.7 cm and d simulated in 14 cm
Fig. 7 Comparing iso-velocity line in latitudinal direction (v): a observed in 0.7 cm; b observed in 14 cm; c simulated in 0.7 cm and d simulated in 14 cm
Fig. 7 Comparing iso-velocity line in latitudinal direction (v): a observed in 0.7 cm; b observed in 14 cm; c simulated in 0.7 cm and d simulated in 14 cm
Fig. 8 3D velocity profiles in x–z plane in the center of the pile (Y = 0): a x = - 1.65D; b x = - 6.59D; c x = 0.69D; d x = 1.32D; e x = 3.69D and f x = 7.60D
Fig. 8 3D velocity profiles in x–z plane in the center of the pile (Y = 0): a x = – 1.65D; b x = – 6.59D; c x = 0.69D; d x = 1.32D; e x = 3.69D and f x = 7.60D
Fig. 9 Comparison of simulated and observed velocity in x–y plane in center of piles
Fig. 9 Comparison of simulated and observed velocity in x–y plane in center of piles

References

  • Akilli AA, Karakus C (2004) Flow characteristics of circular cylinders arranged side-by- side in shallow water. Flow Meas Instrum 15(4):187–189
  • Amini A, Mohammad TM (2017) Local scour prediction in complex pier. Mar Georesour Geotechnol 35(6):857–864
  • Amini A, Melville B, Thamer M, Halim G (2012) Clearwater local scour around pile groups in shallow-water flow. J Hydraul Eng (ASCE) 138(2):177–185
  • Amini A, Mohd TA, Ghazali H, Bujang H, Azlan A (2011) A local scour prediction method for pile cap in complex piers. ICE-water Manag. 164(2):73–80
  • Aslani A (2008) Experimental evaluation of flow pattern around double piles. MSc thesis, Sharif University, Tehran
  • Gu ZF, Sun TF (1999) On interference between two circular cylinders in staged arrangement at high sub-critical Reynolds numbers. J Wind Eng Ind Aerodyn 80:287–309
  • Hang-Wook P, Hyun P, Yang-Ki C (2014) Evaluation of the applicability of pier local scour formulae using laboratory and field data. Mar Georesour Geotechnol. https://doi.org/10.1080/ 1064119X.2014.954658
  • Hannah CR (1978) Scour at pile groups. Research Rep. No. 78-3, Civil Engineering, Univ. of Canterbury, Christchurch
  • Hosseini R, Amini A (2015) Scour depth estimation methods around pile groups. J Civ Eng KSCE 19(7):2144–2156
  • Lanca R, Fael C, Maia R, Peˆgo J, Cardoso A (2013) Clear-water scour at pile groups. J Hydraul Eng. ttps://doi.org/10.1061/ (ASCE)HY.1943-7900.0000770
  • Mohamed HI (2013) Numerical simulation of flow and local scour at two submerged-emergent tandem cylindrical piers. J Eng Sci 41(1):1–19
  • Palau-Salvador G, Stoesser T, Rodi W (2008) LES of the flow around two cylinders in tandem. J Fluids Struct 24(8):1304–1312
  • Papaionannou GV, Yuea DKP, Triantafylloua MS, Karniadakis GE (2008) On the effect of spacing on the vortex-induced vibrations of tandem cylinders. J Fluids Struct 24:833–854
  • Price SJ, Paidoussis MP (1989) The flow induced response of a single flexible cylinder in an in-line array of rigid cylinder. J Fluid Struct 3:61–82
  • Raudkivi AJ (1998) Loose boundary hydraulics. A. A. Balkema, Rotterdam, pp 8–28. https://doi.org/10.1080/02508069608686502
  • Salim MS, Cheah SC (2009) Wall y ? strategy for dealing with wallbounded turbulent flows. In: Proceedings of the international multiconference of engineers and computer scientists, vol II, IMECS, Hong Kong
  • Shin JH, Park HI (2010) Neural network formula for local scour at piers using field data. Mar Georesour Geotechnol 28(1):37–48
  • Sicilian JM, Hirt CW, Harper RP (1987) FLOW-3D. Computational modeling power for scientists and engineers. Report FSI-87-00-Flow Science. Los Alamos, NM
  • Solaimani N, Amini A, Banejad H, Taheri P (2017) The effect of pile spacing and arrangement on bed formation and scour hole dimensions in pile groups. Int J River Basin Manag 15(2):219–225
  • Sumer BM, Fredsøe J (2002) The mechanics of scour in the marine environment. World Scientific, Farrer Road, Singapore
  • Sumer B, Chua L, Cheng N, Fredsøe J (2003) Influence of turbulence on bed load sediment transport. J Hydraul Eng. https://doi.org/ 10.1061/(ASCE)0733-9429(2003)129:8(585)
  • Sun TF, Gu ZF, He DX, Zhang LL (1992) Fluctuating pressure on two circular cylinder at high Reynolds number. J Wind Eng Ind Aero. 42:577–588
Figure 7. Formation of incident and reflected waves.

Investigate Impact Force of Dam-Break Flow against Structures by Both 2D and 3D Numerical Simulations

2D 및 3D 수치 시뮬레이션에 의한 댐 붕괴유동의 구조물 충격력 조사

1 Faculty of Water Resources Engineering, Thuyloi University, 175 Tay Son, Dong Da, Ha Noi 116705, Vietnam
2 Hydraulic Construction Institute, 3/95 Chua Boc, Dong Da, Ha Noi 116705, Vietnam
* Author to whom correspondence should be addressed.
Academic Editor: Costanza Aricò
Water 2021, 13(3), 344;

Abstract

본 논문의 목적은 일부 2D 및 3D 수치 모델이 침수 지역에 고립된 건물 또는 건물 배열이 있는 곳에서 홍수 파동을 시뮬레이션하는 능력을 조사하는 것이었습니다.

먼저, 제안된 2D 수치 모델은 구조화된 메시에서 2D 얕은 물 방정식(2D-SWEs)을 해결하기 위한 유한 볼륨 방법(FVM)을 기반으로 했습니다.

FDS (flux-difference splitting)은 정확한 질량 균형을 얻기 위해 사용되었고 Roe 체계는 Riemann 문제를 근사하기 위해 호출되었습니다.

둘째, 상업적으로 이용 가능한 3D CFD 소프트웨어 패키지가 선택되었으며, 여기에는 두 가지 난류 모델이 포함된 Flow 3D 모델이 포함되어 있습니다.

RNG(Renormalized Group) 및 LES(Large-eddy Simulation)를 사용하는 레이놀즈 평균 Navier-Stokes(RAN)입니다. 댐 붕괴 흐름으로 인한 장애물에 대한 충격력의 수치 결과는 3D 솔루션이 2D 솔루션보다 훨씬 낫다는 것을 보여주었습니다.

건물 배열에 작용하는 충격력의 3D 수치 힘 결과를 보유하고 있는 실험 데이터와 비교함으로써, 속도 유도력이 동적 힘에 미치는 영향은 Froude 숫자의 함수와 사고 파동의 수심 함수에 의해 정량화 되었습니다. 또한, 우리는 힘의 강도의 피크 값의 3D 계산 결과에 대한 초기 물 단계와 댐 붕괴 폭의 영향을 조사했습니다.

The aim of this paper was to investigate the ability of some 2D and 3D numerical models to simulate flood waves in the presence of an isolated building or building array in an inundated area. Firstly, the proposed 2D numerical model was based on the finite-volume method (FVM) to solve 2D shallow-water equations (2D-SWEs) on structured mesh. The flux-difference splitting method (FDS) was utilized to obtain an exact mass balance while the Roe scheme was invoked to approximate Riemann problems. Secondly, the 3D commercially available CFD software package was selected, which contained a Flow 3D model with two turbulent models: Reynolds-averaged Navier-Stokes (RANs) with a renormalized group (RNG) and a large-eddy simulation (LES). The numerical results of an impact force on an obstruction due to a dam-break flow showed that a 3D solution was much better than a 2D one. By comparing the 3D numerical force results of an impact force acting on building arrays with the existence experimental data, the influence of velocity-induced force on a dynamic force was quantified by a function of the Froude number and the water depth of the incident wave. Furthermore, we investigated the effect of the initial water stage and dam-break width on the 3D-computed results of the peak value of force intensity.

Keywords: dam-break wave2D numerical modelFlow 3D modelstructuresimpact force

Introduction

홍수 위험 분석에 따른 도시 계획은 최근에 큰 연구 과제였습니다.

건물 또는 건물 그룹에 대한 홍수 파동의 영향에 대한 연구는 하류 지역에 대한 조기 경고 또는 안전 의식 향상에 중요한 역할을 했습니다. 기본적으로 댐 파괴 흐름에 대한 연구는 실험 측정이나 수치 시뮬레이션을 통해 추정 할 수 있습니다 [1,2,3,4,5,6].

컴퓨터 처리 능력의 증가로 인해 불연속 흐름에 대한 수치 연구가 비용 효율적이되었습니다. 지난 10 년 동안 얕은 물 솔버는 정확성과 계산 능력면에서 크게 향상되었습니다. 침수 가능 지역의 수심 및 속도 프로파일과 같은 유체 역학적 매개 변수에 많은주의를 기울였습니다 [1,2,3,4,5,6,7,8].

Migot et al. [9]는 도시 홍수의 실험적 모델링에 관한 많은 기사를 검토했습니다. 그 논문에 언급 된 45 개의 작품 중 단 4 개의 프로젝트 만이 장애물에 가해지는 일정한 또는 비정상적인 흐름의 힘 또는 압력을 측정했습니다.

또한 물리적 및 2D 수치 모델에서 건물 또는 건물 그룹에 돌발 홍수가 미치는 영향에 대한 연구는 거의 없었습니다. 얕은 물 모델은 [10,11]에서 고립된 장애물에 대한 충격의 힘을 예측하는데 사용되었습니다.

한편 Shige-eda [12]는 액체와 건물 배열 간의 상호 작용을 결정하기 위해 물리적 모델과 2D 수치 체계를 선택했습니다. Aureli와 Shige-eda는 수직 속도와 가속도를 무시하기 때문에 댐 파괴 흐름의 힘을 추정하기 위한 2D 얕은 물 방정식 (SWE)의 단점을 보여주었습니다 [10,12].

Migot [9]은 또한 장애물 주변의 시뮬레이션된 홍수 흐름에 대한 2D SWE에 대한 여러 출판물이 있었지만 이 주제에 대한 3D 수치 모델에 대한 연구는 거의 없다고 지적했습니다. 최근 전산 유체 역학 (CFD) 3D 시뮬레이션은 유체 흐름과 관련된 문제를 해결하기위한 광범위한 도구가되었습니다.

댐 파괴 파의 특성은 [13,14,15,16]에 의해 주목되었고 Issakhov [17]는 다양한 종류의 장애물이 압력 분포에 미치는 영향을 조사하기 위해 CFD 방법을 사용했습니다. 그들은 분포가 댐 표면에서 3 배 더 낮다는 것을 밝혔다.

Aureli [10]는 댐 파괴 파가 구조물에 미치는 영향의 정적 힘을 평가하기 위해 실험 테스트와 2D 및 3D 수치 모델을 사용했습니다. Mokarani [18]는 댐 브레이크 흐름 영향의 VOF 시뮬레이션에서 피크 압력 안정성 조건을 연구했습니다.

앞서 언급한 작품에서 구조물이나 구조물 군에 작용하는 힘은 압력에 의한 정 수력 또는 정력이었다. 한편, 급류에서 속도로 인한 힘은 압력 력보다 크거나 같았습니다 [19]. Armanini [20]는 정상 흐름에 대해이 항을 추정하기 위한 분석적 표현 만을 제시했습니다. 우리가 아는 한, 건물 그룹에 작용하는 비정상 흐름의 동적 힘을 생성하기 위해 2D 및 3D 수학적 모델을 모두 사용하는 작업은 없습니다.

따라서 본 연구에서는 제안된 2D 수치 모델과 3D 수학적 모델 모두에 의해 고립 된 장애물 또는 장애물 그룹에 대한 급격한 비정상 흐름의 테스트 사례를 재현했습니다. 수심 및 유속 수문 그래프와 같은 몇 가지 수력 학적 특성이 추정되었으며 측정 된 데이터와 매우 잘 일치했습니다.

특히 댐 브레이크 흐름이 서로 다른 건물에 가하는 동적인 힘도 시뮬레이션했습니다. 속도 유도 힘이 동적 힘에 미치는 영향 수준을 나타내는 매개 변수는 Froude 수와 입사 파동의 수심의 함수인 것으로 밝혀졌습니다. 또한 붕괴된 댐 사이트 폭 (b)과 초기 수위 (h0)는 충격력의 최대 값에 영향을 미치는 변수로 고려되었습니다.

Figure 1. (a) Configuration of experiment test (dimension in meters); (b) Gauges on the vertical front face of building.
Figure 1. (a) Configuration of experiment test (dimension in meters); (b) Gauges on the vertical front face of building.
Figure 2. (a) Distributed pressure profiles at centerline of front face of column; (b) Comparison of load-time histories simulated by different numerical models
Figure 2. (a) Distributed pressure profiles at centerline of front face of column; (b) Comparison of load-time histories simulated by different numerical models
Figure 3. Group of buildings in flooded area.
Figure 3. Group of buildings in flooded area.
Figure 4. Water depth and u-velocity profiles at gauge b.
Figure 4. Water depth and u-velocity profiles at gauge b.
Figure 5. Water hydrographs at gauges a and c.
Figure 5. Water hydrographs at gauges a and c.
Figure 6. Velocity component profiles at gauges a and c.
Figure 6. Velocity component profiles at gauges a and c.
Figure 7. Formation of incident and reflected waves.
Figure 7. Formation of incident and reflected waves.
Figure 8. Snapshots of streamlines of Froude number at different times: 1.0 s, 2.0 s, 5.0 s and 10 s.
Figure 8. Snapshots of streamlines of Froude number at different times: 1.0 s, 2.0 s, 5.0 s and 10 s.
Figure 9. Force in the flow direction exerted on 6 buildings.
Figure 9. Force in the flow direction exerted on 6 buildings.
Figure 10. The linear regression between forces per unit width (F) and q2b/h0.
Figure 10. The linear regression between forces per unit width (F) and q2b/h0.

Conclusions

댐 붕괴 흐름으로 인한 홍수 파도는 높은 속도 또는 큰 깊이가 관련되었을 때 건물에 큰 영향을 미칩니다. 본 논문에서는 2D 및 3D 수치 모델의 건물 및 건물 그룹에 대한 빠른 흐름에 의해 발생하는 유압 특성과 충격 부하를 추정할 수 있는 능력을 조사했습니다. 얕은 물 방정식에 기초한 2D 수학 모델은 FDS 방법으로 해결되었으며, FDS 방법은 최신 버전의 Flow 3D 유체 역학 모델과 함께 사용되었습니다. 연구의 주요 발견은 다음과 같습니다.
(1) 수심 또는 속도 프로파일을 공식화하기 위해 2D 및 3D 수치 솔루션은 모두 매우 유사합니다. 제안된 2D 수치 모델은 정적 힘의 최대 값 뿐만 아니라 수심 및 속도 구성 요소를 포함하는 유압 특성을 예측하는 데 적합합니다. 그러나 LES 및 RAN 난류 모듈이 포함된 3D 유체역학 모델은 2D 얕은 흐름 모델이 1개만 제공하는 동안 두 개의 최고 충격 부하를 잘 포착할 수 있습니다. 일반적으로 3D 결과는 실험 결과와 더 가깝습니다.
(2) 여러 건물에 대한 정적 및 동적 힘은 모두 LES 모듈을 사용하여 Flow 3D에 의해 계산되었습니다. 건물에서 속도에 의한 힘과 압력의 역할은 위치에 따라 다릅니다. 댐 현장 근처에서, 속도 유도 힘은 댐 파괴 파동의 주 방향에서 멀리 떨어져 있거나 두 번째 배열에서 압력 힘이 더 중요합니다. 속도 유도 힘의 영향은 매개 변수 α에 의해 정량화되며, 이는 사고파의 Froude 숫자와 수심 함수로 수행됩니다. q2b/h0과 정적 힘과 동적 힘의 피크 강도 사이의 선형 회귀 관계는 합리적인 R-제곱 양으로 해결됩니다.

추가 연구에서, 제시된 2D 수치 모델의 견고성과 효과는 더 명확하게 드러날 것입니다. 대규모 도메인에 대한 홍수 흐름을 시뮬레이션하는 데 쉽게 적용할 수 있습니다. 또한, α 매개변수의 제안된 방정식(21)은 실제 사례 연구에서 다운스트림 영역의 건물에 대한 속도 유도 힘의 영향을 정확하게 평가하기 위한 매우 의미가 있습니다. 이 매개 변수의 정확도 수준을 높이려면 서로 다른 조건에서 장애물에 작용하는 여러 가지 힘 실험이 구현되어야 합니다.

References

  1. Testa, G.; Zuccala, D.; Alcrudo, F.; Mulet, J.; Frazao, S.S. Flash flood flow experiment in a simplifed urban district. J. Hydraul. Res. 200745, 37–44. [Google Scholar] [CrossRef]
  2. Soares-Frazao, S.; Zech, Y. Dam-break flow through an idealized city. J. Hydraul. Res. 200846, 648–665. [Google Scholar] [CrossRef]
  3. Soares-Frazão, S.; Zech, Y. Experimental study of dam-break flow against an isolated obstacle. J. Hydraul. Res. 200745, 27–36. [Google Scholar] [CrossRef]
  4. Soares-Frazão, S. Experiments of dam-break wave over a triangular bottom sill. J. Hydraul. Res. 200745, 19–26. [Google Scholar] [CrossRef]
  5. di Cristo, C.; Evangelista, S.; Greco, M.; Iervolino, M.; Leopardi, A.; Vacca, A. Dam-break waves over an erodible embankment: Experiments and simulations. J. Hydraul. Res. 201856, 196–210. [Google Scholar] [CrossRef]
  6. Evangelista, S. Experiments and numerical simulations of dike erosion due to a wave impact. Water 20157, 5831–5848. [Google Scholar] [CrossRef]
  7. Li, Y.L.; Yu, C.H. Research on dam break flow induced front wave impacting a vertical wall based on the CLSVOF and level set methods. Ocean Eng. 2019178, 442–462. [Google Scholar] [CrossRef]
  8. Özgen, I.; Zhao, J.; Liang, D.; Hinkelmann, R. Urban flood modeling using shallow water equations with depth-dependent anisotropic porosity. J. Hydrol. 2016541, 1165–1184. [Google Scholar] [CrossRef]
  9. Mignot, E.; Li, X.; Dewals, B. Experimental modelling of urban flooding: A review. J. Hydrol. 2019568, 334–342. [Google Scholar] [CrossRef]
  10. Aureli, F.; Dazzi, A.; Maranzoni, A.; Mignosa, P.; Vacondio, R. Experimental and numerical evaluation of the force due to the impact of a dam break wave on a structure. Adv. Water Resour. 201576, 29–42. [Google Scholar] [CrossRef]
  11. Milanesi, L.; Pilotti, M.; Belleri, A.; Marini, A.; Fuchs, S. Vulnerability to flash floods: A simplified structural model for masonry buldings. Water Resour. Res. 201854, 7177–7197. [Google Scholar] [CrossRef]
  12. Shige-eda, M.; Akiyama, J. Numerical and experimental study on two dimensional flood flows with and without structures. J. Hydraul. Eng. 2003129, 817–821. [Google Scholar] [CrossRef]
  13. Cagatay, H.O.; Kocaman, S. Dam break flows during initial stage using SWE and RANs approaches. J. Hydraul. Res. 201048, 603–611. [Google Scholar] [CrossRef]
  14. Yang, S.; Yang, W.; Qin, S.; Li, Q.; Yang, B. Numerical study on characteristics of dam break wave. Ocean Eng. 2018159, 358–371. [Google Scholar] [CrossRef]
  15. Robb., D.M.; Vasquez., J.A. Numerical simulation of dam break flows using depth averaged hydrodynamic and three dimensional CFD models. In Proceedings of the 22nd Canadian Hydrotechnical Conference, Ottawa, ON, Canada, 28–30 April 2015. [Google Scholar]
  16. Kocaman, S.; Evangelista, S.; Viccione, G.; Guzel, H. Experimental and Numerical analysis of 3D dam break waves in an enclosed domain with a single oriented obstacles. Environ. Sci. Proc. 20202, 35. [Google Scholar] [CrossRef]
  17. Issakhov, A.; Zhandaulet, Y.; Nogaeva, A. Numerical simulation of dambreak flow for various forms of the obstacle by VOF method. Int. J. Multiph. Flow 2018109, 191–206. [Google Scholar] [CrossRef]
  18. Mokarani, C.; Abadie, S. Conditions for peak pressure stability in VOF simulations of dam break flow impact. J. Fluids Struct. 201662, 86–103. [Google Scholar] [CrossRef]
  19. Liu, L.; Sun, J.; Lin, B.; Lu, L. Building performance in dam break flow—an experimental sudy. Urban Water J. 201815, 251–258. [Google Scholar] [CrossRef]
  20. Armanini, A.; Larcher, M.; Odorizzi, M. Dynamic impact of a debris flow front against a vertical wall. In Proceedings of the 5th international conference on debris-flow hazards mitigation: Mechanics, prediction and assessment, Padua, Italy, 14–17 June 2011. [Google Scholar] [CrossRef]
  21. Hubbard, M.E.; Garcia Navarro, P. Flux difference splitting and the balancing of source terms and flux gradients. J. Comput. Phys. 2000165, 89–125. [Google Scholar] [CrossRef]
  22. Roe, P.L. A basis for upwind differencing of the two-dimensional unsteady Euler equations. In Numerical Methods in Fluids Dynamics II; Oxford University Press: Oxford, UK, 1986. [Google Scholar]
  23. Bradford, S.F.; Sander, B. Finite volume model for shallow water flooding of arbitrary topography. J. Hydraul. Eng. (ASCE) 2002128, 289–298. [Google Scholar] [CrossRef]
  24. Brufau, P.; Garica-Navarro, P. Two dimensional dam break flow simulation. Int. J. Numer. Meth. Fluids 200033, 35–57. [Google Scholar] [CrossRef]
  25. Hien, L.T.T. 2D Numerical Modeling of Dam-Break Flows with Application to Case Studies in Vietnam. Ph.D. Thesis, Brescia University, Brescia, Italy, 2014. [Google Scholar]
  26. Hien, L.T.T.; Tomirotti, M. Numerical modeling of dam break flows over complex topography. Case studies in Vietnam. In Proceedings of the 19th IAHR-APD Congress 2014, Hanoi, Vietnam, 21–24 September 2014; ISBN 978-604-82-1383-1. [Google Scholar]
  27. Flow-3D, Version 12.0; User Mannual; Flow Science Inc.: Santa Fe, NM, USA, 2020.
  28. Guney, M.S.; Tayfur, G.; Bombar, G.; Elci, S. Distorted physical model to study sudden partial dam break flow in an urban area. J. Hydraul. Eng. 2014140, 05014006. [Google Scholar] [CrossRef]
  29. Shige-eda, M.; Akiyama, J. Discussion and Closure to “Numerical and experimental study on two dimensional flood flows with and without structures” by Mirei Shige-eda and Juichiro Akiyama. J. Hydraul. Eng. 2005131, 336–337. [Google Scholar] [CrossRef]
  30. Ritter, A. Die Fortpflanzung der Wasserwelle (Generation of the water wave). Z. Ver. Dtsch. Ing. 189236, 947–954. [Google Scholar]
수자원/수처리/환경분야

수자원 분야

Water & Environmental

FLOW-3D는 작은 하수 처리 시스템부터 대형 수력 발전 프로젝트까지 수처리 및 환경 산업에 직면한 광범위한 문제를 해결할 수 있는 뛰어난 CFD 소프트웨어 입니다. FLOW-3D는 시뮬레이션의 복잡성을 감소시키고 최적의 솔루션에 대해 노력을 집중할 수 있도록 해줍니다. 이를 통해 통해 파악된 가치 있는 통찰력은 귀하의 상당한 시간과 비용을 절약 할 수 있습니다.

실제 지형을 적용하여 3차원 shallow water hybrid model을 이용한 댐 붕괴 시뮬레이션

FLOW-3D는 자유표면 흐름이 있는 수치해석 알고리듬에 의해 유동의 표면이 시공간적으로 변하는 모사를 위한 이상적인 도구라고 할 수 있습니다. 자유 표면은 물과 공기 같은 높은 비율의 밀도 변화를 가지는 유체들 사이의 특정한 경계를 일컫습니다. 자유 표면 흐름을 모델링하는 것은 일반적인 유동방정식과 난류 모델이 결합된 고급 알고리즘을 필요로 합니다. 이 기능은 FLOW-3D로 하여금 침수 구조에 의해 형성된 방수, 수력 점프 및 수면 변화의 흐름의 궤적을 포착 할 수 있습니다.


Bibliography & Technical Data

그림 10. 수문이 고르지 않게 열리는 경우의 시뮬레이션 결과

홍수 시즌에 하수구를 운영할 때 흐름 회로를 제어하는 ​​기술, 푸토코무네 제방을 통해 제방에 적용

요약 대규모 홍수 구호 작업에 대한 일반적인 흐름 회로 현상의 영향은 많은 보고서에서 연구되었으며 비교적 자세하게 연구되었습니다. 그러나 유량 변동이 ...
Figure 3. Computed contour of velocity magnitude (m/s) for Run 1 to Run 15.

FLOW-3D 소프트웨어를 이용한 유입구 및 배플 위치가 침전조 제거 효율에 미치는 영향

Ali Poorkarimi1 Khaled Mafakheri2Shahrzad Maleki2 Journal of Hydraulic StructuresJ. Hydraul. Struct., 2023; 9(4): 76-87DOI: 10.22055/jhs.2024.44817.1265 Abstract 중력에 의한 침전은 부유 ...
Figure 1 | Schematic of the present research model with dimensions and macro-roughnesses installed.

On the hydraulic performance of the inclined drops: the effect of downstreammacro-roughness elements

경사 낙하의 수력학적 성능: 하류 거시 거칠기 요소의 영향 Farhoud Kalateh a,*, Ehsan Aminvash a and Rasoul Daneshfaraz ba Faculty ...
Effects of ramp slope and discharge on hydraulic performance of submerged hump weirs

Effects of ramp slope and discharge on hydraulic performance of submerged hump weirs

Arash Ahmadi a, Amir H. Azimi b Abstract 험프 웨어는 수위 제어 및 배출 측정을 위한 기존의 수력 구조물 중 하나입니다. 상류 및 하류 경사로의 ...
그림 0 - 임계값의 다양한 위치에서 슬라이딩 밸브를 통과하는 흐름의 개략도: a) 밸브 아래, b) 밸브의 하류 측에 접선, c) 밸브의 상류 측에 접선

수직 슬라이딩 밸브의 토출 계수에 대한 형상 및 임계 위치 변화의 영향 평가

Abstract 본 연구의 목적은 다양한 위치에서 임계값을 갖는 슬라이딩 밸브의 유량계수를 조사하는 것입니다. 이 목표를 달성하기 위해 슬라이딩 밸브 아래 ...
Fig. 9 From: An Investigation on Hydraulic Aspects of Rectangular Labyrinth Pool and Weir Fishway Using FLOW-3D

An Investigation on Hydraulic Aspects of Rectangular Labyrinth Pool and Weir Fishway Using FLOW-3D

Research Article-Civil Engineering Open access Published: 04 January 2024 Abstract 웨어의 두 가지 서로 다른 배열(즉, 직선형 웨어와 직사각형 미로 ...
Figura 1 – Mapa de localização da PCH Salto Paraopeba

하천 저수지 물리적 모델의 침적 과정에 대한 전산 유체 역학 모델링(CFD) 기준

Natália Melo da Silva1 1; Jorge Luis Zegarra Tarqui2,Edna Maria de Faria Viana 3 Abstract 저수지 침전은 수력 발전의 지속 ...
Open Channels Flow에서의 콘크리트 캔버스 거동 연구

Study of Concrete Canvas Behavior in Open Channels Flow

Document Type : Research Paper Authors Mohammad Fayyaz saeed ahmadi Mahdi Dehghannejad Sani 1 Imam Hosein Uni 2 Researcher of ...
Figure 4. Rectangular stepped spillway with (a) three baffle arrangement (b) five baffle arrangement

Prediction of Energy Dissipation over Stepped Spillwaywith Baffles Using Machine Learning Techniques

Saurabh Pujari*, Vijay Kaushik, S. Anbu KumarDepartment of Civil Engineering, Delhi Technological University, IndiaReceived February 23, 2023; Revised April 25, ...
Schematic view of the experimental set-up

Short-time numerical simulation of ultrasonically assisted electrochemical removal of strontium from water

September 2023 DOI:10.30955/gnc2023.00436 Conference: 18th International Conference on Environmental Science and Technology CEST2023, 30 August to 2 September 2023, Athens, ...


FLOW-3D Water & Environmental Brochure (FSI) Bibliography

Models

  • Air Entrainment
  • Hybrid Shallow Water/3D Flow
  • Sediment Scour
  • Turbulence
  • More Modeling Capabilities

Case Studies

  • Evaluating Hydraulic Energy Losses and Total Hydraulic Head with FLOW-3D‘s Flux Baffles
  • Modeling Commercial Aquaculture Systems
  • Modeling Local Bridge Scour during Flood Event

Conference Proceedings

FLOW-3D HYDRO – The Complete CFD Solution for the Water & Environmental Industry

물 및 환경 산업을 위한 완벽한 CFD 솔루션인 FLOW-3D HYDRO의 신제품 출시를 알립니다.

Santa Fe, NM, 2020년 10월 29일 – Flow Science는 토목 및 환경 엔지니어링 산업을 위한 완벽한 CFD 모델링 솔루션인 FLOW-3D HYDRO를 출시했습니다. FLOW-3D HYDRO는 사용하기 편리한 수처리 해석 사용자 인터페이스를 갖추고 있으며 효율적인 모델링 워크플로우를 위한 새로운 시뮬레이션 템플릿과 토목 또는 환경 엔지니어의 요구에 맞춘 확장된 교육 자료를 제공합니다. FLOW-3D HYDRO의 진보된 솔버 개발에는 mine tailings, multiphase flows, shallow water models이 포함됩니다. 고성능 컴퓨팅을 위해 병렬 처리되고 모든 모델링 숙련도를 위해 설계된 FLOW-3D HYDRO는 사용자의 손에 뛰어난 시뮬레이션 기능을 제공합니다.
새로운 기능에 대한 자세한 설명은
https://flow3d.co.kr/flow-3d-hydro/
에서 확인할 수 있습니다.

“FLOW-3D HYDRO는 고객의 말을 경청하고 고객의 니즈를 파악한 결과입니다. 수처리 및 환경 고객을 위한 고급 CFD 솔루션을 개발하고 토목 및 환경 엔지니어링 업계에 범용-CFD 플로우-3D를 광범위하게 채택한 것을 바탕으로 소프트웨어 접근성과 사용자 관련성을 높일 수 있는 물 중심 인터페이스를 개발하여, 모델 설정 시간뿐만 아니라 설정 오류도 크게 감소했습니다. 유용성 및 모델링 성공 측면에서 이 신제품이 물과 환경 실무자들에게는 큰 자산이 될 것으로 생각합니다.

일련의 안내된 실습 과정을 통해 새로운 Flow-3D HYDRO소프트웨어를 소개하는 일련의 온라인 워크샵이 예정되어 있습니다. 워크샵 등록에는 참가자들이 소프트웨어와 소프트웨어 기능을 살펴볼 수 있도록 30일 평가 라이센스가 포함되어 있습니다. 등록은 다음 위치에서 사용할 수 있습니다.
https://www.flow3d.com/flow3d-hydro-workshop/에서 확인할 수 있습니다.

사용자 성공을 위해 FLOW-3D HYDRO는 높은 수준의 지원, 비디오 튜토리얼 및 광범위한 예제 시뮬레이션에 대한 액세스 권한을 제공합니다. 또한 고객은 Flow Science의 CFD 서비스를 활용하여 맞춤형 교육 과정, HPC 리소스 및 유연한 클라우드 컴퓨팅 옵션을 포함한 제품 경험을 강화할 수 있습니다.

FLOW-3D HYDRODOR 릴리즈 웨비나는 12월 3일에 열릴 예정입니다. 온라인 등록은 https://zoom.us/webinar/register/WN_pAh7Gi_fQXWc2Y3BGOrg-A에서 가능합니다.

FLOW-3D HYDRO

FLOW-3D HYDRO 2023R2
FLOW-3D HYDRO 2023R2

FLOW-3D HYDRO 2023R2 의 새로운 기능

새로운 결과 파일 형식

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  또는 EXODUS II 파일 형식으로 쓸 수 있습니다 . 새로운 EXODUS II 파일 형식은 각 개체에 대해 유한 요소 메시를 활용하므로 사용자는 다른 호환 가능한 포스트 프로세서 및 FEA 코드를 사용하여 FLOW-3D HYDRO 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. 

혼입 공기 시뮬레이션
FLOW-3D POST 의 새로운 EXODUS II 파일 형식에서 볼륨 렌더링 기능을 사용하여 동반된 공기를 보여주는 예입니다 .

새로운 결과 파일 형식은 hydr3d 솔버의 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 POST 의 새로운 시각화 기능 에 대해 자세히 알아보세요 .

난류 모델 개선

FLOW-3D HYDRO 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 규모를 더 잘 설명하며, 이제 사용자는 고정(물리 기반) 혼합 길이를 설정하는 대신 더 넓은 범위의 흐름에 동적 모델을 적용할 수 있습니다.

접촉식 탱크 혼합 시뮬레이션
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 HYDRO 2023R2는 정수압 솔버의 성능을 크게 향상시켜 전처리 단계에서 최대 6배 빠르게 수렴할 수 있도록 해줍니다.

확장된 지형 표현 지원

GeoTIFF 지원

2023R2 릴리스에서 FLOW-3D HYDRO는 기본적으로 래스터 지형 및 수심 측량을 위한 GeoTIFF(.tif) 파일 형식을 지원합니다. 이제 사용자는 GeoTIFF 파일을 사용자 인터페이스로 직접 가져올 수 있습니다. 

GeoTIFF 래스터 파일의 예
FLOW-3D HYDRO 에서 렌더링된 GeoTIFF(.tif) 래스터 파일의 예

LandXML 지원

측량 데이터가 균일하지 않거나 래스터 표면의 해상도가 충분하지 않은 경우 TIN 표면은 LandXML(.xml) 파일 형식을 통해 향상된 지형 지도를 제공합니다. FLOW-3D HYDRO 2023R2는 기본적으로 LandXML 파일을 가져옵니다. 

래스터 파일과의 향상된 상호 작용

래스터 파일은 고해상도에서 넓은 지형 영역을 다루는 경우가 많으므로 사용자 인터페이스에서 3D 표현의 상호 작용 속도가 느려질 수 있습니다. 이제 사용자는 3D 표현의 품질을 제어하여 렌더링 시간을 크게 줄이고 상호 작용성을 크게 향상시킬 수 있습니다.

FLOW-3D HYDRO 2023R1 의 새로운 기능

FLOW-3D 소프트웨어 제품군의 모든 제품은 2023R1에서 IT 관련 개선 사항을 받았습니다. 

FLOW-3D HYDRO 2023R1은 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.

얕은 물 난류 모델

난류는 물과 환경 흐름장의 주요 측면이며, 특히 얕은 물 근사치로 모델링된 영역에서는 더욱 그렇습니다. 우리는 모델링 위험을 줄이고 더 나은 결과를 제공하기 위해 세 가지 새로운 난류 모델, 일정한 확산도, 혼합 길이 및 Smagorinsky 모델을 포함하도록 얕은 물 모델의 난류 처리를 개선했습니다.

방사제 주변의 흐름

FLOW-3D HYDRO 2022R2 의 새로운 기능

FLOW-3D HYDRO 2022R2 출시로 Flow Science는 FLOW-3D HYDRO 의 워크스테이션과 HPC 버전을 통합하여 단일 노드 CPU 구성에서 다중 노드에 이르기까지 모든 유형의 하드웨어 아키텍처를 활용할 수 있는 단일 솔버 엔진을 제공했습니다. 병렬 고성능 컴퓨팅 실행. 추가 개발에는 향상된 공기 동반 기능과 물 및 환경 응용 분야에 대한 경계 조건 정의 개선이 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.

성능 확장의 예
증가하는 CPU 코어 수를 사용한 성능 확장의 예
메쉬 분해 - 2소켓 워크스테이션
2소켓 워크스테이션에서 OpenMP/MPI 하이브리드 병렬화를 위한 메시 분해의 예

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 런타임은 1.9에서 335까지 더 빨라졌습니다!

점탄성 유체에 대한 로그 형태 텐서 방법

점탄성 유체에 대한 새로운 솔버 옵션을 사용자가 사용할 수 있으며 특히 높은 Weissenberg 수에 효과적입니다.

FLOW-3D HYDRO 경계 조건 개선

FLOW-3D HYDRO 2022R2 에서는 물 적용 경계 조건에 대한 두 가지 개선 사항을 사용할 수 있습니다 . 얕은 물의 유량 경계 조건이 개선되어 보다 현실적이고 공간적으로 변화하는 속도 프로파일을 생성하므로 사용자는 정확도를 잃지 않고 도메인 크기를 줄일 수 있습니다. 자연적인 입구 경계 조건의 경우 정격 곡선 완화 시간 옵션을 사용하여 과도 조건에 대한 응답을 향상시킬 수 있습니다.

스트림 방향의 다양한 속도 프로파일
입구 경계에서 흐름 방향으로 변하는 속도 프로파일의 예

향상된 공기 동반 기능

디퓨저 및 유사한 기포 흐름 응용 분야의 경우 이제 질량 공급원을 사용하여 물기둥에 공기를 유입할 수 있습니다. 또한, 동반 공기 및 용존 산소의 난류 확산에 대한 기본값이 업데이트되었습니다.

디퓨저 모델의 예
디퓨저 모델의 예: 이제 질량 소스를 사용하여 물기둥에 공기를 유입할 수 있습니다.

FLOW-3D HYDRO 아카이브 의 새로운 기능

LOW-3D HYDRO 2022R1 의 새로운 기능

FLOW-3D HYDRO 의 새로운 기능

제품 개요

최근 FLOW Science, Inc에서는 토목 및 환경 엔지니어링 산업을위한 완벽한 CFD 모델링 솔루션인 FLOW-3D HYDRO 제품을 출시했습니다. 기존 FLOW-3D 사용자이거나 유압 엔지니어링 관행에 CFD 모델링 기능을 사용하시는 것에 관심이 있는 경우, 언제든지 아래 연락처로 연락주세요.
연락처 : 02-2026-0442
이메일 : flow3d@stikorea.co.kr

FLOW-3D HYDRO 는 더 높은 수준의 정확도와 모델 해상도를 제공하기 위해 3D 비 유압 모델링 기능이 필요한 경우 고급 모델링 도구로 사용할 수 있습니다. 일반적인 모델링 응용 분야는 소형 댐 / 인프라, 운송 수력학, 복잡한 3D 하천 수력학, 열 부력 연기, 배수구 및 오염 물질 수송과 관련됩니다. 

FLOW-3D HYDRO의 핵심 기능은 전체 3D 모델과 동적으로 연결될 수있는 얕은 물 모델입니다. 

이 기능을 통해 사용자는 멀티 스케일 모델링 애플리케이션을위한 모델 도메인을 확장하여 필요한 모델 해상도로 계산 효율성을 극대화 할 수 있습니다. FLOW-3D HYDRO  또한 강 및 환경 응용 분야에 특화된 추가 기능과 고급 물리학을 포함합니다.

시뮬레이션 템플릿

FLOW-3D HYDRO 의 작업 공간 템플릿으로 시간을 절약하고 실수를 방지하며 일관된 모델을 실행하십시오 . 작업 공간 템플릿은 일반적인 응용 분야에 대한 유체 속성, 물리적 모델, 수치 설정 및 시뮬레이션 출력을 미리로드합니다.

작업 공간 템플릿은 7 가지 모델 클래스에 사용할 수 있습니다.

  • 자유 표면 – TruVOF (기본값)
  • 공기 유입
  • 열 기둥
  • 퇴적물 수송
  • 얕은 물
  • 자유 표면 – 2 유체 VOF
  • 자유 표면 없음

사전로드 된 예제 시뮬레이션

FLOW-3D HYDRO 의 40 개 이상의 사전로드 된 물 중심 예제 시뮬레이션 라이브러리는 애플리케이션 모델링을위한 훌륭한 시작점을 제공합니다. 사전로드 된 예제 시뮬레이션은 모델러에게 모델 설정 및 모범 사례의 로드맵뿐만 아니라 대부분의 애플리케이션에 대한 자세한 시작점을 제공합니다.이전다음

비디오 튜토리얼

비디오 자습서는 새로운 사용자가 다양한 응용 프로그램을 모델링하는 방법을 빠르게 배울 수있는 훌륭한 경로를 제공합니다. FLOW-3D HYDRO 비디오 튜토리얼 기능 :

  • 광범위한 응용 및 물리학을위한 AZ 단계별 기록
  • “사용 방법”정보
  • 모범 사례를위한 팁
  • CAD / GIS 데이터, 시뮬레이션 파일 및 후 처리 파일

고급 솔버 개발

Tailings Model

새로운 Tailings Model은 tailings dam failure로 인한 tailings runout을 시뮬레이션하기위한 고급 기능을 제공합니다. tailings정의에 대한 다층 접근 방식과 함께 미세하고 거친 입자 구성을 나타내는 이중 모드 점도 모델은 모든 방법으로 건설 된 tailings 댐의 모델링을 허용합니다. 

얕은 물, 3D 및 하이브리드 3D / 얕은 물 메싱을 포함한 유연한 메싱을 통해 얕은 지역에서 빠른 솔루션을 제공하면서 다층 tailings의 복잡성을 정확하게 모델링 할 수 있습니다. 점성 경계층의 정확한 표현을 위해 얕은 물 메시에 2 층 Herschel-Bulkley 점도 모델을 사용할 수 있습니다.

모델 하이라이트

  • 미세 입자 및 거친 입자 광미 조성물을위한 이중 모드 점도 모델
  • 침전, 패킹 및 입자 종의 난류 확산을 포함한 Tailings  수송
  • 얕은 물 메시를위한 2 층 Herschel-Bulkley 점도 모델
  • 3D, 얕은 물, 3D / 얕은 물 하이브리드 메시를 포함한 유연한 메시 접근 방식
  • Multi-layer, variable composition tailings for general definition of tailings dam construction

Shallow Water

FLOW-3D HYDRO 의 얕은 물 모델링 기능은 3D 메시를 얕은 물 메시와 결합하여 탁월한 모델링 다양성을 제공하는 고유 한 하이브리드 메시를 사용합니다. 압력 솔버의 수치 개선으로 더 안정적이고 빠른 시뮬레이션이 가능합니다. 하이브리드 메쉬의 하단 전단 응력 계산이 크게 향상되어 정확도가 더욱 향상되었습니다. 지형에 거칠기를 적용하는 새로운 방법에는 Strickler, Chezy, Nikuradse, Colebrook-White, Haaland 및 Ramette 방정식이 포함됩니다.

Two-Fluid VOF Model

sharp 인터페이스가 있거나 없는 압축 가능 또는 비압축성 2 유체 모델은 항상 1 유체 자유 표면 모델과 함께 FLOW-3D 에서 사용할 수 있습니다 . 사실, sharp 인터페이스 처리는 TruVOF 기술을 자유 표면 모델과 공유하며 상용 CFD 소프트웨어에서 고유합니다. 최근 개발에는 2- 필드 온도 및 인터페이스 슬립 모델이 포함되었습니다. 이 모델은 오일 / 물, 액체 / 증기, 물 / 공기 및 기타 2 상 시스템에 성공적으로 적용되었습니다.

FLOW-3D HYDRO 는 2- 유체 솔루션의 정확성과 안정성에서 두 가지 중요한 발전을보고 있습니다. 운동량과 질량 보존 방정식의 강화 된 결합은 특히 액체 / 기체 흐름에서 계면에서 운동량 보존을 향상시킵니다. 연속성 방정식에서 제한된 압축성 항의 확장 된 근사값은 더 빠르고 안정적인 2 유체 압력 솔버를 만듭니다.

예를 들어, 터널 및 드롭 샤프트 설계와 같은 유압 응용 분야에서 공기가 종종 중요한 역할을 하기 때문에 두 개발 모두 FLOW-3D HYDRO 릴리스에 적시에 적용됩니다. 일반적으로 낮은 마하 수로 인해 이러한 경우 물과 공기에 제한된 압축성이 사용됩니다.

고성능 컴퓨팅 및 클라우드

고성능 컴퓨팅 FLOW-3D HYDRO

일반 워크스테이션 또는 랩톱으로 많은 작업을 수행 할 수 있지만, 대형 시뮬레이션과 고화질 시뮬레이션은 더 많은 CPU 코어를 활용함으로써 엄청난 이점을 얻을 수 있습니다. FLOW-3D CLOUD 및 고성능 컴퓨팅은 더 빠르고 정확한 모델을 실행할 수있는 더 빠른 런타임과 더 많은 선택권을 제공합니다.

하천 및 환경 중심 애플리케이션

TRANSPORTATION HYDRAULICS
SMALL DAMS AND DIVERSIONS
RIVER HYDRAULICS
SEDIMENT TRANSPORT AND DEPOSITION
OUTFALLS EFFLUENTS
THERMAL PLUMES BUOYANT FLOWS

Case Studies

자유 표면 모델링 방법

본 자료는 국내 사용자들의 편의를 위해 원문 번역을 해서 제공하기 때문에 일부 오역이 있을 수 있어서 원문과 함께 수록합니다. 자료를 이용하실 때 참고하시기 바랍니다.

Free Surface Modeling Methods

An interface between a gas and liquid is often referred to as a free surface. The reason for the “free” designation arises from the large difference in the densities of the gas and liquid (e.g., the ratio of density for water to air is 1000). A low gas density means that its inertia can generally be ignored compared to that of the liquid. In this sense the liquid moves independently, or freely, with respect to the gas. The only influence of the gas is the pressure it exerts on the liquid surface. In other words, the gas-liquid surface is not constrained, but free.

자유 표면 모델링 방법

기체와 액체 사이의 계면은 종종 자유 표면이라고합니다.  ‘자유’라는 호칭이 된 것은 기체와 액체의 밀도가 크게 다르기 때문입니다 (예를 들어, 물 공기에 대한 밀도 비는 1000입니다).  기체의 밀도가 낮다는 것은 액체의 관성에 비해 기체의 관성은 일반적으로 무시할 수 있다는 것을 의미합니다.  이러한 의미에서, 액체는 기체에 대해 독립적으로, 즉 자유롭게 움직입니다.  기체의 유일한 효과는 액체의 표면에 대한 압력입니다.  즉, 기체와 액체의 표면은 제약되어있는 것이 아니라 자유롭다는 것입니다.

In heat-transfer texts the term ‘Stephen Problem’ is often used to describe free boundary problems. In this case, however, the boundaries are phase boundaries, e.g., the boundary between ice and water that changes in response to the heat supplied from convective fluid currents.

열전달에 관한 문서는 자유 경계 문제를 묘사할 때 “Stephen Problem’”라는 용어가 자주 사용됩니다.  그러나 여기에서 경계는 상(phase) 경계, 즉 대류적인 유체의 흐름에 의해 공급된 열에 반응하여 변화하는 얼음과 물 사이의 경계 등을 말합니다.

Whatever the name, it should be obvious that the presence of a free or moving boundary introduces serious complications for any type of analysis. For all but the simplest of problems, it is necessary to resort to numerical solutions. Even then, free surfaces require the introduction of special methods to define their location, their movement, and their influence on a flow.

이름이 무엇이든, 자유 또는 이동 경계가 존재한다는 것은 어떤 유형의 분석에도 복잡한 문제를 야기한다는 것은 분명합니다. 가장 간단한 문제를 제외한 모든 문제에 대해서는 수치 해석에 의존할 필요가 있습니다. 그 경우에도 자유 표면은 위치, 이동 및 흐름에 미치는 영향을 정의하기 위한 특별한 방법이 필요합니다.

In the following discussion we will briefly review the types of numerical approaches that have been used to model free surfaces, indicating the advantages and disadvantages of each method. Regardless of the method employed, there are three essential features needed to properly model free surfaces:

  1. A scheme is needed to describe the shape and location of a surface,
  2. An algorithm is required to evolve the shape and location with time, and
  3. Free-surface boundary conditions must be applied at the surface.

다음 설명에서는 자유 표면 모델링에 사용되어 온 다양한 유형의 수치적 접근에 대해 간략하게 검토하고 각 방법의 장단점을 설명합니다. 어떤 방법을 사용하는지에 관계없이 자유롭게 표면을 적절히 모델화하는 다음의 3 가지 기능이 필요합니다.

  1. 표면의 형상과 위치를 설명하는 방식
  2. 시간에 따라 모양과 위치를 업데이트 하는 알고리즘
  3. 표면에 적용할 자유 표면 경계 조건

Lagrangian Grid Methods

Conceptually, the simplest means of defining and tracking a free surface is to construct a Lagrangian grid that is imbedded in and moves with the fluid. Many finite-element methods use this approach. Because the grid and fluid move together, the grid automatically tracks free surfaces.

라그랑주 격자 법

개념적으로 자유 표면을 정의하고 추적하는 가장 간단한 방법은 유체와 함께 이동하는 라그랑주 격자를 구성하는 것입니다. 많은 유한 요소 방법이 이 접근 방식을 사용합니다. 격자와 유체가 함께 움직이기 때문에 격자는 자동으로 자유 표면을 추적합니다.

At a surface it is necessary to modify the approximating equations to include the proper boundary conditions and to account for the fact that fluid exists only on one side of the boundary. If this is not done, asymmetries develop that eventually destroy the accuracy of a simulation.

표면에서 적절한 경계 조건을 포함하고 유체가 경계의 한면에만 존재한다는 사실을 설명하기 위해 근사 방정식을 수정해야합니다. 이것이 수행되지 않으면 결국 시뮬레이션의 정확도를 훼손하는 비대칭이 발생합니다.

The principal limitation of Lagrangian methods is that they cannot track surfaces that break apart or intersect. Even large amplitude surface motions can be difficult to track without introducing regridding techniques such as the Arbitrary-Lagrangian-Eulerian (ALE) method. References 1970 and 1974 may be consulted for early examples of these approaches.

라그랑지안 방법의 주요 제한은 분리되거나 교차하는 표면을 추적 할 수 없다는 것입니다. ALE (Arbitrary-Lagrangian-Eulerian) 방법과 같은 격자 재생성 기법을 도입하지 않으면 진폭이 큰 표면 움직임도 추적하기 어려울 수 있습니다. 이러한 접근법의 초기 예를 보려면 참고 문헌 1970 및 1974를 참조하십시오.

The remaining free-surface methods discussed here use a fixed, Eulerian grid as the basis for computations so that more complicated surface motions may be treated.

여기에서 논의된 나머지 자유 표면 방법은 보다 복잡한 표면 움직임을 처리할 수 있도록 고정된 오일러 그리드를 계산의 기준으로 사용합니다.

Surface Height Method

Low amplitude sloshing, shallow water waves, and other free-surface motions in which the surface does not deviate too far from horizontal, can be described by the height, H, of the surface relative to some reference elevation. Time evolution of the height is governed by the kinematic equation, where (u,v,w) are fluid velocities in the (x,y,z) directions. This equation is a mathematical expression of the fact that the surface must move with the fluid:

표면 높이 법

낮은 진폭의 슬로 싱, 얕은 물결 및 표면이 수평에서 너무 멀리 벗어나지 않는 기타 자유 표면 운동은 일부 기준 고도에 대한 표면의 높이 H로 설명 할 수 있습니다. 높이의 시간 진화는 운동학 방정식에 의해 제어되며, 여기서 (u, v, w)는 (x, y, z) 방향의 유체 속도입니다. 이 방정식은 표면이 유체와 함께 움직여야한다는 사실을 수학적으로 표현한 것입니다.

Finite-difference approximations to this equation are easy to implement. Further, only the height values at a set of horizontal locations must be recorded so the memory requirements for a three-dimensional numerical solution are extremely small. Finally, the application of free-surface boundary conditions is also simplified by the condition on the surface that it remains nearly horizontal. Examples of this technique can be found in References 1971 and 1975.

이 방정식의 유한 차분 근사를 쉽게 실행할 수 있습니다.  또한 3 차원 수치 해법의 메모리 요구 사항이 극도로 작아지도록 같은 높이의 위치 값만을 기록해야합니다.  마지막으로 자유 표면 경계 조건의 적용도 거의 수평을 유지하는 표면의 조건에 의해 간소화됩니다.  이 방법의 예는 참고 문헌의 1971 및 1975을 참조하십시오.

Marker-and-Cell (MAC) Method

The earliest numerical method devised for time-dependent, free-surface, flow problems was the Marker-and-Cell (MAC) method (see Ref. 1965). This scheme is based on a fixed, Eulerian grid of control volumes. The location of fluid within the grid is determined by a set of marker particles that move with the fluid, but otherwise have no volume, mass or other properties.

MAC 방법

시간 의존성을 가지는 자유 표면 흐름의 문제에 대해 처음 고안된 수치 법이 MAC (Marker-and-Cell) 법입니다 (참고 문헌 1965 참조).  이 구조는 컨트롤 볼륨 고정 오일러 격자를 기반으로합니다.  격자 내의 유체의 위치는 유체와 함께 움직이고, 그 이외는 부피, 질량, 기타 특성을 갖지 않는 일련의 마커 입자에 의해 결정됩니다.

Grid cells containing markers are considered occupied by fluid, while those without markers are empty (or void). A free surface is defined to exist in any grid cell that contains particles and that also has at least one neighboring grid cell that is void. The location and orientation of the surface within the cell was not part of the original MAC method.

마커를 포함한 격자 셀은 유체로 채워져있는 것으로 간주되며 마커가 없는 격자 셀은 빈(무효)것입니다.  입자를 포함하고, 적어도 하나의 인접 격자 셀이 무효인 격자의 자유 표면은 존재하는 것으로 정의됩니다.  셀 표면의 위치와 방향은 원래의 MAC 법에 포함되지 않았습니다.

Evolution of surfaces was computed by moving the markers with locally interpolated fluid velocities. Some special treatments were required to define the fluid properties in newly filled grid cells and to cancel values in cells that are emptied.

표면의 발전(개선)은 국소적으로 보간된 유체 속도로 마커를 이동하여 계산되었습니다.  새롭게 충전된 격자 셀의 유체 특성을 정의하거나 비어있는 셀의 값을 취소하거나 하려면 특별한 처리가 필요했습니다.

The application of free-surface boundary conditions consisted of assigning the gas pressure to all surface cells. Also, velocity components were assigned to all locations on or immediately outside the surface in such a way as to approximate conditions of incompressibility and zero-surface shear stress.

자유 표면 경계 조건의 적용은 모든 표면 셀에 가스 압력을 할당하는 것으로 구성되었습니다. 또한 속도 성분은 비압축성 및 제로 표면 전단 응력의 조건을 근사화하는 방식으로 표면 위 또는 외부의 모든 위치에 할당되었습니다.

The extraordinary success of the MAC method in solving a wide range of complicated free-surface flow problems is well documented in numerous publications. One reason for this success is that the markers do not track surfaces directly, but instead track fluid volumes. Surfaces are simply the boundaries of the volumes, and in this sense surfaces may appear, merge or disappear as volumes break apart or coalesce.

폭넓게 복잡한 자유 표면 흐름 문제 해결에 MAC 법이 놀라운 성공을 거두고 있는 것은 수많은 문헌에서 충분히 입증되고 있습니다.  이 성공 이유 중 하나는 마커가 표면을 직접 추적하는 것이 아니라 유체의 체적을 추적하는 것입니다.  표면은 체적의 경계에 불과하며, 그러한 의미에서 표면은 분할 또는 합체된 부피로 출현(appear), 병합, 소멸 할 가능성이 있습니다.

A variety of improvements have contributed to an increase in the accuracy and applicability of the original MAC method. For example, applying gas pressures at interpolated surface locations within cells improves the accuracy in problems driven by hydrostatic forces, while the inclusion of surface tension forces extends the method to a wider class of problems (see Refs. 1969, 1975).

다양한 개선으로 인해 원래 MAC 방법의 정확성과 적용 가능성이 증가했습니다. 예를 들어, 셀 내 보간 된 표면 위치에 가스 압력을 적용하면 정 수력으로 인한 문제의 정확도가 향상되는 반면 표면 장력의 포함은 방법을 더 광범위한 문제로 확장합니다 (참조 문헌. 1969, 1975).

In spite of its successes, the MAC method has been used primarily for two-dimensional simulations because it requires considerable memory and CPU time to accommodate the necessary number of marker particles. Typically, an average of about 16 markers in each grid cell is needed to ensure an accurate tracking of surfaces undergoing large deformations.

수많은 성공에도 불구하고 MAC 방법은 필요한 수의 마커 입자를 수용하기 위해 상당한 메모리와 CPU 시간이 필요하기 때문에 주로 2 차원 시뮬레이션에 사용되었습니다. 일반적으로 큰 변형을 겪는 표면의 정확한 추적을 보장하려면 각 그리드 셀에 평균 약 16 개의 마커가 필요합니다.

Another limitation of marker particles is that they don’t do a very good job of following flow processes in regions involving converging/diverging flows. Markers are usually interpreted as tracking the centroids of small fluid elements. However, when those fluid elements get pulled into long convoluted strands, the markers may no longer be good indicators of the fluid configuration. This can be seen, for example, at flow stagnation points where markers pile up in one direction, but are drawn apart in a perpendicular direction. If they are pulled apart enough (i.e., further than one grid cell width) unphysical voids may develop in the flow.

마커 입자의 또 다른 한계는 수렴 / 발산 흐름이 포함된 영역에서 흐름 프로세스를 따라가는 작업을 잘 수행하지 못한다는 것입니다. 마커는 일반적으로 작은 유체 요소의 중심을 추적하는 것으로 해석됩니다. 그러나 이러한 유체 요소가 길고 복잡한 가닥으로 당겨지면 마커가 더 이상 유체 구성의 좋은 지표가 될 수 없습니다. 예를 들어 마커가 한 방향으로 쌓여 있지만 수직 방향으로 떨어져 있는 흐름 정체 지점에서 볼 수 있습니다. 충분히 분리되면 (즉, 하나의 그리드 셀 너비 이상) 비 물리적 공극이 흐름에서 발생할 수 있습니다.

Surface Marker Method

One way to limit the memory and CPU time consumption of markers is to keep marker particles only on surfaces and not in the interior of fluid regions. Of course, this removes the volume tracking property of the MAC method and requires additional logic to determine when and how surfaces break apart or coalesce.

표면 마커 법

마커의 메모리 및 CPU 시간의 소비를 제한하는 방법 중 하나는 마커 입자를 유체 영역의 내부가 아니라 표면에만 보존하는 것입니다.  물론 이는 MAC 법의 체적 추적 특성이 배제되기 때문에 표면이 분할 또는 합체하는 방식과 시기를 특정하기위한 논리를 추가해야합니다.

In two dimensions the marker particles on a surface can be arranged in a linear order along the surface. This arrangement introduces several advantages, such as being able to maintain a uniform particle spacing and simplifying the computation of intersections between different surfaces. Surface markers also provide a convenient way to locate the surface within a grid cell for the application of boundary conditions.

2 차원의 경우 표면 마커 입자는 표면을 따라 선형으로 배치 할 수 있습니다.  이 배열은 입자의 간격을 균일하게 유지할 수있는 별도의 표면이 교차하는 부분의 계산이 쉽다는 등 몇 가지 장점이 있습니다.  또한 표면 마커를 사용하여 경계 조건을 적용하면 격자 셀의 표면을 간단한 방법으로 찾을 수 있습니다.

Unfortunately, in three-dimensions there is no simple way to order particles on surfaces, and this leads to a major failing of the surface marker technique. Regions may exist where surfaces are expanding and no markers fill the space. Without markers the configuration of the surface is unknown, consequently there is no way to add markers. Reference 1975 contains examples that show the advantages and limitations of this method.

불행히도 3 차원에서는 표면에 입자를 정렬하는 간단한 방법이 없으며 이로 인해 표면 마커 기술이 크게 실패합니다. 표면이 확장되고 마커가 공간을 채우지 않는 영역이 존재할 수 있습니다. 마커가 없으면 표면의 구성을 알 수 없으므로 마커를 추가 할 방법이 없습니다.
참고 문헌 1975이 방법의 장점과 한계를 보여주는 예제가 포함되어 있습니다.

Volume-of-Fluid (VOF) Method

The last method to be discussed is based on the concept of a fluid volume fraction. The idea for this approach originated as a way to have the powerful volume-tracking feature of the MAC method without its large memory and CPU costs.

VOF (Volume-of-Fluid) 법

마지막으로 설명하는 방법은 유체 부피 분율의 개념을 기반으로합니다. 이 접근 방식에 대한 아이디어는 대용량 메모리 및 CPU 비용없이 MAC 방식의 강력한 볼륨 추적 기능을 갖는 방법에서 시작되었습니다.

Within each grid cell (control volume) it is customary to retain only one value for each flow quantity (e.g., pressure, velocity, temperature, etc.) For this reason it makes little sense to retain more information for locating a free surface. Following this reasoning, the use of a single quantity, the fluid volume fraction in each grid cell, is consistent with the resolution of the other flow quantities.

각 격자 셀 (제어 체적) 내에서 각 유량 (예 : 압력, 속도, 온도 등)에 대해 하나의 값만 유지하는 것이 일반적입니다. 이러한 이유로 자유 표면을 찾기 위해 더 많은 정보를 유지하는 것은 거의 의미가 없습니다. 이러한 추론에 따라 각 격자 셀의 유체 부피 분율인 단일 수량의 사용은 다른 유량의 해상도와 일치합니다.

If we know the amount of fluid in each cell it is possible to locate surfaces, as well as determine surface slopes and surface curvatures. Surfaces are easy to locate because they lie in cells partially filled with fluid or between cells full of fluid and cells that have no fluid.

각 셀 내의 유체의 양을 알고 있는 경우, 표면의 위치 뿐만 아니라  표면 경사와 표면 곡률을 결정하는 것이 가능합니다.  표면은 유체 가 부분 충전 된 셀 또는 유체가 전체에 충전 된 셀과 유체가 전혀없는 셀 사이에 존재하기 때문에 쉽게 찾을 수 있습니다.

Slopes and curvatures are computed by using the fluid volume fractions in neighboring cells. It is essential to remember that the volume fraction should be a step function, i.e., having a value of either one or zero. Knowing this, the volume fractions in neighboring cells can then be used to locate the position of fluid (and its slope and curvature) within a particular cell.

경사와 곡률은 인접 셀의 유체 체적 점유율을 사용하여 계산됩니다.  체적 점유율은 계단 함수(step function)이어야 합니다, 즉, 값이 1 또는 0 인 것을 기억하는 것이 중요합니다.  이 것을 안다면, 인접 셀의 부피 점유율을 사용하여 특정 셀 내의 유체의 위치 (및 그 경사와 곡률)을 찾을 수 있습니다.

Free-surface boundary conditions must be applied as in the MAC method, i.e., assigning the proper gas pressure (plus equivalent surface tension pressure) as well as determining what velocity components outside the surface should be used to satisfy a zero shear-stress condition at the surface. In practice, it is sometimes simpler to assign velocity gradients instead of velocity components at surfaces.

자유 표면 경계 조건을 MAC 법과 동일하게 적용해야 합니다.  즉, 적절한 기체 압력 (및 대응하는 표면 장력)을 할당하고, 또한 표면에서 제로 전단 응력을 충족 시키려면 표면 외부의 어떤 속도 성분을 사용할 필요가 있는지를 확인합니다.  사실, 표면에서의 속도 성분 대신 속도 구배를 지정하는 것이보다 쉬울 수 있습니다.

Finally, to compute the time evolution of surfaces, a technique is needed to move volume fractions through a grid in such a way that the step-function nature of the distribution is retained. The basic kinematic equation for fluid fractions is similar to that for the height-function method, where F is the fraction of fluid function:

마지막으로, 표면의 시간 변화를 계산하려면 분포의 계단 함수의 성질이 유지되는 방법으로 격자를 통과하고 부피 점유율을 이동하는 방법이 필요합니다.  유체 점유율의 기본적인 운동학방정식은 높이 함수(height-function) 법과 유사합니다.  F는 유체 점유율 함수입니다.

A straightforward numerical approximation cannot be used to model this equation because numerical diffusion and dispersion errors destroy the sharp, step-function nature of the F distribution.

이 방정식을 모델링 할 때 간단한 수치 근사는 사용할 수 없습니다.  수치의 확산과 분산 오류는 F 분포의 명확한 계단 함수(step-function)의 성질이 손상되기 때문입니다.

It is easy to accurately model the solution to this equation in one dimension such that the F distribution retains its zero or one values. Imagine fluid is filling a column of cells from bottom to top. At some instant the fluid interface is in the middle region of a cell whose neighbor below is filled and whose neighbor above is empty. The fluid orientation in the neighboring cells means the interface must be located above the bottom of the cell by an amount equal to the fluid fraction in the cell. Then the computation of how much fluid to move into the empty cell above can be modified to first allow the empty region of the surface-containing cell to fill before transmitting fluid on to the next cell.

F 분포가 0 또는 1의 값을 유지하는 같은 1 차원에서이 방정식의 해를 정확하게 모델링하는 것은 간단합니다.  1 열의 셀에 위에서 아래까지 유체가 충전되는 경우를 상상해보십시오.  어느 순간에 액체 계면은 셀의 중간 영역에 있고, 그 아래쪽의 인접 셀은 충전되어 있고, 상단 인접 셀은 비어 있습니다.  인접 셀 내의 유체의 방향은 계면과 셀의 하단과의 거리가 셀 내의 유체 점유율과 같아야 한다는 것을 의미합니다.  그 다음 먼저 표면을 포함하는 셀의 빈 공간을 충전 한 후 다음 셀로 유체를 보내도록 위쪽의 빈 셀에 이동하는 유체의 양의 계산을 변경할 수 있습니다.

In two or three dimensions a similar procedure of using information from neighboring cells can be used, but it is not possible to be as accurate as in the one-dimensional case. The problem with more than one dimension is that an exact determination of the shape and location of the surface cannot be made. Nevertheless, this technique can be made to work well as evidenced by the large number of successful applications that have been completed using the VOF method. References 1975, 1980, and 1981 should be consulted for the original work on this technique.

2 차원과 3 차원에서 인접 셀의 정보를 사용하는 유사한 절차를 사용할 수 있지만, 1 차원의 경우만큼 정확하게 하는 것은 불가능합니다.  2 차원 이상의 경우의 문제는 표면의 모양과 위치를 정확히 알 수없는 것입니다.  그래도 VOF 법을 사용하여 달성 된 다수의 성공 사례에서 알 수 있듯이 이 방법을 잘 작동시킬 수 있습니다.  이 기법에 관한 초기의 연구 내용은 참고 문헌 1975,1980,1981를 참조하십시오.

The VOF method has lived up to its goal of providing a method that is as powerful as the MAC method without the overhead of that method. Its use of volume tracking as opposed to surface-tracking function means that it is robust enough to handle the breakup and coalescence of fluid masses. Further, because it uses a continuous function it does not suffer from the lack of divisibility that discrete particles exhibit.

VOF 법은 MAC 법만큼 강력한 기술을 오버 헤드없이 제공한다는 목표를 달성 해 왔습니다.  표면 추적이 아닌 부피 추적 기능을 사용하는 것은 유체 질량의 분할과 합체를 처리하는 데 충분한 내구성을 가지고 있다는 것을 의미합니다.  또한 연속 함수를 사용하기 때문에 이산된 입자에서 발생하는 숫자를 나눌 수 없는 문제를 겪지 않게 됩니다.

Variable-Density Approximation to the VOF Method

One feature of the VOF method that requires special treatment is the application of boundary conditions. As a surface moves through a grid, the cells containing fluid continually change, which means that the solution region is also changing. At the free boundaries of this changing region the proper free surface stress conditions must also be applied.

VOF 법의 가변 밀도 근사

VOF 법의 특수 처리가 필요한 기능 중 하나는 경계 조건의 적용입니다.  표면이 격자를 통과하여 이동할 때 유체를 포함하는 셀은 끊임없이 변화합니다.  즉, 계산 영역도 변화하고 있다는 것입니다.  이 변화하고있는 영역의 자유 경계에는 적절한 자유 표면 응력 조건도 적용해야합니다.

Updating the flow region and applying boundary conditions is not a trivial task. For this reason some approximations to the VOF method have been used in which flow is computed in both liquid and gas regions. Typically, this is done by treating the flow as a single fluid having a variable density. The F function is used to define the density. An argument is then made that because the flow equations are solved in both liquid and gas regions there is no need to set interfacial boundary conditions.

유체 영역의 업데이트 및 경계 조건의 적용은 중요한 작업입니다.  따라서 액체와 기체의 두 영역에서 흐름이 계산되는 VOF 법에 약간의 근사가 사용되어 왔습니다.  일반적으로 가변 밀도를 가진 단일 유체로 흐름을 처리함으로써 이루어집니다.  밀도를 정의하려면 F 함수를 사용합니다.  그리고, 흐름 방정식은 액체와 기체의 두 영역에서 계산되기 때문에 계면의 경계 조건을 설정할 필요가 없다는 논증이 이루어집니다.

Unfortunately, this approach does not work very well in practice for two reasons. First, the sensitivity of a gas region to pressure changes is generally much greater than that in liquid regions. This makes it difficult to achieve convergence in the coupled pressure-velocity solution. Sometimes very large CPU times are required with this technique.

공교롭게도 이 방법은 두 가지 이유로 인해 실제로는 그다지 잘 작동하지 않습니다.  하나는 압력의 변화에 대한 기체 영역의 감도가 일반적으로 액체 영역보다 훨씬 큰 것입니다.  따라서 압력 – 속도 결합 해법 수렴을 달성하는 것은 어렵습니다.  이 기술은 필요한 CPU 시간이 매우 커질 수 있습니다.

The second, and more significant, reason is associated with the possibility of a tangential velocity discontinuity at interfaces. Because of their different responses to pressure, gas and liquid velocities at an interface are usually quite different. In the Variable-Density model interfaces are moved with an average velocity, but this often leads to unrealistic movement of the interfaces.

두 번째 더 중요한 이유는 계면에서 접선 속도가 불연속이되는 가능성에 관련이 있습니다.  압력에 대한 반응이 다르기 때문에 계면에서 기체와 액체의 속도는 일반적으로 크게 다릅니다.  가변 밀도 모델은 계면은 평균 속도로 동작하지만, 이는 계면의 움직임이 비현실적으로 되는 경우가 많습니다.

Even though the Variable-Density method is sometimes referred to as a VOF method, because is uses a fraction-of-fluid function, this designation is incorrect. For accurately tracking sharp liquid-gas interfaces it is necessary to actually treat the interface as a discontinuity. This means it is necessary to have a technique to define an interface discontinuity, as well as a way to impose the proper boundary conditions at that interface. It is also necessary to use a special numerical method to track interface motions though a grid without destroying its character as a discontinuity.

가변 밀도 방법은 유체 분율 함수를 사용하기 때문에 VOF 방법이라고도하지만 이것은 올바르지 않습니다. 날카로운 액체-가스 인터페이스를 정확하게 추적하려면 인터페이스를 실제로 불연속으로 처리해야합니다. 즉, 인터페이스 불연속성을 정의하는 기술과 해당 인터페이스에서 적절한 경계 조건을 적용하는 방법이 필요합니다. 또한 불연속성으로 특성을 훼손하지 않고 격자를 통해 인터페이스 동작을 추적하기 위해 특수한 수치 방법을 사용해야합니다.

Summary

A brief discussion of the various techniques used to numerically model free surfaces has been given here with some comments about their relative advantages and disadvantages. Readers should not be surprised to learn that there have been numerous variations of these basic techniques proposed over the years. Probably the most successful of the methods is the VOF technique because of its simplicity and robustness. It is this method, with some refinement, that is used in the FLOW-3D program.

여기에서는 자유 표면을 수치적으로 모델링 할 때 사용하는 다양한 방법에 대해 상대적인 장점과 단점에 대한 설명을 포함하여 쉽게 설명하였습니다.  오랜 세월에 걸쳐 이러한 기본적인 방법이 많이 제안되어 온 것을 알고도 독자 여러분은 놀라지 않을 것입니다.  아마도 가장 성과를 거둔 방법은 간결하고 강력한 VOF 법 입니다.  이 방법에 일부 개량을 더한 것이 현재 FLOW-3D 프로그램에서 사용되고 있습니다.

Attempts to improve the VOF method have centered on better, more accurate, ways to move fluid fractions through a grid. Other developments have attempted to apply the method in connection with body-fitted grids and to employ more than one fluid fraction function in order to model more than one fluid component. A discussion of these developments is beyond the scope of this introduction.

VOF 법의 개선은 더 나은, 더 정확한 방법으로 유체 점유율을 격자를 통과하여 이동하는 것에 중점을 두어 왔습니다.  기타 개발은 물체 적합 격자(body-fitted grids) 관련 기법을 적용하거나 여러 유체 성분을 모델링하기 위해 여러 유체 점유율 함수를 채용하기도 했습니다.  이러한 개발에 대한 논의는 여기에서의 설명 범위를 벗어납니다.

References

1965 Harlow, F.H. and Welch, J.E., Numerical Calculation of Time-Dependent Viscous Incompressible Flow, Phys. Fluids 8, 2182.

1969 Daly, B.J., Numerical Study of the Effect of Surface Tension on Interface Instability, Phys. Fluids 12, 1340.

1970 Hirt, C.W., Cook, J.L. and Butler, T.D., A Lagrangian Method for Calculating the Dynamics of an Incompressible Fluid with Free Surface, J. Comp. Phys. 5, 103.

1971 Nichols, B.D. and Hirt, C.W.,Calculating Three-Dimensional Free Surface Flows in the Vicinity of Submerged and Exposed Structures, J. Comp. Phys. 12, 234.

1974 Hirt, C.W., Amsden, A.A., and Cook, J.L.,An Arbitrary Lagrangian-Eulerian Computing Method for all Flow Speeds, J. Comp. Phys., 14, 227.

1975 Nichols, B.D. and Hirt, C.W., Methods for Calculating Multidimensional, Transient Free Surface Flows Past Bodies, Proc. of the First International Conf. On Num. Ship Hydrodynamics, Gaithersburg, ML, Oct. 20-23.

1980 Nichols, B.D. and Hirt, C.W., Numerical Simulation of BWR Vent-Clearing Hydrodynamics, Nucl. Sci. Eng. 73, 196.

1981 Hirt, C.W. and Nichols, B.D., Volume of Fluid (VOF) Method for the Dynamics of Free Boundaries, J. Comp. Phys. 39, 201.

FLOW-3D What’s New Ver.12.0

FLOW-3D v12는 그래픽 사용자 인터페이스 (GUI)의 설계 및 기능에서 매우 큰 변화를 이룬 제품으로 모델 설정을 단순화하고 사용자 워크 플로를 향상시킵니다. 최첨단 Immersed Boundary Method(침수경계 방법)은 FLOW-3D v12 솔루션의 정확성을 높여줍니다. 다른 주요 기능으로는 슬러지 침강 모델, 2-Fluid 2-Temperature 모델 및 Steady State Accelerator가 있으며,이를 통해 사용자는 자유 표면 흐름을 더욱 빠르게 모델링 할 수 있습니다.

Physical and Numerical Model

Immersed boundary method

힘과 에너지 손실에 대한 정확한 예측은 고체 주위의 흐름과 관련된 많은 엔지니어링 문제를 모델링하는 데 중요합니다. 새 릴리스 FLOW-3D v12에는 이러한 문제점 해결을 위해 설계된 새로운 고스트 셀 기반 Immersed Boundary Method (IBM)가 있습니다. IBM은 내 외부 흐름 해석을 위해, 벽 근처에서 보다 정확한 해를 제공하여 드래그 앤 리프트 힘의 계산을 향상시킵니다.힘과 에너지 손실의 정확한 예측은 고체 주위의 흐름을 포함하는 많은 공학적 문제를 모델링 하는데 중요합니다.

Two-field temperature for the two-fluid model

2 유체 열전달 모델은 각 유체에 대한 에너지 전달 방정식을 분리하기 위해 확장되었습니다. 각 유체는 이제 자체 온도 변수를 가지므로 인터페이스 근처의 열 및 물질 전달 솔루션의 정확도가 향상됩니다. 인터페이스에서의 열전달은 이제 시간의 표 함수가 될 수 있는 사용자 정의 열전달 계수에 의해 제어됩니다.

블로그 보기

Sludge settling model

새로운 슬러지 정착 모델은 수처리 애플리케이션에 부가되어 사용자들이 수 처리 탱크와 클래리퍼의 고형 폐기물 역학을 모델링 할 수 있게 해 줍니다. 침전 속도가 분산상의 액적 크기의 함수 인 드리프트-플럭스 모델과 달리, 침전 속도는 슬러지 농도의 함수이며 기능 및 표 형식으로 입력 할 수 있습니다.

개발노트 읽기

Steady-state accelerator for free surface flows

이름에서 알 수 있듯이 정상 상태 가속기는 정상 상태 솔루션에 대한 접근을 빠르게합니다.
이것은 작은 진폭 중력과 모세관 표면파를 감쇠시킴으로써 달성되며 자유 표면 흐름에만 적용 할 수 있습니다.

개발노트 읽기

Void particles

Void particles 가 기포 및 상 변화 모델에 추가되었습니다. Void particles는 붕괴 된 Void 영역을 나타내며, 항력 및 압력을 통해 유체와 상호 작용하는 작은 기포로 작용합니다. 주변 유체 압력에 따라 크기가 변하고 시뮬레이션이 끝날 때의 최종 위치는 공기 유입 가능성을 나타냅니다.

Sediment scour model

퇴적물 수송 및 침식 모델은 정확성과 안정성을 향상시키기 위해 정비되었습니다. 특히 퇴적물 종의 질량 보존이 크게 개선되었습니다.

개발 노트 읽기>

Outflow pressure boundary condition

고정 압력 경계 조건에는 압력 및 유체 분율을 제외한 모든 유량이 해당 경계의 상류의 유량 조건을 반영하는 ‘유출’옵션이 포함됩니다. 유출 압력 경계 조건은 고정 압력 및 연속 경계 조건의 하이브리드입니다.

Moving particle sources

시뮬레이션 중에 입자 소스를 이동할 수 있습니다. 시간에 따른 병진 및 회전 속도는 표 형식으로 정의됩니다. 입자 소스의 운동은 소스에서 방출 된 입자의 초기 속도에 추가됩니다.

Variable center of gravity

기변 무게중심은 중력 및 비관 성 기준 프레임 모델에서, 시간의 함수로서 무게 중심의 위치는 외부 파일에서 테이블로서 정의 될 수있다. 이 기능은 연료를 소비하고 분리 단계를 수행하는 로켓과 같은 모형을 모델링 할 때 유용합니다.

공기 유입 모델

가장 간단한 부피 기반 공기 유입 모델 옵션이 기존 질량 기반 모델로 대체되었습니다. 질량 기반 모델은 부피와 달리 주변 유체 압력에 따라 부피가 변화하는 동안 흡입된 공기량이 보존되기 때문에 물리학적 모델입니다.

Tracer diffusion

유동 표면에서 생성된 추적 물질은 분자 및 난류 확산 과정에 의해 확산될 수 있으며, 예를 들어 실제 오염 물질의 동작을 모방한다.

Model Setup

Simulation units

온도를 포함하여 단위 시스템은 완전히 정의해야하는데 표준 단위 시스템이 제공됩니다. 또한 사용자는 다양한 옵션 중에서 질량, 시간 및 길이 단위를 정의 할 수 있으므로 사용자 정의가 가능한 편리한 단위를 사용할 수 있습니다. 사용자는 압력이 게이지 또는 절대 단위로 정의되는지 여부도 지정해야합니다. 기본 시뮬레이션 단위는 기본 설정에서 설정할 수 있습니다. 단위를 완전히 정의하면 FLOW-3D 가 물리량의 기본값을 정의하고 범용 상수를 설정하여 사용자가 요구하는 작업량을 최소화 할 수 있습니다.

Shallow water model

Manning’s roughness in shallow water model

Manning의 거칠기 계수는 지형 표면의 전단 응력 평가를 위해 얕은 물 모델에서 구현되었습니다. 표면 결함의 크기를 기반으로 기존 거칠기 모델을 보완하며 이 모델과 함께 사용할 수 있습니다. 표준 거칠기와 마찬가지로 매닝 계수는 구성 요소 또는 하위 구성 요소의 속성이거나 지형 래스터 데이터 세트에서 가져올 수 있습니다.

Mesh generation

하단 및 상단 경계 좌표의 정의만으로 수직 방향의 메시 설정이 단순화되었습니다.

Component transformations

사용자는 이제 여러 하위 구성 요소로 구성된 구성 요소에 회전, 변환 및 스케일링 변환을 적용하여 복잡한 형상 어셈블리 설정 프로세스를 단순화 할 수 있습니다. GMO (General Moving Object) 구성 요소의 경우, 이러한 변환을 구성 요소의 대칭 축과 정렬되도록 신체에 맞는 좌표계에 적용 할 수 있습니다.

Changing the number of threads at runtime

시뮬레이션 중에 솔버가 사용하는 스레드 수를 변경하는 기능이 런타임 옵션 대화 상자에 추가되어 사용 가능한 스레드를 추가하거나 다른 태스크에 자원이 필요한 경우 스레드 수를 줄일 수 있습니다.

Probe-controlled heat sources

활성 시뮬레이션 제어가 형상 구성 요소와 관련된 heat sources로 확장되었습니다. 히스토리 프로브로 열 방출을 제어 할 수 있습니다.

Time-dependent temperature at sources     

질량 및 질량 / 운동량 소스의 유체 온도는 이제 테이블 입력을 사용하여 시간의 함수로 정의 할 수 있습니다.

Emissivity coefficients

공극으로의 복사 열 전달을위한 방사율 계수는 이제 사용자가 방사율과 스테판-볼츠만 상수를 지정하도록 요구하지 않고 직접 정의됩니다. 후자는 이제 단위 시스템을 기반으로 솔버에 의해 자동으로 설정됩니다.

Output

  • 등속 필드 솔버 옵션을 사용할 때 유량 속도를 선택한 데이터 로 출력 할 수 있습니다 .
  • 벽 접착력으로 인한 지오메트리 구성 요소의 토크 는 기존 벽 접착력의 출력 외에도 일반 이력 데이터에 별도의 수량으로 출력됩니다.
  • 난류 모델 출력이 요청 될 때 난류 에너지 및 소산과 함께 전단 속도 및 y +가 선택된 데이터로 자동 출력됩니다 .
  • 공기 유입 모델 출력에 몇 가지 수량이 추가되었습니다. 자유 표면을 포함하는 모든 셀에서 혼입 된 공기 및 빠져 나가는 공기의 체적 플럭스가 재시작 및 선택된 데이터로 출력되어 사용자에게 공기가 혼입 및 탈선되는 위치 및 시간에 대한 자세한 정보를 제공합니다. 전체 계산 영역 및 각 샘플링 볼륨 에 대해이 두 수량의 시간 및 공간 통합 등가물 이 일반 히스토리 로 출력됩니다.
  • 솔버의 출력 파일 flsgrf 의 최종 크기 는 시뮬레이션이 끝날 때보 고됩니다.
  • 2 유체 시뮬레이션의 경우, 기존의 출력 수량 유체 체류 시간 및 유체 가 이동 한 거리는 이제 유체 # 1 및 # 2와 유체의 혼합물에 대해 별도로 계산됩니다.
  • 질량 입자의 경우 각 종의 총 부피와 질량이 계산되어 전체 계산 영역, 샘플링 볼륨 및 플럭스 표면에 대한 일반 히스토리 로 출력되어 입자 종 수에 대한 현재 출력을 보완합니다.
  • 예를 들어 사용자가 가스 미순환을 식별하고 연료 탱크의 환기 시스템을 설계하는 데 도움이 되도록 마지막 국부적 가스 압력이 옵션 출력량으로 추가되었습니다. 이 양은 유체가 채워지기 전에 셀의 마지막 간극 압력을 기록하며, 단열 버블 모델과 함께 사용됩니다.

New Customizable Source Routines

사용자 정의 가능한 새로운 소스 루틴이 추가되었으며 사용자의 개발 환경에서 액세스 할 수 있습니다.

소스 루틴 이름설명
cav_prod_cal캐비 테이션 생산 및 확산 속도
sldg_uset슬러지 정착 속도
phchg_mass_flux증발 및 응축에 의한 질량 흐름
flhtccl유체#1과#2사이의 열 전달 계수
dsize_cal2상 유동에서의 동적 낙하 크기 모델의 충돌 및 이탈율
elstc_custom.점탄성 유체에 대한 응력 방정식의 소스 용어

Brand New User Interface

FLOW-3D의 사용자 인터페이스가 완전히 재설계되어 사용자의 작업 흐름을 획기적으로 간소화하는 최신의 타일 구조를 제공합니다.

Dock widgets 설정

Physics, Fluids, Mesh 및 FAVOR ™를 포함한 모든 설정 작업이 형상 창 주위의 dock widgets으로 변환되어 모델 설정을 단일 탭으로 압축 할 수 있습니다. 이 전환을 통해 이전 버전의 복잡한 트리가 훨씬 깔끔하고 효율적인 메뉴 표시로 바뀌어 모델 설정 탭을 떠나지 않고도 모든 매개 변수에 쉽게 액세스 할 수 있습니다.

New Model Setup icons
With our new Model Setup design comes new icons, representing each step of the setup process.
New Physics icons
Our Physics icons are designed to be easily differentiated from one another at a glance, while providing clear visual representation of each model’s purpose and use.

RSS feed

새 RSS 피드부터 FLOW-3D v12.0 의 시뮬레이션 관리자 탭이 개선되었습니다 . FLOW-3D 를 시작하면 사용자에게 Flow Science의 최신 뉴스, 이벤트 및 블로그 게시물이 표시됩니다.

Configurable simulation monitor

시뮬레이션을 실행할 때 중요한 작업은 모니터링입니다. FLOW-3Dv12.0에서는 사용자가 시뮬레이션을 더 잘 모니터링할 수 있도록 Simulation Manager의 플로팅 기능이 향상되었습니다. 사용자는 시뮬레이션 런타임 그래프를 통해 모니터링할 사용 가능한 모든 일반 기록 데이터 변수를 선택하고 각 그래프에 여러 변수를 추가할 수 있습니다. 이제 런타임에서 사용할 수 있는 일반 기록 데이터는 다음과 같습니다.

  • 최소/최대 유체 온도
  • 프로브 위치의 온도
  • 유동 표면 위치에서의 유량
  • 시뮬레이션 진단(예:시간 단계, 안정성 한계)
Runtime plots of the flow rate at the gates of the large dam / Large dam with flux surfaces at the gates

Conforming mesh visualization

사용자는 이제 새로운 FAVOR ™ 독 위젯을 통해 적합한 메쉬 블록을 시각화 할 수 있습니다 .

Large raster and STL data

데이터를 처리하는 데 걸리는 시간으로 인해 큰 형상 데이터를 처리하는 것은 어려울 수 있습니다. 대형 지오메트리 데이터를 처리하는 데 여전히 상당한 시간이 소요될 수 있지만 FLOW-3D는 이제 이러한 대형 데이터 세트를 백그라운드 작업으로로드하여 사용자가 데이터를 처리하는 동안 완벽하게 응답하고 중단없는 인터페이스에서 계속 작업 할 수 있습니다.

FLOW-3D 교육 안내

education_banner

HIGH-END TOP CLASS
FLOW-3D CFD EDUCATION

FLOW-3D 분야별 교육 과정 안내


  • 교육 과정명 : 수리 분야

댐, 하천의 여수로, 수문 등 구조물 설계 및 방류, 월류 등 흐름 검토를 하기 위한 유동 해석 방법을 소개하는 교육 과정입니다. 유입 조건(수위, 유량 등)과 유출 조건에 따른 방류량 및 유속, 압력 분포 등 유체의 흐름을 검토를 할 수 있도록 관련 예제를 통해 적절한 기능을 습득하실 수 있습니다.

  • 교육 과정명 : 수처리 분야

정수처리 및 하수처리 공정에서 각 시설물들의 특성에 맞는 최적 운영조건 검토 및 설계 검토을 위한 유동해석 방법을 소개하는 교육 과정입니다. 취수부터 시작하여 혼화지, 분배수로, 응집지, 침전지, 여과지, 정수지, 협기조, 호기조, 소독조 등 각 공정별 유동 특성을 검토하기 위한 해석 모델을 설정하는 방법에 대해 알려드립니다.

  • 교육 과정명 : 주조 분야

주조 분야 사용자들이 쉽게 접근할 수 있도록 각 공정별로 해석 절차 및 해석 방법을 소개하는 교육 과정입니다. 고압다이캐스팅, 저압다이캐스팅, 경동주조, 중력주조, 원심주조, 정밀주조 등 주조 공법 별 관련 예제를 통해 적절한 기능들을 습득할 수 있도록 도와 드립니다.

  • 교육 과정명 : Micro/Bio/Nano Fluidics 분야

점성력 및 모세관력 같은 유체 표면에 작용하는 힘이 지배적인 미세 유동의 특성을 정확하게 표현할 수 있는 해석 방법에 대해 소개하는 교육 과정입니다. 열적, 전기적 물리 현상을 구현할 수 있도록 관련 예제와 함께 해석 방법을 알려드립니다.

  • 교육 과정명 : 코팅 분야 과정

코팅 공정에 따른 코팅액의 두께, 균일도, 유동 특성 분석을 위한 해석 방법을 소개하는 교육 과정입니다. Slide coating, Dip coating, Spin coating, Curtain coating, Slot coating, Roll coating, Gravure coating 등 각 공정별 예제와 함께 적절한 기능을 습득하실 수 있도록 도와 드립니다.

  • 교육 과정명 : 레이저 용접 분야

레이저 용접 해석을 하기 위한 물리 모델과 용접 조건들을 설정하는 방법에 대해 소개하는 교육 과정입니다. 해석을 통해 용접 공정을 최적화할 수 있도록 관련 예제와 함께 적절한 기능들을 습득할 수 있도록 도와 드립니다.

  • 교육 과정명 : 3D프린팅 분야 과정

Powder Bed Fusion(PBF)와 Directed Energy Deposition(DED) 공정에 대한 해석 방법을 소개하는 교육 과정입니다. 파우더 적층 및 레이저 빔을 조사하면서 동시에 금속 파우더 용융지가 적층되는 공정을 해석하는 방법을 관련 예제와 함께 습득하실 수 있습니다.

  • 교육 과정명 : 해양/항만 분야

해안, 항만, 해양 구조물에 대한 파랑의 영향 및 유체의 수위, 유속, 압력의 영향을 예측할 수 있는 해석 방법을 소개하는 과정입니다. 항주파, 슬로싱, 계류 등 해안, 해양, 에너지, 플랜트 분야 구조물 설계 및 검토에 필요한 유동해석을 하실 수 있는 방법을 알려드립니다. 각 현상에 대한 적절한 예제를 통해 기능을 습득하실 수 있습니다.

  • 교육 과정명 : 우주/항공 분야

항공기 및 우주선의 연료 탱크와 추진체 관리장치의 내부 유동, 엔진 및 터빈 노즐 내부의 유동해석을 하실 수 있도록 관련 메뉴에 대한 설명, 설정 방법을 소개하는 과정입니다. 경계조건 설정, Mesh 방법 등 유동해석을 위한 기본적인 내용과 함께 관련 예제를 통해 기능들을 습득하실 수 있습니다.

고객 맞춤형 과정


상기 과정 이외의 경우 고객의 사업 업무 환경에 적합한 사례를 중심으로 맞춤형 교육을 실시합니다. 필요하신 부분이 있으시면 언제든지 교육 담당자에게 연락하여 협의해 주시기 바랍니다.

고객센터 및 교육 담당자

  • 전화 : 02)2026-0450, 02)2026-0455
  • 이메일 : flow3d@stikorea.co.kr

교육 일정 안내


Education Banner

교육은 매월 정해진 일정에 시행되는 정기 교육과 고객의 요청에 의해 시행되는 비정기 교육이 있습니다. 비정기 교육은 별도문의 바랍니다.

1. 연간교육 일정


2. 교육 내용 : FLOW-3D Basic

  1. FLOW-3D 소개 및 이론
    • FLOW-3D 소개  – 연혁, 특징 등
    • FLOW-3D 기본 개념
      • VOF
      • FAVOR
    • 해석사례 리뷰
  2. GUI 소개 및 사용법
    • 해석 모델 작성법  – 물리 모델 설정
      • 모델 형상 정의
      • 격자 분할
      • 초기 유체 지정
      • 경계 조건 설정
    • 해석 결과 분석 방법  – 해석 모델 설명
  3. 해석 모델 작성 실습
    • 해석 모델 작성 실습  – 격자 분할
      • 물리 모델 설정
      • 모델 형상 및 초기 조건 정의
      • 경계 조건 설정
      • 해석 과정 모니터링
      • 해석 결과 분석
    • 질의 응답 및 토의

3. 교육 과정 : FLOW-3D Advanced

  1. Physics Ⅰ
    • Density evaluation
    • Drift flux
    • Scalars
    • Sediment scour
    • Shallow water
  2. Physics Ⅱ
    • Gravity and non-inertial reference frame
    • Heat transfer
    • Moving objects
    • Solidification
  3. FLOW-3D POST (Post-processor)
    • FLOW-3D POST 소개
    • Interface Basics
    • 예제 실습

FLOW-3D 교육 신청 방법 안내


  • 교육 신청은 홈페이지의 교육 신청 창에서 최소 3일 전에 신청합니다.
  • 모든 교육과정은 신청 인원이 2인 이상일때 개설되며, 선착순 마감입니다.
  • 교육 신청을 완료하시면, 신청시 입력하신 메일주소로 교육 담당자가 확인 메일을 보내드립니다.
  • 교육 시간은 Basic : 오전10시~오후5시, Advanced : 오후1시30분~오후5시30분까지입니다.
  • 교육비 안내
    • FLOW-3D, FLOW-3D CAST, FLOW-3D HYDRO Basic (2일) : 기업 66만원, 학생 55만원
    • FLOW-3D WELD/AM Basic 레이저용접, 3D 프린팅(2일) : 기업 88만원, 학생 66만원
    • FLOW-3D Advanced (1일) : 기업 33만원, 학생 25만원
    • 상기 가격은 부가세 포함 가격입니다.
  • 교육비는 현금(계좌이체)로 납부 가능하며, 교재 및 중식이 제공됩니다.
  • 세금계산서 발급을 위해 사업자등록증 또는 신분증 사본을 함께 첨부하여 신청해 주시기 바랍니다.
  • 교육 종료 후 이메일로 수료증이 발급됩니다.
고객센터 및 교육 담당자
  • 전화 : 02)2026-0450, 02)2026-0455
  • 이메일 : flow3d@stikorea.co.kr
교육 장소 안내
  • 지하철 1호선/가산디지털단지역 (8번출구), 지하철 7호선/가산디지털단지역 (5번출구)
  • 우림라이온스밸리 B동 302호 또는 교육장
  • 당사 건물에 주차할 경우 무료 주차 1시간만 지원되오니, 가능하면 대중교통을 이용해 주시기 바랍니다.
오시는 길

조선/해양 분야

Coastal & Maritime

FLOW-3D 는 선박설계, 슬로싱 동역학, 파도에 미치는 영향 및 환기를 포함하여 해안 및 해양 관련 분야에 사용할 수 있는 이상적인 소프트웨어입니다.

자유 표면 유체 역학, 파동 생성, 움직이는 물체, 계선 및 용접 공정과 관련한 FLOW-3D 의 기능은 해양 및 해양 산업에서 CFD 공정을 모델링하는 데 매우 적합한 도구입니다. 해안 응용 분야의 경우  FLOW-3D  해안 응용 분야의 경우 FLOW-3D  는 해안 구조물에 대한 심한 폭풍 및 쓰나미 파동의 세부 사항을 정확하게 예측하고 돌발 홍수 및 중요 구조물 홍수 및 피해 분석에 사용됩니다. 기능은 다음과 같습니다.

  • 자유 표면 – 파동 유체 역학 및 오버 토핑 : 규칙 및 불규칙파 및 파동 스펙트럼 (Pierson Moskowitz, JONSWAP)
  • Seakeeping – slamming, planing, porpoising 및 선체 선체 변위 : 완전히 결합된 선박 및 수중 차량 유체 역학
  • 선체 – Resistance, stability and dynamics: surging, heaving, pitching and rolling motion (response amplitude operators or RAOs)
  • 슬로싱 – LNG / 밸러스트 탱크
  • 해양 공학 – 파동 에너지 변환기
 

해안 응용 분야의 경우, FLOW-3D 는 강력한 폭풍과 쓰나미 현상에 의한 해안 구조물이 받는 영향에 대한 세부 사항 예측, 돌발 홍수에 의한 중요한 시설물에 대한 정확한 피해 분석 등을 위해 사용됩니다.

Mooring Lines, Springs and Ropes

FLOW-3D (계류선 및 스프링 등)의 특수 물체를 다른 움직이는 물체에 부착하면 엔지니어가 선박 런칭, 부유 장애물 역학, 부표, 파도에너지 변환기 등을 정확하게 포착할 수 있습니다.

Welding

FLOW-3D 용접 모듈이 추가되면서 조선업계의 용접분야에서는 다공성 등 용접 결함을 최소화할 수 있어 선체의 품질을 크게 높이는 동시에 생산 시간을 최적화할 수 있습니다.

Coastal & Maritime Case Studies

FLOW-3D 사용자들은 연약한 해안선 보호, 구조물에 대한 파장 시뮬레이션, 선체 설계 최적화, 선박 내 환기 연구 등 해안 및 해양 애플리케이션에 FLOW-3D를 사용합니다.

우리는 보트가 세계 항해를 하면서 마주칠 것 같은 다양한 조건에서 항해를 할 수 있는지를 볼 수 있었습니다. 그리고 속도뿐만 아니라 연료 효율과 안전도 고려하도록 설계를 수정할 수 있었습니다.
– Pete Bethune, skipper of Earthrace

Lateral wave impact in waterWave resultsEarthrace vessel
Validation of Sloshing Simulations in Narrow Tanks / Aerial Landslide Generated Wave Simulations / Earthrace: Speed, Fuel Efficiency and Safety
Wave impact vertical displacementEmerged breakwater accropodeStokes theory horizontal velocity
Wave Impact on Offshore StructuresInteraction Between Waves and BreakwatersWave Forces on Coastal Bridges

기타

Bibliography

Models

  • Hybrid Shallow Water/3D Flow
  • Moving Objects
  • Sediment Scour
  • More Modeling Capabilities

관련 기술자료

Evaluation of Pedestrian Safety for Wave Overtopping by Ship-Induced Waves in Waterfront Revetment

Evaluation of Pedestrian Safety for Wave Overtopping by Ship-Induced Waves in Waterfront Revetment

Young-Ki Moon, Chang-Ill Yoo, Jong-Min Lee, Sang-Hyub Lee, Han-Sam Yoon Author Affiliations +J. of Coastal Research, 116(sp1):314-318 (2024). https://doi.org/10.2112/JCR-SI116-064.1 Abstract Moon, ...
비선형 파력의 영향에 따른 잔해 언덕 방파제 형상의 효과에 대한 수치 분석

비선형 파력의 영향에 따른 잔해 언덕 방파제 형상의 효과에 대한 수치 분석

Numerical Analysis of the Effects of Rubble Mound Breakwater Geometry Under the Effect of Nonlinear Wave Force Research Article-Civil Engineering ...
Study on the critical sediment concentration determining the optimal transport capability of submarine sediment flows with different particle size composition

Study on the critical sediment concentration determining the optimal transport capability of submarine sediment flows with different particle size composition

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 ...
Fig. 7.Simulation results by single external force (left: rainfall, right: storm surge)

연안 지역의 복합 외력에 의한 침수 특성 분석

Analysis on inundation characteristics by compound external forces in coastal areas 연안 지역의 복합 외력에 의한 침수 특성 분석 Taeuk ...
Figure 1. Three-dimensional finite element model of local scouring of semi-exposed submarine cable.

반노출 해저케이블의 국부 정련과정 및 영향인자에 대한 수치적 연구

Numerical Study of the Local Scouring Process and Influencing Factors of Semi-Exposed Submarine Cables by Qishun Li,Yanpeng Hao *,Peng Zhang,Haotian Tan,Wanxing Tian,Linhao ...
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

Performance Comparison of NAMI DANCE and FLOW-3D® Models in Tsunami Propagation, Inundation and Currents using NTHMP Benchmark Problems

NTHMP 벤치마크 문제를 사용하여 쓰나미 전파, 침수 및 해류에서 NAMI DANCE 및 FLOW-3D® 모델의 성능 비교 Deniz Velioglu Sogut & ...
Figure 4 Snapshots of the trimaran model during the tests. a Inboard side hulls in the Tri-1confguration, b Outboard side hulls in the Tri-4 confguration, c Symmetric side hulls in the Tri-4confguration

조파식 3동선의 선체측면대칭이 저항성능에 미치는 영향에 관한 실험적 연구

Abolfath Askarian Khoob, Atabak Feizi, Alireza Mohamadi, Karim Akbari Vakilabadi, Abbas Fazeliniai & Shahryar Moghaddampour Abstract 이 논문은 비대칭 인보드, ...
Strain rate magnitude at the free surface, illustrating Kelvin-Helmoltz (KH) shear instabilities.

On the reef scale hydrodynamics at Sodwana Bay, South Africa

Atish Deoraj, Calvin Wells, Justin Pringle & Derek Stretch Environmental Fluid Mechanics (2022)Cite this article 277 Accesses Metricsdetails Abstract The ...
Figure 2: 3D (left) and 2D (right) views of wave elevation using case C

CFD 접근법을 사용하여 파도에서 하이드로포일의 SEAKEEPING 성능

SYAFIQ ZIKRYAND FITRIADHY*Faculty of Ocean Engineering Technology and Informatics, Universiti Malaysia Terengganu, 21030 KualaTerengganu, Terengganu, Malaysia*Corresponding author: naoe.afit@gmail.com http://doi.org/10.46754/umtjur.2021.07.017 Abstract ...
Numerical Simulation of Local Scour Around Square Artificial Reef

사각 인공어초 주변 국지세굴 수치모의

Abstract 인공어초(Artificial Reef, ARs)는 연안 어업 자원을 복원하고 생태 환경을 복원하기 위한 핵심 인공 구조물 중 하나입니다. 그러나 많은 AR이 세굴로 ...

FLOW-3D 제품소개

About FLOW-3D


FLOW-3D 2022R2
FLOW-3D 2022R2

FLOW-3D 개발 회사

Flow Science Inc Logo Green.svg
IndustryComputational Fluid Dynamics Software
Founded1980
FounderDr. C.W. “Tony” Hirt
Headquarters
Santa Fe, New Mexico, USA
United States
Key people
Dr. Amir Isfahani, President & CEO
ProductsFLOW-3D, FLOW-3D CAST, FLOW-3D AM, FLOW-3D CLOUD, FlowSight
ServicesCFD consultation and services

FLOW-3D 개요

FLOW-3D는 미국 뉴멕시코주(New Mexico) 로스알라모스(Los Alamos)에 있는 Flow Scicence, Inc에서 개발한 범용 전산유체역학(Computational Fluid Dynamics) 프로그램입니다. 로스알라모스 국립연구소의 수치유체역학 연구실에서 F.Harlow, B. Nichols 및 T.Hirt 등에 의해 개발된 MAC(Marker and Cell) 방법과 SOLA-VOF 방식을 기초로 하여, Hirt 박사가 1980년에 Flow Science, Inc사를 설립하여 계속 프로그램을 발전시켰으며 1985년부터 FLOW-3D를 전세계에 배포하였습니다.

유체의 3차원 거동 해석을 수행하는데 사용되는 CFD모형은 몇몇 있으나, 유동해석에 적용할 물리모델 선정은 해석의 정밀도와 밀접한 관계가 있으므로, 해석하고자 하는 대상의 유동 특성을 분석하여 신중하게 결정하여야 합니다.

FLOW-3D는 자유표면(Free Surface) 해석에 있어서 매우 정확한 해석 결과를 제공합니다. 해석방법은 자유표면을 포함한 비정상 유동 상태를 기본으로 하며, 연속방정식, 3차원 운동량 보전방정식(Navier-Stokes eq.) 및 에너지 보존방정식 등을 적용할 수 있습니다.

FLOW-3D는 유한차분법을 사용하고 있으며, 유한요소법(FEM, Finite Element Method), 경계요소법(Boundary Element Method)등을 포함하여 자유표면을 포함하는 유동장 해석(Fluid Flow Analysis)에서 공기와 액체의 경계면을 정밀하게 표현 가능합니다.

유체의 난류 해석에 대해서는 혼합길이 모형, 난류 에너지 모형, RNG(Renormalized Group Theory)  k-ε 모형, k-ω 모형, LES 모형 등 6개 모형을 적용할 수 있으며, 자유표면 해석을 위하여 VOF(Volume of Fluid) 방정식을 사용하고, 격자 생성시 사용자가 가장 쉽게 만들 수 있는 직각형상격자는 형상을 더욱 정확하게 표현하기 위해 FAVOR(Fractional Area Volume Obstacle Representation) 기법을 각 방정식에 적용하고 있습니다.

FLOW-3D는 비압축성(Incompressible Fluid Flow), 압축성 유체(Compressible Fluid Flow)의 유동현상 뿐만 아니라 고체와의 열전달 현상을 해석할 수 있으며, 비정상 상태의 해석을 기본으로 합니다.

FLOW-3D v12.0은 모델 설정을 간소화하고 사용자 워크 플로우를 개선하는 GUI(그래픽 사용자 인터페이스)의 설계 및 기능에 있어 중요한 변화를 가져왔습니다. 최첨단 Immersed Boundary Method는 FLOW-3Dv12.0솔루션의 정확도를 높여 줍니다. 다른 특징적인 주요 개발에는 슬러지 안착 모델, 2-유체 2-온도 모델, 사용자가 자유 표면 흐름을 훨씬 더 빠르게 모델링 할 수 있는 Steady State Accelerator등이 있습니다.

물리 및 수치 모델

Immersed Boundary Method

힘과 에너지 손실에 대한 정확한 예측은 솔리드 바디 주변의 흐름과 관련된 많은 엔지니어링 문제를 모델링하는 데 중요합니다. FLOW-3D v12.0의 릴리스에는 이러한 문제 해결을 위해 설계된 새로운 고스트 셀 기반 Immersed Boundary Method (IBM)가 포함되어 있습니다. IBM은 내부 및 외부 흐름을 위해 벽 근처 해석을 위해 보다 정확한 솔루션을 제공하여 드래그 앤 리프트 힘의 계산을 개선합니다.

Two-field temperature for the two-fluid model

2유체 열 전달 모델은 각 유체에 대한 에너지 전달 공식을 분리하도록 확장되었습니다. 이제 각 유체에는 고유한 온도 변수가 있어 인터페이스 근처의 열 및 물질 전달 솔루션의 정확도를 향상시킵니다. 인터페이스에서의 열 전달은 시간의 표 함수가 될 수 있는 사용자 정의 열 전달 계수에 의해 제어됩니다.

슬러지 침전 모델 / Sludge settling model

중요 추가 기능인 새로운 슬러지 침전 모델은 도시 수처리 시설물 응용 분야에 사용하면 수처리 탱크 및 정화기의 고형 폐기물 역학을 모델링 할 수 있습니다. 침전 속도가 확산된 위상의 방울 크기에 대한 함수인 드리프트-플럭스 모델과 달리, 침전 속도는 슬러지 농도의 함수이며 기능적인 형태와 표 형태로 모두 입력 할 수 있습니다.

Steady-state accelerator for free surface flows

이름이 암시하듯이, 정상 상태 가속기는 안정된 상태의 솔루션에 대한 접근을 가속화합니다. 이는 작은 진폭의 중력과 모세관 현상을 감쇠하여 이루어지며 자유 표면 흐름에만 적용됩니다.

꾸준한 상태 가속기

Void particles

보이드 입자가 버블 및 위상 변경 모델에 추가되었습니다. 보이드 입자는 항력과 압력 힘을 통해 유체와 상호 작용하는 작은 기포의 역할을 하는 붕괴된 보이드 영역을 나타냅니다. 주변 유체 압력에 따라 크기가 변경되고 시뮬레이션이 끝난 후 최종 위치는 공기 침투 가능성을 나타냅니다.

Sediment scour model

침전물의 정확성과 안정성을 향상시키기 위해 침전물의 운반과 침식 모델을 정밀 조사하였다. 특히, 침전물 종에 대한 질량 보존이 크게 개선되었습니다.

Outflow pressure boundary condition

고정 압력 경계 조건에는 이제 압력 및 유체 비율을 제외한 모든 유량이 해당 경계의 상류에 있는 흐름 조건을 반영하는 ‘유출’ 옵션이 포함됩니다. 유출 압력 경계 조건은 고정 압력 및 연속성 경계 조건의 혼합입니다.

Moving particle sources

시뮬레이션 중에 입자 소스는 이동할 수 있습니다. 시간에 따른 변환 및 회전 속도는 표 형식으로 정의됩니다. 입자 소스의 운동은 소스에서 방출 된 입자의 초기 속도에 추가됩니다.

Variable center of gravity

중력 및 비 관성 기준 프레임 모델에서 시간 함수로서의 무게 중심의 위치는 외부 파일의 표로 정의할 수 있습니다. 이 기능은 연료를 소모하는 로켓을 모델링하고 단계를 분리할 때 유용합니다.

공기 유입 모델

가장 간단한 부피 기반 공기 유입 모델 옵션이 기존 질량 기반 모델로 대체되었습니다.  질량 기반 모델은 부피와 달리 주변 유체 압력에 따라 부피가 변화하는 동안 흡입된 공기량이 보존되기 때문에 물리학적 모델입니다.

Air entrainment model in FLOW-3D v12.0

Tracer diffusion / 트레이서 확산

유동 표면에서 생성된 추적 물질은 분자 및 난류 확산 과정에 의해 확산될 수 있으며, 예를 들어 실제 오염 물질의 거동을 모방합니다.

모델 설정

시뮬레이션 단위

이제 온도를 포함하여 단위계 시스템을 완전히 정의해야 합니다. 표준 단위 시스템이 제공됩니다. 또한 사용자는 선택한 옵션에서 질량, 시간 및 길이 단위를 정의하여 편리하며, 사용자 정의된 단위를 사용할 수 있습니다. 사용자는 또한 압력이 게이지 단위로 정의되는지 절대 단위로 정의되는지 여부를 지정해야 합니다. 기본 시뮬레이션 단위는 Preferences(기본 설정)에서 설정할 수 있습니다. 단위를 완벽하게 정의하면 FLOW-3D는 물리적 수량에 대한 기본 값을 정의하고 범용 상수를 설정할 수 있으므로 사용자가 필요로 하는 작업량을 최소화할 수 있습니다.

Shallow water model

얕은 물 모델에서 매닝의 거칠기

Manning의 거칠기 계수는 지형 표면의 전단 응력 평가를 위해 얕은 물 모델에서 구현되었습니다. 표면 결함의 크기를 기반으로 기존 거칠기 모델을 보완하며이 모델과 함께 사용할 수 있습니다. 표준 거칠기와 마찬가지로 매닝 계수는 구성 요소 또는 하위 구성 요소의 속성이거나 지형 래스터 데이터 세트에서 가져올 수 있습니다.

메시 생성

하단 및 상단 경계 좌표의 정의만으로 수직 방향의 메시 설정이 단순화되었습니다.

구성 요소 변환

사용자는 이제 여러 하위 구성 요소로 구성된 구성 요소에 회전, 변환 및 스케일링 변환을 적용하여 복잡한 형상 어셈블리 설정 프로세스를 단순화 할 수 있습니다. GMO (General Moving Object) 구성 요소의 경우, 이러한 변환을 구성 요소의 대칭 축과 정렬되도록 신체에 맞는 좌표계에 적용 할 수 있습니다.

런타임시 스레드 수 변경

시뮬레이션 중에 솔버가 사용하는 스레드 수를 변경하는 기능이 런타임 옵션 대화 상자에 추가되어 사용 가능한 스레드를 추가하거나 다른 태스크에 자원이 필요한 경우 스레드 수를 줄일 수 있습니다.

프로브 제어 열원

활성 시뮬레이션 제어가 형상 구성 요소와 관련된 heat sources로 확장되었습니다.  history probes로 열 방출을 제어 할 수 있습니다.

소스에서 시간에 따른 온도

질량 및 질량/모멘트 소스의 유체 온도는 이제 테이블 입력을 사용하여 시간의 함수로 정의 할 수 있습니다.

방사율 계수

공극으로의 복사 열 전달을위한 방사율 계수는 이제 사용자가 방사율과 스테판-볼츠만 상수를 지정하도록 요구하지 않고 직접 정의됩니다. 후자는 이제 단위 시스템을 기반으로 솔버에 의해 자동으로 설정됩니다.

Output

  • 등속 필드 솔버 옵션을 사용할 때 유량 속도를 선택한 데이터로 출력 할 수 있습니다.
  • 벽 접착력으로 인한 지오메트리 구성 요소의 토크는 기존 벽 접착력 출력과 함께 별도의 수량으로 일반 이력 데이터에 출력됩니다.
  • 난류 모델 출력이 요청 될 때 난류 에너지 및 소산과 함께 전단 속도 및 y +가 선택된 데이터로 자동 출력됩니다.
  • 공기 유입 모델 출력에 몇 가지 수량이 추가되었습니다. 자유 표면을 포함하는 모든 셀에서 혼입 된 공기 및 빠져 나가는 공기의 체적 플럭스가 재시작 및 선택된 데이터로 출력되어 사용자에게 공기가 혼입 및 탈선되는 위치 및 시간에 대한 자세한 정보를 제공합니다. 전체 계산 영역 및 각 샘플링 볼륨 에 대해이 두 수량의 시간 및 공간 통합 등가물이 일반 히스토리 로 출력됩니다.
  • 솔버의 출력 파일 flsgrf 의 최종 크기는 시뮬레이션이 끝날 때 보고됩니다.
  • 2 유체 시뮬레이션의 경우, 기존의 출력 수량 유체 체류 시간 및 유체 가 이동 한 거리는 이제 유체 # 1 및 # 2와 유체의 혼합물에 대해 별도로 계산됩니다.
  • 질량 입자의 경우, 각 종의 총 부피 및 질량이 계산되어 전체 계산 영역, 샘플링 볼륨 및 플럭스 표면에 대한 일반 히스토리 로 출력되어 입자 종 수에 대한 현재 출력을 보완합니다.
  • 최종 로컬 가스 압력 은 사용자가 가스 포획을 식별하고 연료 탱크의 배기 시스템 설계를 지원하는 데 도움이되는 선택적 출력량으로 추가되었습니다. 이 양은 유체로 채워지기 전에 셀의 마지막 공극 압력을 기록하며 단열 버블 모델과 함께 사용됩니다.

새로운 맞춤형 소스 루틴

새로운 사용자 정의 가능 소스 루틴이 추가되었으며 사용자의 개발 환경에서 액세스 할 수 있습니다.

소스 루틴 이름기술
cav_prod_calCavitation 생성과 소산 비율
sldg_uset슬러지 침전 속도
phchg_mass_flux증발 및 응축으로 인한 질량 플럭스
flhtccl유체 # 1과 # 2 사이의 열전달 계수
dsize_cal2 상 흐름에서 동적 액적 크기 모델의 응집 및 분해 속도
elstc_custom점탄성 유체에 대한 응력 방정식의 Source Terms

새로운 사용자 인터페이스

FLOW-3D 사용자 인터페이스는 완전히 새롭게 디자인되어 현대적이고 평평한 구조로 사용자의 작업 흐름을 획기적으로 간소화합니다.

Setup dock widgets

Physics, Fluids, Mesh 및 FAVOR ™를 포함한 모든 설정 작업이 지오 메트리 윈도우 주변에서 독 위젯으로 변환되어 모델 설정을 단일 탭으로 요약할 수 있습니다. 이러한 전환으로 인해 이전 버전의 복잡한 접이식 트리가 훨씬 깨끗하고 효율적인 메뉴 프레젠테이션으로 대체되어 사용자는 ModelSetup탭을 떠나지 않고도 모든 매개 변수에 쉽게 액세스 할 수 있습니다.

New Model Setup icons

새로운 모델 설정 디자인에는 설정 프로세스의 각 단계를 나타내는 새로운 아이콘이 있습니다.

Model setup icons - FLOW-3D v12.0

New Physics icons

RSS feed

새 RSS 피드부터 FLOW-3D v12.0의 시뮬레이션 관리자 탭이 개선되었습니다. FLOW-3D 를 시작하면 사용자에게 Flow Science의 최신 뉴스, 이벤트 및 블로그 게시물이 표시됩니다.

RSS feed - FLOW-3D

Configurable simulation monitor

시뮬레이션을 실행할 때 중요한 작업은 모니터링입니다. FLOW-3Dv1.0에서는 사용자가 시뮬레이션을 더 잘 모니터링할 수 있도록 SimulationManager의 플로팅 기능이 향상되었습니다. 사용자는 시뮬레이션 런타임 그래프를 통해 모니터링할 사용 가능한 모든 일반 기록 데이터 변수를 선택하고 각 그래프에 여러 변수를 추가할 수 있습니다. 이제 런타임에서 사용할 수 있는 일반 기록 데이터는 다음과 같습니다.

  • 최소/최대 유체 온도
  • 프로브 위치의 온도
  • 유동 표면 위치에서의 유량
  • 시뮬레이션 진단(예:시간 단계, 안정성 한계)
출입문에 유동 표면이 있는 대형 댐
Runtime plots of the flow rate at the gates of the large dam

Conforming 메쉬 시각화

용자는 이제 새로운 FAVOR ™ 독 위젯을 통해 적합한 메쉬 블록을 시각화 할 수 있습니다.Visualize conforming mesh blocks

Large raster and STL data

데이터를 처리하는 데 걸리는 시간 때문에 큰 지오 메트리 데이터를 처리하는 것은 수고스러울 수 있습니다. 대형 지오 메트리 데이터를 처리하는 데는 여전히 상당한 시간이 걸릴 수 있지만, FLOW-3D는 이제 이러한 대규모 데이터 세트를 백그라운드 작업으로 로드하여 사용자