Fig. 11. Velocity vectors along x-direction through the center of the box culvert for B0, B30, B50, and B70 respectively.

Numerical investigation of scour characteristics downstream of blocked culverts

막힌 암거 하류의 세굴 특성 수치 조사

NesreenTahabMaged M.El-FekyaAtef A.El-SaiadaIsmailFathya
aDepartment of Water and Water Structures Engineering, Faculty of Engineering, Zagazig University, Zagazig 44519, Egypt
bLab Manager, Faculty of Engineering, Zagazig University, Zagazig 44519, Egypt


횡단 구조물을 통한 막힘은 안정성을 위협하는 위험한 문제 중 하나입니다. 암거의 막힘 형상 및 하류 세굴 특성에 미치는 영향에 관한 연구는 거의 없습니다.

이 연구의 목적은 수면과 세굴 모두에서 상자 암거를 통한 막힘의 작용을 수치적으로 논의하는 것입니다. 이를 위해 FLOW 3D v11.1.0을 사용하여 퇴적물 수송 모델을 조사했습니다.

상자 암거를 통한 다양한 차단 비율이 연구되었습니다. FLOW 3D 모델은 실험 데이터로 보정되었습니다. 결과는 FLOW 3D 프로그램이 세굴 다운스트림 상자 암거를 정확하게 시뮬레이션할 수 있음을 나타냅니다.

막힌 경우에 대한 속도 분포, 최대 세굴 깊이 및 수심을 플롯하고 비차단된 사례(기본 사례)와 비교했습니다.

그 결과 암거 높이의 70% 차단율은 상류의 수심을 암거 높이의 2.3배 증가시키고 평균 유속은 기본 경우보다 3배 더 증가시키는 것으로 입증되었다. 막힘 비율의 함수로 상대 최대 세굴 깊이를 추정하는 방정식이 만들어졌습니다.

Blockage through crossing structures is one of the dangerous problems that threaten its stability. There are few researches concerned with blockage shape in culverts and its effect on characteristics of scour downstream it.

The study’s purpose is to discuss the action of blockage through box culvert on both water surface and scour numerically. A sediment transport model has been investigated for this purpose using FLOW 3D v11.1.0. Different ratios of blockage through box culvert have been studied. The FLOW 3D model was calibrated with experimental data.

The results present that the FLOW 3D program was capable to simulate accurately the scour downstream box culvert. The velocity distribution, maximum scour depth and water depths for blocked cases have been plotted and compared with the non-blocked case (base case).

The results proved that the blockage ratio 70% of culvert height makes the water depth upstream increases by 2.3 times of culvert height and mean velocity increases by 3 times more than in the base case. An equation has been created to estimate the relative maximum scour depth as a function of blockage ratio.

1. Introduction

Local scour is the removal of granular bed material by the action of hydrodynamic forces. As the depth of scour hole increases, the stability of the foundation of the structure may be endangered, with a consequent risk of damage and failure [1]. So the prediction and control of scour is considered to be very important for protecting the water structures from failure. Most previous studies were designed to study the different factors that impact on scour and their relationship with scour hole dimensions like fluid characteristics, flow conditions, bed properties, and culvert geometry. Many previous researches studied the effect of flow rate on scour hole by information Froude number or modified Froude number [2][3][4][5][6]. Cesar Mendoza [6] found a good correlation between the scour depth and the discharge Intensity (Qg−.5D−2.5). Breusers and Raudkiv [7] used shear velocity in the outlet-scour prediction procedure. Ali and Lim [8] used the densimetric Froude number in estimation of the scour depth [1][8][9][10][11][12][13][14]. “The densimetric Froude number presents the ratio of the tractive force on sediment particle to the submerged specific weight of the sediment” [15](1)Fd=uρsρ-1gD50

Ali and Lim [8] pointed to the consequence of tailwater depth on scour behavior [1][2][8][13]. Abida and Townsend [2] indicated that the maximum depth of local scour downstream culvert was varying with the tailwater depth in three ways: first, for very shallow tailwater depths, local scouring decreases with a decrease in tailwater depth; second, when the ratio of tailwater depth to culvert height ranged between 0.2 and 0.7, the scour depth increases with decreasing tailwater depth; and third for a submerged outlet condition. The tailwater depth has only a marginal effect on the maximum depth of scour [2]. Ruff et al. [16] observed that for materials having similar mean grain sizes (d50) but different standard deviations (σ). As (σ) increased, the maximum scour hole depth decreased. Abt et al. [4] mentioned to role of soil type of maximum scour depth. It was noticed that local scour was more dangerous for uniform sands than for well-graded mixtures [1][2][4][9][17][18]. Abt et al [3][19] studied the culvert shape effect on scour hole. The results evidenced that the culvert shape has a limited effect on outlet scour. Under equivalent discharge conditions, it was noted that a square culvert with height equal to the diameter of a circular culvert would reduce scour [16][20]. The scour hole dimension was also effected by the culvert slope. Abt et al. [3][21] showed that the culvert slope is a key element in estimating the culvert flow velocity, the discharge capacity, and sediment transport capability. Abt et al. [21][22] tested experimentally culvert drop height effect on maximum scour depth. It was observed that as the drop height was increasing, the depth of scour was also increasing. From the previous studies, it could have noticed that the most scour prediction formula downstream unblocked culvert was the function of densimetric Froude number, soil properties (d50, σ), tailwater depth and culvert opening size. Blockage is the phenomenon of plugging water structures due to the movement of water flow loaded with sediment and debris. Water structures blockage has a bad effect on water flow where it causes increasing of upstream water level that may cause flooding around the structure and increase of scour rate downstream structures [23][24]. The blockage phenomenon through was studied experimentally and numerical [15][25][26][27][28][29][30][31][32][33]. Jaeger and Lucke [33] studied the debris transport behavior in a natural channel in Australia. Froude number scale model of an existing culvert was used. It was noticed that through rainfall event, the mobility of debris was impressed by stream shape (depth and width). The condition of the vegetation (size and quantities) through the catchment area was the main factor in debris transport. Rigby et al. [26] reported that steep slope was increasing the ability to mobilize debris that form field data of blocked culverts and bridges during a storm in Wollongong city.

Streftaris et al. [32] studied the probability of screen blockage by debris at trash screens through a numerical model to relate between the blockage probability and nature of the area around. Recently, many commercial computational fluid programs (CFD) such as SSIIM, Fluent, and FLOW 3D are used in the analysis of the scour process. Scour and sediment transport numerical model need to validate by using experimental data or field data [34][35][36][37][38]. Epely-Chauvin et al. [36] investigated numerically the effect of a series of parallel spur diked. The experimental data were compared by SSIIM and FLOW 3D program. It was found that the accuracy of calibrated FLOW 3D model was better than SSIIM model. Nielsen et al. [35] used the physical model and FLOW 3D model to analyze the scour process around the pile. The soil around the pile was uniform coarse stones in the physical models that were simulated by regular spheres, porous media, and a mixture of them. The calibrated porous media model can be used to determine the bed shear stress. In partially blocked culverts, there aren’t many studies that explain the blockage impact on scour dimensions. Sorourian et al. [14][15] studied the effect of inlet partial blockage on scour characteristics downstream box culvert. It resulted that the partial blockage at the culvert inlet could be the main factor in estimating the depth of scour. So, this study is aiming to investigate the effects of blockage through a box culvert on flow and scour characteristics by different blockage ratios and compares the results with a non-blocked case. Create a dimensionless equation relates the blockage ratio of the culvert with scour characteristics downstream culvert.

2. Experimental data

The experimental work of the study was conducted in the Hydraulics and Water Engineering Laboratory, Faculty of Engineering, Zagazig University, Egypt. The flume had a rectangular cross-section of 66 cm width, 65.5 cm depth, and 16.2 m long. A rectangular culvert was built with 0.2 m width, 0.2 m height and 3.00 m long with θ = 25° gradually outlet and 0.8 m fixed apron. The model was located on the mid-point of the channel. The sediment part was extended for a distance 2.20 m with 0.66 m width and 0.20 m depth of coarse sand with specific weight 1.60 kg/cm3, d50 = 2.75 mm and σ (d90/d50) = 1.50. The particle size distribution was as shown in Fig. 1. The experimental model was tested for different inlet flow (Q) of 25, 30, 34, 40 l/s for different submerged ratio (S) of 1.25, 1.50, 1.75.

3. Dimensional analysis

A dimensional analysis has been used to reduce the number of variables which affecting on the scour pattern downstream partial blocked culvert. The main factors affecting the maximum scour depth are:(2)ds=f(ρ.ρ

Fig. 2 shows a definition sketch of the experimental model. The maximum scour depth can be written in a dimensionless form as:(3)dsh=f(B.Fd.S)where the ds/h is the relative maximum scour depth.

4. Numerical work

The FLOW 3D is (CFD) program used by many researchers and appeared high accuracy in solving hydrodynamic and sediment transport models in the three dimensions. Numerical simulation with FLOW 3D was performed to study the impacts of blockage ratio through box culvert on shear stress, velocity distribution and the sediment transport in terms of the hydrodynamic features (water surface, velocity and shear stress) and morphological parameters (scour depth and sizes) conditions in accurately and efficiently. The renormalization group (RNG) turbulence model was selected due to its high ability to predict the velocity profiles and turbulent kinetic energy for the flow through culvert [39]. The one-fluid incompressible mode was used to simulate the water surface. Volume of fluid (VOF) method was employed in FLOW 3D to tracks a liquid interface through arbitrary deformations and apply the correct boundary conditions at the interface [40].1.

Governing equations

Three-dimensional Reynolds-averaged Navier Stokes (RANS) equation was applied for incompressible viscous fluid motion. The continuity equation is as following:(4)VF∂ρ∂t+∂∂xρuAx+∂∂yρvAy+∂∂zρwAz=RDIF(5)∂u∂t+1VFuAx∂u∂x+vAy∂u∂y+ωAz∂u∂z=-1ρ∂P∂x+Gx+fx(6)∂v∂t+1VFuAx∂v∂x+vAy∂v∂y+ωAz∂v∂z=-1ρ∂P∂y+Gy+fy(7)∂ω∂t+1VFuAx∂ω∂x+vAy∂ω∂y+ωAz∂ω∂z=-1ρ∂P∂z+Gz+fz

ρ is the fluid density,

VF is the volume fraction,

(x,y,z) is the Cartesian coordinates,

(u,v,w) are the velocity components,

(Ax,Ay,Az) are the area fractions and

RDIF is the turbulent diffusion.

P is the average hydrodynamic pressure,

(Gx, Gy, Gz) are the body accelerations and

(fx, fy, fz) are the viscous accelerations.

The motion of sediment transport (suspended, settling, entrainment, bed load) is estimated by predicting the erosion, advection and deposition process as presented in [41].

The critical shields parameter is (θcr) is defined as the critical shear stress τcr at which sediments begin to move on a flat and horizontal bed [41]:(8)θcr=τcrgd50(ρs-ρ)

The Soulsby–Whitehouse [42] is used to predict the critical shields parameter as:(9)θcr=0.31+1.2d∗+0.0551-e(-0.02d∗)(10)d∗=d50g(Gs-1ν3where:

d* is the dimensionless grain size

Gs is specific weight (Gs = ρs/ρ)

The entrainment coefficient (0.005) was used to scale the scour rates and fit the experimental data. The settling velocity controls the Soulsby deposition equation. The volumetric sediment transport rate per width of the bed is calculated using Van Rijn [43].2.

Meshing and geometry of model

After many trials, it was found that the uniform cell size with 0.03 m cell size is the closest to the experimental results and takes less time. As shown in Fig. 3. In x-direction, the total model length in this direction is 700 cm with mesh planes at −100, 0, 300, 380 and 600 cm respectively from the origin point, in y-direction, the total model length in this direction is 66 cm at distances 0, 23, 43 and 66 cm respectively from the origin point. In z-direction, the total model length in this direction is 120 cm. with mesh planes at −20, 0, 20 and 100 cm respectively.3.

Boundary condition

As shown in Fig. 4, the boundary conditions of the model have been defined to simulate the experimental flow conditions accurately. The upstream boundary was defined as the volume flow rate with a different flow rate. The downstream boundary was defined as specific pressure with different fluid elevation. Both of the right side, the left side, and the bottom boundary were defined as a wall. The top boundary defined as specified pressure with pressure value equals zero.

5. Validation of experimental results and numerical results

The experimental results investigated the flow and scour characteristics downstream culvert due to different flow conditions. The measured value of maximum scour depth is compared with the simulated depth from FLOW 3D model as shown in Fig. 5. The scour results show that the simulated results from the numerical model is quite close to the experimental results with an average error of 3.6%. The water depths in numerical model results is so close to the experimental results as shown in Fig. 6 where the experiment and numerical results are compared at different submerged ratios and flow rates. The results appear maximum error percentage in water depths upstream and downstream the culvert is about 2.37%. This indicated that the FLOW 3D is efficient for the prediction of maximum scour depth and the flow depths downstream box culvert.

6. Computation time

The run time was chosen according to reaching to the stability limit. Hydraulic stability was achieved after 50 s, where the scour development may still go on. For run 1, the numerical simulation was run for 1000 s as shown in Fig. 7 where it mostly reached to scour stability at 800 s. The simulation time was taken 500 s at about 95% of scour stability.

7. Analysis and discussions

Fig. 8 shows the study sections where sec 1 represents to upstream section, sec2 represents to inside section and sec3 represents to downstream stream section. Table 1 indicates the scour hole dimensions at different blockage case. The symbol (B) represents to blockage and the number points to blockage ratio. B0 case signifies to the non-blocked case, B30 is that blockage height is 30% to the culvert height and so on.

Table 1. The scour results of different blockage ratio.

Casehb cmB = hb/hQ lit/sSFdd50 mmds/h measuredls/hdd/hld/hds/h estimated

7.1. Scour hole geometry

The scour hole geometry mainly depends on the properties of soil of the bed downstream the fixed apron. From Table 1, the results show that the maximum scour depth in B0 case is about 0.58 of culvert height while the maximum deposition in B0 is 0.27 culvert height. There is a symmetric scour hole as shown in Fig. 9 in B0 case. An asymmetric scour hole is created in B50 and B70 due to turbulences that causes the deviation of the jet direction from the center of the flume where appear in Fig. 11 and Fig. 19.

7.2. Flow water surface

Fig. 10 presents the relative free surface water (hw/h) along the x-direction at center of the box culvert. From the mention Figure, it is easy to release the effect of different blockage ratios. The upstream water level rises by increasing the blockage ratio. Increasing upstream water level may cause flooding over the banks of the waterway. In the 70% blockage case, the upstream water level rises to 2.3 times of culvert height more than the non-blocked case at the same discharge and submerged ratio. The water surface profile shows an increase in water level upstream the culvert due to a decrease in transverse velocity. Because of decreasing velocity downstream culvert, there is an increase in water level before it reaches its uniform depth.

7.3. Velocity vectors

Scour downstream hydraulic structures mainly affects by velocities distribution and bed shear stress. Fig. 11 shows the velocity vectors and their magnitude in xz plane at the same flow conditions. The difference in the upstream water level due to the different blockage ratios is so clear. The maximum water level is in B70 and the minimum level is in B0. The inlet mean velocity value is about 0.88 m/s in B0 increases to 2.86 m/s in B70. As the blockage ratio increases, the inlet velocity increases. The outlet velocity in B0 case makes downward jet causes scour hole just after the fixed apron in the middle of the bed while the blockage causes upward water flow that appears clearly in B70. The upward jet decreases the scour depth to 0.13 culvert height less than B0 case. After the scour hole, the velocity decreases and the flow becomes uniform.

7.4. Velocity distribution

Fig. 12 represents flow velocity (Vx) distribution along the vertical depth (z/hu) upstream the inlet for the different blockage ratios at the same flow conditions. From the Figure, the maximum velocity creates closed to bed in B0 while in blocked case, the maximum horizontal velocity creates at 0.30 of relative vertical depth (z/hu). Fig. 13 shows the (Vz) distribution along the vertical depth (z/hu) upstream culvert at sec 1. From the mentioned Figure, it is easy to note that the maximum vertical is in B70 which appears that as the blockage ratio increases the vertical ratio also increases. In the non-blocked case. The vertical velocity (Vz) is maximum at (z/hu) equals 0.64. At the end of the fixed apron (sec 3), the horizontal velocity (Vx) is slowly increasing to reach the maximum value closed to bed in B0 and B30 while the maximum horizontal velocity occurs near to the top surface in B50 and B70 as shown in Fig. 14. The vertical velocity component along the vertical depth (z/hd) is presented in Fig. 15. The vertical velocity (Vz) is maximum in B0 at vertical depth (z/hd) 0.3 with value 0.45 m/s downward. Figs. 16 and 17 observe velocity components (Vx, Vz) along the vertical depth just after the end of blockage length at the centerline of the culvert barrel. It could be noticed the uniform velocity distribution in B0 case with horizontal velocity (Vx) closed to 1.0 m/s and vertical velocity closed to zero. In the blocked case, the maximum horizontal velocity occurs in depth more than the blockage height.

7.5. Bed velocity distribution

Fig. 18 presents the x-velocity vectors at 1.5 cm above the bed for different blockage ratios from the velocity vectors distribution and magnitude, it is easy to realize the position of the scour hole and deposition region. In B0 and B30, the flow is symmetric so that the scour hole is created around the centerline of flow while in B50 and B70 cases, the flow is asymmetric and the scour hole creates in the right of flow direction in B50. The maximum scour depth is found in the left of flow direction in B70 case where the high velocity region is found.

8. Maximum scour depth prediction

Regression analysis is used to estimate maximum scour depth downstream box culvert for different ratios of blockage by correlating the maximum relative scour by other variables that affect on it in one formula. An equation is developed to predict maximum scour depth for blocked and non-blocked. As shown in the equation below, the relative maximum scour depth(ds/hd) is a function of densimetric Froude number (Fd), blockage ratio (B) and submerged ratio (S)(11)dsh=0.56Fd-0.20B+0.45S-1.05

In this equation the coefficient of correlation (R2) is 0.82 with standard error equals 0·08. The developed equation is valid for Fd = [0.9 to 2.10] and submerged ratio (S) ≥ 1.00. Fig. 19 shows the comparison between relative maximum scour depths (ds/h) measured and estimated for different blockage ratios. Fig. 20 clears the comparison between residuals and ds/h estimated for the present study. From these figures, it could be noticed that there is a good agreement between the measured and estimated relative scour depth.

9. Comparison with previous scour equations

Many previous scour formulae have been produced for calculation the maximum scour depth downstream non-blockage culvert. These equations have been included the effect of flow regime, culvert shape, soil properties and the flow rate on maximum scour depth. Two of previous experimental studies data have been chosen to be compared with the present study results in non-blocked study data. Table 2 shows comparison of culvert shape, densmetric Froude number, median particle size and scour equations for these previous studies. By applying the present study data in these studies scour formula as shown in Fig. 21, it could be noticed that there are a good agreement between present formula results and others empirical equations results. Where that Lim [44] and Abt [4] are so closed to the present study data.

Table 2. Comparison of some previous scour formula.

ResearchersFdCulvert shaped50(mm)Proposed equationSubmerged ratio
Present study0.9–2.11square2.75dsh=0.56Fd-0.20B+0.45S-1.051.25–1.75
Lim [44]1–10Circular1.65dsh=0.45Fd0.47
Abt [4]Fd ≥ 1Circular0.22–7.34-dsh=3.67Fd0.57∗D500.4∗σ-0.4

10. Conclusions

The present study has shown that the FLOW 3D model can accurately simulate water surface and the scour hole characteristics downstream the box culvert with error percentage in water depths does not exceed 2.37%. Velocities distribution through and outlets culvert barrel helped on understanding the scour hole shape.

The blockage through culvert had caused of increasing of water surface upstream structure where the upstream water level in B70 was 2.3 of culvert height more than non-blocked case at the same discharge that could be dangerous on the stability of roads above. The depth averaged velocity through culvert barrel increased by 3 times its value in non-blocked case.

On the other hand, blockage through culvert had a limited effect on the maximum scour depth. The little effect of blockage on maximum scour depth could be noticed in Fig. 11. From this Figure, it could be noted that the residual part of culvert barrel after the blockage part had made turbulences. These turbulences caused the deviation of the flow resulting in the formation of asymmetric scour hole on the side of channel. This not only but in B70 the blockage height caused upward jet which made a wide far scour hole as cleared from the results in Table 1.

An empirical equation was developed from the results to estimate the maximum scour depth relative to culvert height function of blockage ratio (B), submerged ratio (S), and densimetric Froude number (Fd). The equation results was compared with some scour formulas at the same densimetric Froude number rang where the present study results was in between the other equations results as shown in Fig. 21.

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.


[1]P. Sarathi, M. Faruque, R. BalachandarInfluence of tailwater depth, sediment size and densimetric Froude number on scour by submerged square wall jetsJ. Hydraul. Res., 46 (2) (2008), pp. 158-175CrossRefView Record in ScopusGoogle Scholar[2]H. Abida, R. TownsendLocal scour downstream of box-culvert outletsJ. Irrig. Drain. Eng., 117 (3) (1991), pp. 425-440CrossRefView Record in ScopusGoogle Scholar[3]S.R. Abt, C.A. Donnell, J.F. Ruff, F.K. DoehringCulvert Slope and Shape Effects on Outlet ScourTransp. Res. Rec., 1017 (1985), pp. 24-30View Record in ScopusGoogle Scholar[4]S.R. Abt, R.L. Kloberdanz, C. MendozaUnified culvert scour determinationJ. Hydraul. Eng., 110 (10) (1984), pp. 1475-1479CrossRefView Record in ScopusGoogle Scholar[5]J.P. Bohan, Erosion And Riprap Requirements At Culvert And Storm-Drain Outlets, ARMY ENGINEER WATERWAYS EXPERIMENT STATION VICKSBURG MISS1970.Google Scholar[6]C. Mendoza, S.R. Abt, J.F. RuffHeadwall influence on scour at culvert outletsJ. Hydraul. Eng., 109 (7) (1983), pp. 1056-1060CrossRefView Record in ScopusGoogle Scholar[7]H. Breusers, A. Raudkivi, Scouring, hydraulic structures design manual, vol. 143, IAHR, AA Balkema, Rotterdam, 1991.Google Scholar[8]K. Ali, S. LimLocal scour caused by submerged wall jetsProc. Inst. Civ. Eng., 81 (4) (1986), pp. 607-645CrossRefView Record in ScopusGoogle Scholar[9]O. Aderibigbe, N. RajaratnamEffect of sediment gradation on erosion by plane turbulent wall jetsJ. Hydraul. Eng., 124 (10) (1998), pp. 1034-1042View Record in ScopusGoogle Scholar[10]F.W. Blaisdell, C.L. AndersonA comprehensive generalized study of scour at cantilevered pipe outletsJ. Hydraul. Res., 26 (4) (1988), pp. 357-376CrossRefView Record in ScopusGoogle Scholar[11]Y.-M. Chiew, S.-Y. LimLocal scour by a deeply submerged horizontal circular jetJ. Hydraul. Eng., 122 (9) (1996), pp. 529-532View Record in ScopusGoogle Scholar[12]R.A. Day, S.L. Liriano, W.R. WhiteEffect of tailwater depth and model scale on scour at culvert outletsProc. Instit. Civil Eng. – Water Marit. Eng., 148 (3) (2001), pp. 189-198 Record in ScopusGoogle Scholar[13]S. Emami, A.J. SchleissPrediction of localized scour hole on natural mobile bed at culvert outletsScour and Erosion (2010), pp. 844-853CrossRefView Record in ScopusGoogle Scholar[14]S. Sorourian, A. Keshavarzi, J. Ball, B. SamaliStudy of Blockage Effect on Scouring Pattern Downstream of a Box Culvert under Unsteady FlowAustr. J Water Resor. (2013)Google Scholar[15]S. Sorourian, Turbulent Flow Characteristics At The Outlet Of Partially Blocked Box Culverts, in: 36th IAHR World Congress, The Hague, the Netherlands, 2015.Google Scholar[16]J. Ruff, S. Abt, C. Mendoza, A. Shaikh, R. KloberdanzScour at culvert outlets in mixed bed materialsUnited States. Federal Highway Administration. Office of Research and Development (1982)Google Scholar[17]S.A. Ansari, U.C. Kothyari, K.G.R. RajuInfluence of cohesion on scour under submerged circular vertical jetsJ. Hydraul. Eng., 129 (12) (2003), pp. 1014-1019View Record in ScopusGoogle Scholar[18]B. Crookston B. Tullis, Scour and Riprap Protection in a Bottomless Arch Culvert, in: World Environmental and Water Resources Congress 2008: Ahupua’A, 2008, pp. 1–10.Google Scholar[19]S.R. Abt, J. Ruff, F. Doehring, C. DonnellInfluence of culvert shape on outlet scourJ. Hydraul. Eng., 113 (3) (1987), pp. 393-400View Record in ScopusGoogle Scholar[20]Y.H. Chen, Scour at outlets of box culverts, Colorado State University, 1970.Google Scholar[21]S. Abt, P. Thompson, T. LewisEnhancement of the culvert outlet scour estimation equationsTransp. Res. Rec. J. Transp. Res. Board, 1523 (1996), pp. 178-185View Record in ScopusGoogle Scholar[22]F.K. Doehring, S.R. AbtDrop height influence on outlet scourJ. Hydraul. Eng., 120 (12) (1994), pp. 1470-1476CrossRefView Record in ScopusGoogle Scholar[23]W. Weeks, A. Barthelmess, E. Rigby, G. Witheridge, R. Adamson, Australian rainfall and runoff revison project 11: blockage of hydraulic structures, 2009.Google Scholar[24]W. Weeks, G. Witheridge, E. Rigby, A. BarthelmessProject 11: blockage of hydraulic structuresEngineers Australia (2013)Google Scholar[25]S.R. Abt, T.E. Brisbane, D.M. Frick, C.A. McKnightTrash rack blockage in supercritical flowJ. Hydraul. Eng., 118 (12) (1992), pp. 1692-1696View Record in ScopusGoogle Scholar[26]E. Rigby, M. Boyd, S. Roso, P. Silveri, A. Davis, Causes and effects of culvert blockage during large storms, in: Global solutions for urban drainage, 2002, pp. 1–16.Google Scholar[27]S. Roso, M. Boyd, E. Rigby, R. VanDrie“Prediction of increased flooding in urban catchments due to debris blockage and flow diversionsProceedings Novatech (2004), pp. 8-13View Record in ScopusGoogle Scholar[28]C.-D. Jan, C.-L. ChenDebris flows caused by Typhoon Herb in Taiwanin Debris-Flow Hazards and Related Phenomena, Springer (2005), pp. 539-563CrossRefGoogle Scholar[29]L.W. Zevenbergen, P.F. Lagasse, P.E. Clopper, Effects of debris on bridge pier scour, in: World Environmental and Water Resources Congress 2007: Restoring Our Natural Habitat, 2007, pp. 1–10.Google Scholar[30]A. Barthelmess, E. Rigby, Estimating Culvert and Bridge Blockages-a Simplified Procedure, 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, Engineers Australia, 2011, pp. 39.Google Scholar[31]E. Rigby, A. Barthelmess, Culvert Blockage Mechanisms and their Impact on Flood Behaviour, 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, Engineers Australia, 2011, pp. 380.Google Scholar[32]G. Streftaris, N. Wallerstein, G. Gibson, S. ArthurModeling probability of blockage at culvert trash screens using Bayesian approachJ. Hydraul. Eng., 139 (7) (2012), pp. 716-726Google Scholar[33]R. Jaeger, T. LuckeInvestigating the relationship between rainfall intensity, catchment vegetation and debris mobilityInt. J. GEOMATE, 12 (33) (2017), pp. 22-29 Download PDFView Record in ScopusGoogle Scholar[34]S. Amiraslani, J. Fahimi, H. Mehdinezhad, The Numerical Investigation of Free Falling Jet’s Effect On the Scour of Plunge Pool, in: XVIII International conference on water resources, Tehran University, Iran, 2008.Google Scholar[35]A.W. Nielsen, X. Liu, B.M. Sumer, J. FredsøeFlow and bed shear stresses in scour protections around a pile in a currentCoast. Eng., 72 (2013), pp. 20-38ArticleDownload PDFView Record in ScopusGoogle Scholar[36]G. Epely-Chauvin, G. De Cesare, S. SchwindtNumerical modelling of plunge pool scour evolution in non-cohesive sedimentsEng. Appl. Comput. Fluid Mech., 8 (4) (2014), pp. 477-487 Download PDFCrossRefView Record in ScopusGoogle Scholar[37]H. Karami, H. Basser, A. Ardeshir, S.H. HosseiniVerification of numerical study of scour around spur dikes using experimental dataWater Environ. J., 28 (1) (2014), pp. 124-134CrossRefView Record in ScopusGoogle Scholar[38]S.-H. Oh, K.S. Lee, W.-M. JeongThree-dimensional experiment and numerical simulation of the discharge performance of sluice passageway for tidal power plantRenew. Energy, 92 (2016), pp. 462-473ArticleDownload PDFView Record in ScopusGoogle Scholar[39]M.A. Khodier, B.P. TullisExperimental and computational comparison of baffled-culvert hydrodynamics for fish passageJ. Appl. Water Eng. Res. (2017), pp. 1-9CrossRefView Record in ScopusGoogle Scholar[40]F.S. Inc., FLOW-3D user’s manual, Flow Science, Inc., 2009.Google Scholar[41]G. Wei, J. Brethour, M. Grünzner, J. BurnhamSedimentation scour modelFlow Science Report, 7 (2014), pp. 1-29View Record in ScopusGoogle Scholar[42]R. Soulsby, R. Whitehouse, Threshold of sediment motion in coastal environments, in: Pacific Coasts and Ports’ 97: Proceedings of the 13th Australasian Coastal and Ocean Engineering Conference and the 6th Australasian Port and Harbour Conference, vol. 1, Centre for Advanced Engineering, University of Canterbury, 1997, pp. 145.Google Scholar[43]L.C.v. RijnSediment transport, part II: suspended load transportJ. Hydraul. Eng., 110 (11) (1984), pp. 1613-1641View Record in ScopusGoogle Scholar[44]S Y LIMScour below unsubmerged full-flowing culvert outletsProc. Instit. Civil Eng. – Water Marit. Energy, 112 (2) (1995), pp. 136-149 Record in ScopusGoogle Scholar

Peer review under responsibility of Faculty of Engineering, Alexandria University.

Figure 6. Scour depth (in negative value) at different views around pier

Three-dimensional numerical simulation of local scour around circular bridge pier using Flow-3D software

Flow-3D 소프트웨어를 이용한 원형 교각 주변 지역 scour의 3 차원 수치 시뮬레이션

To cite this article: Halah Kais Jalal and Waqed H. Hassan 2020 IOP Conf. Ser.: Mater. Sci. Eng. 745 012150

Halah Kais Jalal1
, Waqed H. Hassan2
1 Graduate student, Civil Engineering Department, University of Kerbala, Kerbala, Iraq.
2 Professor, University of Kerbala, Kerbala, Iraq.


주어진 값의 내부 드리프트를 나타내는 다항식 순서 또는 자체 정의 함수 목록을 제공 할 수 있습니다. 이 드리프트는 kriging 보간 동안 내부적으로 적합합니다. 다음에서는 선형 드리프트가 추가된 인공 데이터를 생성합니다. 그런 다음 결과 샘플은 Universal kriging의 입력으로 사용됩니다. 그런 다음 보간 중에 “선형”드리프트가 추정됩니다. 추정된 평균 / 드리프트에만 액세스하기 위해 호출 루틴에 스위치 only_mean을 제공합니다. 원형 교각 주변의 국부 수색 문제는 Flow-3D 모델을 사용하여 전산 유체 역학 (CFD)에서 국부적 진화를 나타냅니다. 교각 설계에서 중요한 scour 및 scour 구멍의 최대 깊이. 이 연구의 목적은 교각 주변의 수색 깊이를 정확하게 시뮬레이션하고 예측하는 수치 시뮬레이션 모델 Flow-3D의 능력을 검증하는 것입니다. 이 검증은 수치 결과를 Melville 실험실 실험 모델과 비교하여 수행됩니다. 30 분후 수치 결과에서 얻은 원형 부두 주변의 최대 scour 깊이는 3.6cm이고 Melville 모델에서 얻은 scour 깊이는 4cm입니다. 이 결과에 따르면 수치 모델과 실험 모델 간의 오류율 비율은 10 %에 가깝습니다. 결과는 실험 결과와 함께 좋은 검증을 보여주었습니다. 마지막으로 제안 된 Flow-3D 모델은 교각 주변의 수색 깊이를 예측하고 시뮬레이션 하는데 효과적인 도구를 고려하고 잠재적인 결과를 예측하는 경제적인 방법을 고려했습니다.

The problem of local scouring around circular bridge pier has been studied numerically
by Computational Fluid Dynamics (CFD) using Flow-3D model to represent the evolution of local
scour and the maximum depth of the scour hole which is important in the bridge pier design. The
aim of this study is to verify the ability of the numerical simulation model Flow-3D to accurately
simulate and predict the scour depth around the bridge pier. This verification is conducted by
comparison the numerical results with Melville laboratory experimental model. The maximum
scours depth around the circular pier obtained from numerical results after 30 min is 3.6 cm, while
the scouring depth obtained from Melville model is 4 cm. According to these results, the error rate
ratio between the numerical and experimental models is close to 10%. The results showed a good
validation with experimental results. Finally, the proposed Flow-3D model considered an effective
tool in predicting and simulating the scour depth around bridge pier and considered an economic
method to predict potential results.
Keywords: Local scour, Flow-3D, CFD, Verfication

scour은 흐르는 물의 침식 작용으로 정의 할 수 있으며, 이는 가까운 교각 및 교각에서 베드를 제거하고 침식합니다 [1]. 다리의 교각 주변을 scour하는 것은 다리의 실패 원인이 충돌 및 과부하와 함께 엄청난 인명 손실과 경제적 영향으로 이어지는 주요 원인 중 하나로 간주됩니다 [2], 지역 scour 예측, 특히 최대 scour 깊이는 다음과 같습니다.

교량 설계, 유지 보수 및 평가에 필수적입니다. 전 세계의 많은 연구자들은 다양한 관점과 다양한 조건에서 광범위하게 scour 문제를 연구했습니다.

교량 부지에서 만든 scour에는 일반적으로 세 가지 유형이 포함되어 있습니다. 일반 scour, 수축 scour 및 국부 scour [3], 세 가지 scour 유형 중, scour는 다리와 관련된 위험에서 가장 중요한 역할을 하기 때문에, local scour는 이 연구의 중요한 부분으로 간주됩니다.

많은 선행 연구가 경험적 테스트를 사용하여 교량의 국부 scour을 분석하는 기술과 방법론을 목표로 했습니다 [4], [5], [6], [7], [8], [9], [10], [11] . 이러한 경험적 scour 테스트의 대부분은 비용이 많이 들고 노동 집약적이기 때문에 크고 중요한 교량에서 종종 수행됩니다.

그러나 가장 인기 있는 고속도로 교량의 경우 경험적 테스트가 적용되지 않지만 이러한 일반 교량에서 scour이 자주 발생하지만 일부 연구에서는 경제적이고 실용적인 목적으로 교량 scour에 대한 분석 솔루션을 조사했습니다.

지난 몇 년 동안 전산 유체 역학 (CFD를 사용하여 산업 및 환경 응용 분야에서 유체 흐름 동작을 결정하는 데 사용)을 더 많이 사용할 수 있는 컴퓨터 및 소프트웨어의 기능이 증가함에 따라 scour의 3 차원 시뮬레이션 방법이 더욱 널리 보급되었습니다.

FLUENT, CFX, PHOENIX와 같은 CFD 소프트웨어는 실험 설정과 여러면에서 유사하므로 이 수치 시뮬레이션의 원래 개념은 속도계와 같은 확장된 부속품을 사용하여 물리적 모델을 설계하고 구성하는 것입니다. 복잡한 모델 실험실 조건에서 모델링하기 어려운 모델은 수치 시뮬레이션을 사용하여 간단하게 모델링 할 수 있습니다.

좋은 수치 모델은 확실히 모델 테스트를 보완 할 수 있으며 설계 엔지니어가 모델 테스트를 수행 할 수 있는 가장 중요한 사례를 식별하는 데 도움이 될 수 있다는 것이 널리 알려져 있습니다.

복잡한 문제와 대규모 모델 연구를 해결할 수 있는 매력적인 아이디어입니다. 실제 결과를 결정하기 위해 추가 작업자 또는 기존의 대규모 설정이 필요하지 않습니다.

CFD (Computational Fluid Dynamics) 방법은 Navier-Stokes의 이산화 및 해석과 계산 셀의 연속성 방정식을 통해 유동 프로세스 시뮬레이션에 항상 사용됩니다. 현재 연구에서 상용 코드 Flow-3D는 교각 주변의 scour 깊이를 모델링하는 데 사용됩니다.

Flow-3D 모델은 유압 공학 응용을 위한 특수 장치가 있는 CFD 패키지입니다. 수치 기법은 다중 스케일 다중 물리 흐름 문제를 얻기 위해 과도 및 3 차원 솔루션에 대한 유체 운동 방정식을 해결하는 데 사용됩니다.

물리적 옵션과 수치 옵션의 조합을 통해 사용자는 Flow-3D를 광범위한 유체 흐름 및 열 전달 현상에 적용 할 수 있으며 다양한 유압 문제를 해결하는 데 널리 사용됩니다 [12]. Flow-3D에 의한 scour의 수치 시뮬레이션은 많은 연구자들에 의해 제안 되었습니다.

Flow-3D에 의한 Scour의 수치 시뮬레이션은 많은 연구자들에 의해 제안 되었습니다.

예를 들어, [13]은 Scour Hole 내의 원형 브리지 부두의 기초에서 발생하는 흐름을 시뮬레이션하기 위해 Flow-3D를 사용했고, [14]는 조수 아래의 복잡한 브리지 피어에서 국소 스캐닝을 시뮬레이션하기 위해 숫자 모델을 사용했고 [15]는 Flow-3D를 사용했습니다.다양한 조건에서 국부적 골절 깊이의 더미 모양과 [16] CFD 코드를 사용하여 3D 흐름과 다양한 모양의 교량 부두 주위의 국부적 스캐닝을 시뮬레이션했습니다.

이 모든 연구는 맑은 물 조건에서 흐르는 물이 주로 흐름과 강바닥 사이의 대부분의 상호 작용으로 이어진다는 가설을 세웠습니다.

본 논문에서는 [4]의 실험실 모델에 의한 수치 시뮬레이션 검증을 통해 교량 주변의 국부 scour 실험 결과를 CFD 코드 Flow-3D의 수치 시뮬레이션 결과와 비교하여 검증을 목적으로 합니다. 이 검증의 주요 목적은 교량 부두 주변의 scour 깊이를 예측할 때 수치 모델 Flow-3D의 효과를 테스트하는 것입니다.

Figure 1. Plan view of Melville experimental setup [4]
Figure 1. Plan view of Melville experimental setup [4]
Figure 2. Geometry of the numerical model configured by the FLOW-3D
Figure 2. Geometry of the numerical model configured by the FLOW-3D
Figure 3. Effect of Cell Size on Scour Depth
Figure 3. Effect of Cell Size on Scour Depth
Figure 4. Meshing Plane Structure Around a Circular Pier
Figure 4. Meshing Plane Structure Around a Circular Pier
Figure 6. Scour depth (in negative value) at different views around pier
Figure 6. Scour depth (in negative value) at different views around pier
Figure 7. Contour Lines Represented the Depth of Scour Around Circular Bridge Pier for Melville Model
Figure 7. Contour Lines Represented the Depth of Scour Around Circular Bridge Pier for Melville Model
Figure 8. Contour Lines Represented the Depth of Scour Around the bridge Pier for the Numerical model
Figure 8. Contour Lines Represented the Depth of Scour Around the bridge Pier for the Numerical model
Figure 9. Scour depth against time around cylindrical pier.
Figure 9. Scour depth against time around cylindrical pier.
Figure 10. Contour map of flow velocity around a pier at 30 min resulted by Melville [4]
Figure 10. Contour map of flow velocity around a pier at 30 min resulted by Melville [4]
Figure 11. Contour map of flow velocity distribution around a pier at 30 min resulted by numerical simulation.
Figure 11. Contour map of flow velocity distribution around a pier at 30 min resulted by numerical simulation.


이 연구는 교각에서 scour깊이의 발달을 예측하는 데 있어 이 수치 시뮬레이션의 효과를 검증하는 것을 목표로 합니다. 검증은 30 분의 scour 깊이 공식화 후 Flow-3D의 수치 결과를 Melville 실험 모델과 비교하여 결론을 내립니다.

결과의 비교는 최대 수세공 깊이에 대한 오류율이 10 %임을 나타내며,이 관찰은 수치 및 실험 작업 사이에 좋은 검증을 보여 주므로 수치 시뮬레이션은 scour 깊이를 성공적으로 재현합니다.

이러한 결과에 따르면 제안된 수치 모델 Flow-3D는 교각 주변의 scour 깊이와 유동장을 시뮬레이션하고 예측하는데 효과적인 도구로 간주되었습니다.

[1] Breusers Nicollet and Shen 1977 Local scour around cylindrical piers Journal of Hydraulic
Research, IAHR,15 (3): 211-252.
[2] Shepherd R. and Frost J D 1995 Failures in civil engineering: Structural, foundation and
geoenvironmental case studies Journal of Hydraulic Engineering, Puolisher ASCE.
[3] Cheremisinoff N P and Cheng S L 1987 Hydraulic mechanics 2 Civil Engineering Practice,
Technomic Published Company, Lancaster, Pennsylvania, U.S.A. 780 p.
[4] Melville B W 1975 Local scour at bridge sites University of Auckland, New Zealand, phd. Thesis,
Dept. of Civil eng., Rep. No. 117.
[5] Abdul-Nour M 1990 Scouring depth around multiple M.Sc. Thesis , Department of Irrigation and
Drainage , University of Baghdad.
[6] Hosny M M 1995 Experimental study of local scour around circular bridge piers in cohesive soils
Colorado State University, Fort Collins.
[7] Ansari S A Kothyari U C and Ranga Raju K G 2002 Influence of cohesion on scour around bridge
piers Journal of Hydraulic Research, IAHR, pp. 40(6): 717-729.
[8] Khsaf S I 2010 A study of scour around Al-Kufa bridge piers Kufa Engineering
Journal.Vol.1No.1,2010, University of Kufa / College Engineering / Civil Department.
[9] Hassan W H Jassem M H and Mohammed S S 2018 A GA-HP Model for the Optimal Design of
Sewer Networks Water Resour. Manag., vol. 32, no. 3, pp. 865–879.
[10] Hassan W H 2017 Application of a genetic algorithm for the optimization of a cutoff wall under
hydraulic structures J. Appl. Water Eng. Res., vol. 5, no. 1, pp. 22–30, Jan.
[11] Ataie-Ashtiani B 2013 Flow field around single and tandem piers Flow Turbulence and Combustion
Journal of Hydraulic Engineering,volume 9429.
[12] Flow -3D manual 2014 Flow-3D user manual version 11, Flow Science Santa Fe, NM.
[13] Richardson J E and Panchang V G 1998 Three-Dimensional Simulation of Scour Inducing Flow at
Bridge Piers Journal of Hydraulic Engineering, 124(5), pp. 530–540. doi: 10.1061/(asce)0733-
[14] Vasquez J and Walsh B 2009 CFD simulation of local scour in complex piers under tidal flow
Proceedings of the thirty-third IAHR Congress: Water Engineering for a Sustainable Environment,
(604), pp. 913–920.
[15] W H H and Halah k Jalal 2019 Effect of Bridge Pier Shape on Depth of Scour Iop, Conf. Ser.,(under
[16] Obeid Z H 2016 3D numerical simulation of local scouring and velocity distributions around bridge
piers with different shapes A Peer Reviewed International Journal of Asian Academic Research
Associates, 20(16), p. 2801. doi: 10.1186/1757-7241-20-67.
[17] Drikakis D 2003 Advances in turbulent flow computations using high-resolution methods Progress
in Aerospace Sciences, 39(6–7), pp. 405–424. doi: 10.1016/S03760421(03)00075-7.
[18] Yakhot and Orszag 1986 Renormalization Group Analysis of Turbulence, Basic Theory Journal of
Scientific Computing, pp. 3–51. 1, pp. 3–51.
[19] Mastbergen D R and Van Den Berg J H 2003 Breaching in fine sands and the generation of
sustained turbidity currents in submarine canyons Sedimentology, 50(4), pp. 625–637. doi:
[20] Soulsby R L and Whitehouse R J S W 1997 Threshold of sediment motion in Coastal Environments
Proc. Combined Australian Coastal Engineering and Port Conference, EA, pp. 149-154.
[21] Meyer-Peter E and Müller R 1948 Formulas for bed-load transport Proceedings of the 2nd Meeting
of the International Association for Hydraulic Structures Research, 39– 64.
[22] Wei G Brethour J Grünzner M and Burnham J 2014 Sedimentation Scour Model Flow Science
Report 03-14.

The Sedimentation Scour Model [침전 세굴(쇄굴) 모델]

1. Introduction
The three-dimensional sediment scour model for non-cohesive soils was first introduced to FLOW-3D in Version 8.0 to simulate sediment erosion and deposition (Brethour, 2003). It was coupled with the three-dimensional fluid dynamics and considered entrainment, drifting and settling of sediment grains. In Version 9.4 the model was improved by introducing bedload transport and multiple sediment species (Brethour and Burnham, 2010). Although applications were successfully simulated, a major limitation of the model was the approximate treatment of the interface between the packed and suspended sediments. The packed bed was represented by scalars rather than FAVORTM (Fractional Area Volume Obstacle Representation, the standard treatment for solid components in FLOW-3D). As a result, limited information about the packed bed interface was available. That made accurate calculation of bed shear stress, a critical factor determining the model accuracy, challenging.

In this work, the 3D sediment scour model is mostly redeveloped and rewritten. The model is still fully coupled with fluid flow, allows multiple non-cohesive species and considers entrainment, deposition, bedload transport and suspended load transport. The fundamental difference from the old model is that the packed bed is described by the FAVORTM technique. At each time step, area and volume fractions describing the packed sediments are calculated throughout the domain. In the mesh cells at the bed interface, the location, orientation and area of the interface are calculated and used to determine the bed shear stress, the critical Shields parameter, the erosion rate and the bedload transport rate. Bed shear stress is evaluated using the standard wall function with consideration of bed surface roughness that is related to the median grain size d50. A sub-mesh method is developed and implemented to calculate bedload transport. Computation of erosion considers entrainment and deposition simultaneously in addition to bedload transport.

Furthermore, a shallow-water sediment scour model is developed in this work by adapting the new 3D model. It is coupled with the 2D shallow water flows to calculate depth-averaged properties for both suspended and packed sediments. Its main differences from the 3D model are 1) the settling velocity of grains is calculated using an existing equation instead of the drift-flux approach in the 3D model, and 2) turbulent bed shear stress is calculated using a well-accepted quadratic law rather than the log wall function. The drag coefficient for the bed shear stress is either user-given or locally evaluated using the water depth and the bed surface roughness that is proportional to d50 of the bed material. The following sections present the sediment theory used in the model and application and validation cases.