Development of an analysis methodology for pressure flow scour under flooded bridge decks using commercial CFD software

Bridges are a significant component of the ground transportation infrastructure in the United States. With about sixty percent of bridge failures due to hydraulic causes, primarily scour, application of computational fluid dynamics (CFD) analysis techniques to the assessment of risk of bridge failure under flood conditions can provide increased accuracy in scour risk assessment at a relatively low cost. The analysis can be used to make optimum use of limited federal and state funds available to maintain and replace bridges and ensure public safety while traveling on the nation's roads and highways during and after floods. Scour is the erosion of riverbed material during high flow conditions, such as floods. When scouring of the supporting soil around the piers and abutments of bridges takes place, risk of bridge failure increases. A simulation methodology to conservatively predict equilibrium shape and size of the scour hole under pressure flow conditions for flooded bridge decks using commercial CFD software was developed. The computational methodology has been developed using C++ to compute changes in the bed contour outside of the CFD software and generate a re-meshing script to change the bed boundary contour. STAR-CD was used to run the hydrodynamic analysis to obtain bed shear stress, and a BASH script was developed to automate cycling between computing bed shear stress with the CFD software and computing changes in the bed contour due to scour predicted using the computed shear stress for the current bed contour. A single-phase moving boundary formulation has been developed to compute the equilibrium scour hole contour that proceeds through a series of quasi-steady CFD computations. It is based on CFD analysis of the flow fields around the flooded bridge deck and shear stress computed at the bed modeled as a rough wall. A high Reynolds number κ-ε turbulence model with standard wall functions, based on a Reynolds-Averaged Navier-Stokes (RANS) turbulence model, was used to compute bed shear stress. The scour sites on the bed were identified as those sites where the computed shear stress exceeded the critical shear stress computed from a published correlation for flat bed conditions. Comparison with experimental data obtained from the Turner-Fairbank Highway Research Center (TFHRC), McLean, VA, USA, revealed larger discrepancies than anticipated between the bridge inundation ratio and the scour hole depth. Although scour hole slopes were small for the cases tested, a correction to critical shear stress to account for bed slope was also tested. It did not significantly improve the correlation between CFD prediction and experimental observations. These results may be a consequence of using only excess shear stress above critical as a criteria for scour when other physical mechanisms also contribute to the initiation of scour. Prediction of scour depth using federal guidelines over predicts scour depth by as much as an order of magnitude in some cases. Over prediction is acceptable for purposes of ensuring bridge safety. CFD methods for scour prediction can be a significant improvement of current methods as long as under prediction of scour depth is avoided. Conservative scour prediction using CFD methods can be achieved by using conservative values of parameters such as critical shear stress and effective bed roughness.