Study on Prediction Method for Compression Scour Depth of River-Crossing Bridge

Abstract

Riverbed deformation caused by river-crossing bridge construction can be divided into compression scour and local scour. Compared with local scour, fewer studies have been made on the compression scour caused by bridge piers. It is noteworthy that, the compression scour can lead to riverbed scour of the whole cross section along a bridge site, which is obviously detrimental to the bridge foundation safety. Based on a summary of existing research findings, a prediction model for the compression scour of bridge piers is constructed, and the model is applied in predicting the compression scour depth of Shiyezhou River Bridge in the lower reaches of the Yangtze River. Firstly, the pier boundary treatment methods at different spatial scales are discussed. Subsequently, the selection method of flow and sediment processes is proposed from the engineering safety point of view, according to the flow and sediment characteristics on the lower reaches of the Yangtze River. Finally, the depth of compression scour around the upstream and downstream of Shiyezhou Bridge piers are predicted, and comparisons were made between the prediction depth of Shiyezhou Bridge and other existing bridges in the lower reaches of the Yangtze River. Comparisons show that the compression scour depth of Shiyezhou Bridge was basically equivalent to that of other bridges downstream the Yangtze River. The results indicate that the method for predicting the compression scour depth of bridge piers is reasonable and feasible, and the prediction of compression scour depth can provide technical basis for determining the embedment depth of the bridge pier foundation.

1 Introduction

In the lower reaches of the Yangtze River, many river-crossing bridges are built to strengthen the communication between cities on both river sides, and to mitigate the river-crossing traffic pressure. Due to the narrowing action of the bridge substructure on the water flow, the riverbed deforms accordingly (WUHES 1981). The riverbed changes caused by bridge can be divided into compression scour and local scour, where the former refers to the scour occurring after the sectional contraction of piers, including natural riverbed scour and the scour arising from the contraction of water flow. As for local scour, it is the erosion, i.e., scour hole, just around piers formed by flow vortex (Zhang et al. 1993). At present, studies mainly focus on local pier scour (Olsen and Kjellesvig 1998, Karim and Ali 2000, Roulund et al. 2005, Liu and Garcia 2006, Chen 2008 and Sumer et al. 2001), while much less attention has been paid on compression scour (Yong and Blair 2010; Fenocchi and Natale 2016; Guo and Qi 2011). In most case, if a pier experiences compression scour apparently, the cross-section of the bridge site will be subjected to overall scour along the bridge, which is disadvantageous for the safety of the building foundation (Guo 2013). Especially on the middle and lower reaches of the Yangtze River, the sediment concentration declines sharply after the impoundment of the Three Gorges Reservoir, which accelerates the downstream riverbed scour. Therefore, the foundation depth of bridge piers should be chosen with an overall consideration. In the design of river-crossing bridges, the prediction of compression scour of piers is one of the key technical problems that must be solved.

Previously, empirical prediction formulas for the compression scour of piers have been acquired through model tests or field measurements. There are some commonly used formulas, including formulas published in railway and highway specifications (MRPRC 1999; MTPRC 2002), American HEC-18 formula (Shirole and Holt 1991), Soviet Baldakov, Lestervan, and Andreev formulas (Lu and Gao 1996). However, those formulas can only reflect the equilibrium value of compression scour under the prolonged action of specific flow and sediment conditions, along with the disadvantage of coefficient uncertainty. In recent years, domestic and foreign scholars have started predicting the compression scour of piers through numerical models. For instance, Liu et al. (1993) simulated the scour of railway bridge piers in the midstream of the Yellow River during a once-in-a-century catastrophic flood, using a method of adjusting the local head loss coefficient of grids. Results showed that the scouring-silting trend on the riverbed surface basically accorded with the test, but the measured values and calculated values were different. Lai et al. (2010), Guo and Qi (2011), Guo (2013), and Due and Rodi (2008) verified rectangular flume experiments on long contraction segments by using numerical simulation, and the model could simulate the variation of water level along the river and the scour distribution in the contraction segment, but sediment process and suspended load were not considered. For the lower reaches of the Yangtze River, due to the impoundment of the Three Gorges Reservoir, flow and sediment processes are obviously different from the natural situation, leading to much complex scouring-silting conditions. Compared with river-crossing tunnels, more factors should be considered for river-crossing bridge (Zhang et al. 2011; Wei et al. 2016; Yue et al. 2010). Therefore, predicting the compression scour of piers, a model should meet the following conditions: (a) flow and sediment processes should be reasonably selected; (b) parameters should be calibrated and verified so that annual and interannual scouring-silting changes of riverways can be rechecked; (c) models should reflect the flow and sediment movements before and after bridge construction well.

Taking a newly built bridge as an example, a mathematical model for compression scour of bridge piers was constructed in this study. The model parameters were verified, the pier boundary was proved, the flow and sediment processes were selected, and then the scour depth was predicted, expecting to provide a reference for solving similar problems.

2 Project Overview

The newly built Lianyungang-to-Zhenjiang Railway, which is 249 km in overall length, continues Xinyi-Changxing Railway in the north, connects Nanjing-Qidong Railway in the middle, and meets Beijing-Shanghai High-Speed Railway and Shanghai-Nanjing Intercity Railway in the south, being the main north-south longitudinal railway passage linking northern, central, and southern Jiangsu regions. This railway is planned to span the Yangtze River at the river reach from Shiyezhou to Wufengshan. Two bridge positions were initially designed, i.e., Shiyezhou bridge site scheme and Wufengshan bridge site scheme. And at last, the bridge scheme crossing over the central bar named Shiyezhou was chosen. The suspension bridge scheme was adopted, where the left-branch main span was 406 m and the right-branch main span was 910 m. The plane layout and elevation layout are as shown in Fig. 1 and Fig. 2.

Fig. 1.

Plan layout of the scheme of Shiyezhou bridge location

Fig. 2.

Elevation layout of the right branch

3 Recent Evolution of the River Reach

Shiyezhou river reach is gently bent and branched, which is divided into left and right branches by Shiyezhou in the middle of the river (Fig. 1). The left branch is the secondary branch, with the measured water diversion ratio of 38.8% in December 2012, and the right branch is the main branch. The overall river regime remains relatively stable due to the levees and the nodes alongside the riverway. But the local evolution is still obvious, which is mainly manifested by the change in the diversion pattern of branches and the adjustment of riverbed.

• (1) The changes in the diversion pattern of branches are mainly embodied in the increasing diversion ratio of the left branch and the declining diversion ratio of the right branch. From the middle 1970s to the early 1990s, the water diversion ratio of the left branch grew slowly, with an annual average increase amplitude of 0.1%. After 1995, the growth rate was evidently elevated because of continuous floods, and the annual growth rate reached 2.8% during 1997–1999. But the growth rate was slowed down again after 2000. The water diversion rate of the left branch in Shiyezhou rose from 20% in the 1970s to about 40% in recent years.

• (2) As for the adjustment of riverbed, the low elevated beach of Shiyezhou experienced sustained scouring and regression; the left branch was subjected to overall scouring, with the local scour depth reaching 4 m; the entrance section of the right branch in Shiyezhou was characterized by side beach scouring at the left side and deep channel silting at the right side, while the middle-lower section of the right branch was manifested as side beach silting at the right edge of Shiyezhou and deep channel scouring. Variation of the transverse profiles at the bridge site are showed in Fig. 3.

Fig. 3.

Transverse profiles at the Shiyezhou bridge site

4 Modeling and Verification

4.1 Governing Equations Set

The governing equations set for the 2D numerical planar flow and sediment model under the orthogonal curvilinear coordinate system is as follows:

$$ \frac{\partial Z}{{\partial t}} + \frac{1}{{C_{\xi } C_{\eta } }}\left[ {\frac{{\partial \left( {C_{\eta } Hu} \right)}}{\partial \xi }} \right. + \left. {\frac{{\partial \left( {C_{\xi } Hv} \right)}}{\partial \eta }} \right] = 0 $$

(1)

• A. Continuity equation of water flow:

  $$ \begin{array}{l} \quad \frac{{\partial \left( {Hu} \right)}}{\partial t} + \frac{1}{{C_{\xi } C_{\eta } }}\left[ {\frac{\partial }{\partial \xi }} \right.\left( {C_{\eta } Huu} \right) + \frac{\partial }{\partial \eta }\left( {C_{\xi } Hvu} \right) + Hvu\frac{{\partial C_{\xi } }}{\partial \eta }\left. { - Hv^{2} \frac{{\partial C_{\eta } }}{\partial \xi }} \right] \hfill \\ = - \frac{{gu\sqrt {u^{2} + v^{2} } }}{{C^{2} }} - \frac{gH}{{C_{\xi } }}\frac{\partial Z}{{\partial \xi }} + \frac{1}{{C_{\xi } C_{\eta } }}\left[ {\frac{\partial }{\partial \xi }} \right.\left( {C_{\eta } H\sigma_{\xi \xi } } \right) + \frac{\partial }{\partial \eta }\left( {C_{\xi } H\sigma_{\eta \xi } } \right) + H\sigma_{\xi \eta } \frac{{\partial C_{\xi } }}{\partial \eta } - \left. {H\sigma_{\eta \eta } \frac{{\partial C_{\eta } }}{\partial \xi }} \right] \hfill \\ \end{array} $$

  (2)

• B. Motion equation of water flow:

  $$ \begin{array}{l} \;\frac{{\partial \left( {Hv} \right)}}{\partial t} + \frac{1}{{C_{\xi } C_{\eta } }}\left[ {\frac{\partial }{\partial \xi }} \right.\left( {C_{\eta } Huv} \right) + \frac{\partial }{\partial \eta }\left( {C_{\xi } Hvv} \right) + Huv\frac{{\partial C_{\eta } }}{\partial \xi }\left. { - Hu^{2} \frac{{\partial C_{\xi } }}{\partial \eta }} \right] \hfill \\ = - \frac{{gv\sqrt {u^{2} + v^{2} } }}{{C^{2} }} - \frac{gH}{{C_{\eta } }}\frac{\partial Z}{{\partial \eta }} + \frac{1}{{C_{\xi } C_{\eta } }}\left[ {\frac{\partial }{\partial \xi }} \right.\left( {C_{\eta } H\sigma_{\xi \eta } } \right) + \frac{\partial }{\partial \eta }\left( {C_{\xi } H\sigma_{\eta \eta } } \right) + H\sigma_{\eta \xi } \frac{{\partial C_{\eta } }}{\partial \xi } - \left. {H\sigma_{\xi \xi } \frac{{\partial C_{\xi } }}{\partial \eta }} \right] \hfill \\ \end{array} $$

  (3)

  where $\left( {\xi ,\eta } \right)$ represents the coordinates of the curvilinear coordinate system; $\left( {x,y} \right)$ stands for physical coordinates; $u$ and $v$ are the flow velocity components in the directions $\xi$ and $\eta$, respectively; $Z$ means the water level; $t$ is time; $H$ refers to the water depth; $C$ signifies the Chezy coefficient ($C = \frac{1}{n}H^{{{1 \mathord{\left/ {\vphantom {1 6}} \right. \kern-\nulldelimiterspace} 6}}}$); $n$ denotes the roughness coefficient; $\sigma_{\xi \xi } ,\,\,\sigma_{\eta \eta } ,\,\,\sigma_{\xi \eta }$ and $\sigma_{\eta \xi }$ indicate turbulent shear stresses; $\nu_{t}$ is the turbulent viscosity coefficient.

• C. Unbalanced sediment transport equation of suspended loads

  $$ \frac{{\partial (HS{}_{i})}}{\partial t} + \frac{1}{{C_{\xi } C_{\eta } }}[\frac{\partial }{\partial \xi }(C_{\eta } HuS{}_{i}) + \frac{\partial }{\partial \eta }(C_{\xi } HvS{}_{i})] = \frac{1}{{C_{\xi } C_{\eta } }}[\frac{\partial }{\partial \xi }(\frac{{\varepsilon_{\xi } }}{{\sigma_{s} }}\frac{{C_{\eta } }}{{C_{\xi } }}H\frac{{\partial S{}_{i}}}{\partial \xi }) + \frac{\partial }{\partial \eta }(\frac{{\varepsilon_{\eta } }}{{\sigma_{s} }}\frac{{C_{\xi } }}{{C_{\eta } }}H\frac{{\partial S{}_{i}}}{\partial \eta })] + \alpha_{i} \omega_{i} (S_{i}^{*} - S{}_{i}) $$

  (4)

  where $S_{i}$ is the sediment content of suspended loads in group $i$; $S_{i}^{*}$ represents the sediment transport capacity of suspended loads in group $i$; $\varepsilon_{\xi }$ and $\varepsilon_{\eta }$ stand for the sediment diffusion coefficients in the directions $\xi$ and $\eta$, respectively; $\sigma_{s}$ is a constant, taken as 1.0; $\alpha_{i}$ denotes the saturation recovery coefficient of suspended loads in group $i$; $\omega_{i}$ means the sediment deposition velocity in group $i$.

• D. Unbalanced transport equation of bed loads:

  $$ \frac{{\partial HS_{{{\text{b}}L}} }}{\partial t} + \frac{1}{{C_{\xi } C_{\eta } }}[\frac{\partial }{\partial \xi }(C_{\eta } HuS_{{{\text{b}}L}} ) + \frac{\partial }{\partial \eta }(C_{\xi } HvS_{{{\text{b}}L}} )] = \alpha_{{{\text{b}}L}} \omega_{{{\text{b}}L}} (S_{{{\text{b}}L}}^{*} - S_{{{\text{b}}L}} ) $$

  (5)

  where $S_{{{\text{b}}L}}$ denotes the sediment concentration of bed loads in group $L$; $S_{{{\text{b}}L}}^{*}$ is the sediment transport capacity of bed loads in group $L$, $S_{{{\text{b}}L}}^{*} = g_{{{\text{b}}L}}^{*} /\left( {\sqrt {u^{2} + v^{2} } h} \right)$, and $g_{{{\text{b}}L}}^{*}$ refers to the sediment transport rate of different bed load groups; $\alpha_{{{\text{b}}L}}$ is the saturation recovery coefficient of bed loads in group $L$; $\omega_{{{\text{b}}L}}$ stands for the deposition velocity of bed loads in group $L$.

• E. Riverbed deformation equation:

  Riverbed deformation caused by suspended load scouring-silting: $\gamma_{0i} \frac{{\partial Z_{i} }}{\partial t} = \alpha_{i} \omega_{i} (S_{i} - S_{i}^{*} )$.

  Riverbed deformation arising from bed load scouring-silting: $\gamma_{{{\text{0b}}L}} \frac{{\partial Z_{{{\text{b}}L}} }}{\partial t} = \alpha_{{{\text{b}}L}} \omega_{{{\text{b}}L}} (S_{{{\text{b}}L}} - S_{{{\text{b}}L}}^{*} )$.

  Total riverbed scouring-silting thickness: $Z = \sum\limits_{i = 1}^{m} {Z_{i} } + \sum\limits_{L = 1}^{n} {Z_{L} }$.

  Where $m$ and $n$ represent the numbers of suspended load groups and bed load groups, respectively; $\gamma_{0i}$ denotes the dry bulk density of sediments in suspended load group $i$; $Z_{i}$ is the riverbed scouring-silting thickness induced by the suspended loads in group $i$; $\gamma_{0bL}$ stands for the dry bulk density of sediments in bed load group $L$; $Z_{L}$ is the total riverbed scouring-silting thickness.

4.2 Model Discretization and Solving

To numerically solve the water flow motion Eqs. (1–3), the equations were discretized using the control volume method with integral conservation, the coupled equations were solved via SIMPLER formula, and such a process was repeatedly iterated until the convergence of the flow field. After discretization, the suspended load transport Eq. (4) and bed load transport Eq. (5) were implicitly solved through the underrelaxation technology and progressive scanning time division multiple address (TDMA) technology. Next, the riverbed deformation Eq. (6) was explicitly solved through finite difference discretization.

4.3 Relevant Parameter Settings

The main parameters of water flow motion equations include roughness coefficient $n$ and turbulent viscosity coefficient $\nu_{t}$, where the former reflects the resistance of natural river, which can be determined by measuring the water level and verifying the flow velocity. The turbulent viscosity coefficient is calculated through $\nu_{t} = k_{t} u_{*} H$, wherein $u_{*}$ is the friction velocity and $k_{t}$ is a constant.

The main parameters of sediment motion equations include sediment transport capacity $S_{i}^{*}$ of different suspended load groups and that ($g_{{{\text{b}}L}}^{*}$) of different bed load groups. $S_{i}^{*} = P_{*} S_{*}$, where $P_{*}$ is the gradation of sediment transport capacity in different groups and $S_{*}$ is the total sediment transport capacity of water flow, which are generally determined through the following Zhang R.J. formula:

$$ S_{*} = k\left( {\frac{{\sqrt {u^{2} + v^{2} } }}{gh\omega }} \right)^{m} $$

(6)

where $\omega$ is the average sediment deposition velocity of suspended loads; $k$ and $m$ are constants; $P_{*}$ can be determined through the method rendered in Appendix B of Literature (Zhang et al. 2011). $g_{{{\text{b}}L}}^{*} = P_{{{\text{b}}L}} \eta_{L} g_{{\text{b}}}^{*}$, where $P_{{{\text{b}}L}}$ is the percentage of sediments (group $L$) in riverbed sediments, $\eta_{L}$ is the camouflage coefficient, and $g_{{\text{b}}}^{*}$ is the total sediment transport rate of bed loads, which can be calculated through the formulas listed in Appendix A of Literature (Zhang et al. 2011).

4.4 Considerations in Pier Calculation

In the past calculations, piers are generalized mainly using local terrain correction, additional roughness coefficient, etc., which can only simulate the water-resisting effect of piers while failing to give the detailed flow field changes near piers or calculate the compression scour of piers. As computer performance is enhanced and numerical methods are improved, piers can be processed by densifying the grids so that the grids are consistent with the pier scale. How to process the pier boundary is the key to accurately simulating the flow and sediment motions near piers, and the models can be divided into small-scale and large-scale models according to the spatial scale simulated.

Small-scale models are usually applied to laboratory pier simulation, with a small grid size (mm level). In this process, the pier boundary can be processed by combining the large coefficient method and wall shear rate, i.e., the grids within the pier (internal grids encircled by grid nodes 1–16 in Fig. 4) are processed through the large coefficient method (if the source item is a large coefficient, its calculation result is taken as 0), while those adjacent to the pier wall (grid nodes 1–13) are processed using wall shear rate.

Large-scale models are generally used for pier simulation in natural riverways with complicated riverway boundaries and meter-level grid scale. When wall shear ratio is used, the distance from the first computing node on the wall surface to the wall surface is $z^{ + } = \frac{{z_{l} u_{*} }}{\upsilon }$, $30 < z^{ + } < 100$; for the lower reaches of the Yangtze River, $z_{l}$ is 0.001 m under the flow velocity of 2 m/s, water depth of 10 m, and roughness coefficient of 0.025, so the scale needed in the calculation is much smaller than the pier scale. Therefore, the wall shear ratio is rarely used in the actual calculation, but instead, the inaccessible conditions and non-slip conditions are directly stipulated on the wall surface of the pier. The practice has proved that this method can simulate the detouring flow phenomenon nearby the pier very well.

Fig. 4.

Schematic diagram of bridge pier boundary treatment

4.5 Model Verification

In this study, a 2D planar flow and sediment model of Shiyazhou river reach was constructed, which reached the three rivers estuary upward and extended Liuxu estuary downward. The overall length of the river section simulated was about 40 km. To improve the model prediction accuracy, the model parameters were repeatedly calibrated and verified as seen in Table 1. The verification results of water level along the river and typical cross-sectional flow velocity are presented in Fig. 4. Restricted by the length of this paper, only the medium and long-term verification results of riverbed scouring-silting were given. In this model, the terrain in October 2011 was taken as the initial terrain and that in November 2015 as the verification terrain. The inlet flow rate and sediment content were the measured data from Datong Station, the outlet water level was acquired by the water-level-flow rate relation curves from Zhenjiang Station, and then this model verified the riverbed scouring-silting of this river reach after 4 complete hydrological years. The calculation results accorded well with the measured scouring-silting position and distribution.

Table 1. Calibration results of 2D flow and sediment model parameters for Shiyezhou

5 Prediction of the Compression Scour of Piers

5.1 Selection of Flow and Sediment Processes

The rationality of prediction results is closely related to the selection of flow and sediment processes. In this study, the determination method of flow and sediment processes was proposed from the partially safe angle according to the inflow flow and sediment characteristics on the lower reaches of the Yangtze River and considering the flow and sediment changes of a downstream hydrological control station (Datong) after the impoundment of the Three Gorges Reservoir.

• (1) Flow and sediment processes in the years of the catastrophic flood: The year of once-in-three hundred years flood was selected as the year of catastrophic flood according to the design criteria (once-in-a-century flood design and once-in-three hundred years flood check) of the Shiyezhou bridge site scheme. The flow process was based on the year 1998 (once-in-a-century), and the peak discharge was amplified according to the frequency. The sediment process corresponding to the year of the catastrophic flood was derived (Fig. 5) using lower envelopes from the angle of engineering safety according to the flow rate-sediment transport rate relation curves of Datong Station after the impoundment of the Three Gorges Reservoir.

• (2) Flow and sediment processes in serial hydrological years: Typical serial years were adopted, and the combination of the years of catastrophic floods was considered. According to the flow and sediment data of Datong Station during 1950–2014, the flow rate of Datong Station changed little since the impoundment of the Three Gorges Reservoir, the years of medium and small floods played a dominant role, but the sediment discharge was evidently smaller than that before the impoundment, which reflected the impounding and sediment trapping effects of the Three Gorges Reservoir. Before the impoundment of the Three Gorges Reservoir, the multi-year sediment discharge was 427 million tons in Datong Station. After the impoundment, the annual average sediment discharge was 143 million tons during 2004–2014, which was 66% lower than that before the impoundment, as shown in Fig. 6. The period of 2007–2010 represented the flow and sediment characteristics during 2004–2014 very well after the construction of the Three Gorges Reservoir. According to statistics, the runoff was 837,200 million m³ during 2007–2010, which was basically equivalent to the average runoff (832,100 million m³) after the construction of the Three Gorges Reservoir, with an annual maximum flow rate of 64,600 m³/s and minimum flow rate of 10,000 m³/s (Fig. 6). Considering the influence of the 1998 flood, the serial hydrological years were finally determined as 2007–2010+1998 (the sediment process was expressed by the lower envelope in Fig. 5(a)).

Fig. 5.

Runoff and sediment of Datong Station

Fig. 6.

Discharge and sediment discharge rate

5.2 Cross-Sectional Hydrodynamic Changes of the Bridge Site After Contraction of Piers

Due to the bridge construction, the water level and flow velocity distribution within a certain range of the bridge area were changed, which would have a direct bearing on the cross-sectional flow rate per unit width of the bridge site. Generally, the discharge per unit width between piers was elevated somehow after the bridge construction (Fig. 7). Impacted by water damming and covering of pier studs, the upstream and downstream flow rate per unit width was reduced. At the left side of the right branch in Shiyazhou, the flow rate per unit width between piers was reduced due to the small pier spacing and large resistance, and meanwhile, the increased amplitude of the flow rate per unit width of the right riverway was large.

Fig. 7.

Transverse distribution of the unit discharge of bridge site sections

5.3 Prediction of the Compression Scour of Piers

After encountering the year of a catastrophic flood, the upstream flow velocity was decelerated due to the water-resisting effect of piers, and riverbed silting dominated; the downstream riverbed experienced both scouring and silting due to the water flow extruding and covering effects of piers; given the presence of piers on the cross-section of this bridge site, the discharge area was narrowed, the flow rate per unit width was elevated, compression scour took place, and the scour amplitude was greater at the left-branch, e.g., N3# pier (9.46 m) and N2# pier (5.92 m), while the cross-sectional scour amplitude was relatively smaller at the right branch, e.g., 2# auxiliary bridge pier (2.81 m) and 3# main pier (3.70 m) (see Fig. 8 and Fig. 9).

Fig. 8.

Erosion and siltation caused by bridge piers after a catastrophic flood year

Fig. 9.

Changes of transverse section after a catastrophic flood year

DaTong-WuSong Estuary sailing charts of the Yangtze River in 1998 and 2013 (Lu 2016) were used to calculate the cross-sectional compression scour (Table 2) at typical bridge sites on the lower reaches of the Yangtze River. It could be seen that due to the different bridge locations and cross-sectional morphologies, the depth of compression scour varied from the bridge to bridge, and the scour amplitude ranged from 2.6 m to 8.0 m, which was equivalent to the prediction result of the Shiyezhou bridge site model.

Table 2. Statistical measured maximum depth of cross-sectional compression scour at typical bridge sites on the lower reaches of the Yangtze river (1998–2013)

6 Conclusions

• (1) A prediction model for the compression scour of piers was constructed. Small-scale models can be processed by combining the large coefficient method and wall shear rate. Wall shear rate is inapplicable to natural large-scale models, and thus inaccessible conditions and non-slip conditions can be directly stipulated on the wall surface of piers. Repeated calibration and validation allow the model parameters to apply to recheck the annual and interannual scouring-silting deformation of riverways.

• (2) The determination method of flow and sediment conditions was proposed and the adverse flow and sediment processes were determined from the angle of engineering safety based on the flow and sediment characteristics on the lower reaches of the Yangtze River and the changes in the flow and sediment conditions of Datong Station since the impoundment of the Three Gorges Reservoir.

• (3) Impacted by bridge piers, the cross-section of the bridge site was subjected to compression scour, where the scour amplitude was great at left-branch N3# pier (9.46 m) and N2# pier (5.92 m), while the cross-sectional scour amplitude was relatively smaller at the right branch. Through the statistical cross-sectional compression scour at the already built bridge sites on the lower reaches of the Yangtze River, the scour depth ranged from 2.6 m to 8.0 m, which was basically equivalent to the model prediction result.