LAUSR

Bridge pier scouring measurement in the field, especially during flood season, is very difficult. This study experimentally investigated the pier scour pattern to find better alternatives to represent the actual field conditions. For this purpose, piers of different shapes (circular and square) and different sizes were modeled in the laboratory. The contour maps were drawn to check the extent of possible damage caused by scouring. Under the same laboratory flow conditions and sediment properties, scour depth resulted from square-shaped piers was more evident compared to circular piers and pier geometry. It was also evident that the scour depth increases with an increase in pier size. The contour maps could reflect different flow conditions, sediment, and pier geometry. The scouring process could identify the extent of remedial measures needed. The models developed were regression, artificial neural network (ANN), and genetic function (GF) based, and the results obtained from these models were compared with the experimental data. From the models comparison based on coefficient of determination ($$R^{2})$$R2), the dimensional variables-based regression, ANN, and GF models depicted $$R^{2}$$R2 values as 0.38, 0.64, and 0.67, respectively, which were inferior to the corresponding $$R^{2}$$R2 values of 0.80, 0.95, and 0.97 for models developed using non-dimensional input variables. Overall, GF-based model performed better than the rest of the models not only because it gave higher $$R^{2}$$R2 values and less measured error but also because it resulted in more simple, compact, and explicit expression for bridge pier scour.