Evaluation of empirical equations to predict bridge pier scour in a non-cohesive bed under clear-water conditions

Abstract

Pier scour has been cited as the main mechanism responsible for the failure of bridges spanning rivers. Despite extensive research since the 1950s, there is no universally agreed upon procedure to accurately predict the equilibrium scour depth. Experimental data was generated by 48 tests with four flows and three pier shapes to evaluate the capability of 30 empirical equations to predict the local scour depth. Fine sand and crushed peach pips were used to address the scaling challenges of the equations by means of an equivalent movability number. The equations yielded a wide range of mostly unreliable results, particularly for the non-cylindrical pier shapes. Nevertheless, the HEC-18 models are recommended, in conjunction with Shen et al (1969), and Ali and Karim (2002), because they rely on the pier Reynolds number, a parameter which is significant in the vortex formation. Prediction models taking the horseshoe vortex into consideration could offer better scour depth predictions. Field data was analysed to improve the HEC-18 equation with new factors for pier shape and armouring for different confidence intervals. The armouring factor is based on the particle Reynolds number as opposed to the widely adopted critical velocity, and achieves considerably less scatter about the line of equality despite under-predictions for the cylindrical piers. Alternatively, a diagram comparable to the Modified Liu Diagram has the potential to predict bridge pier scour even though the pier structure parameters are omitted. Further research and improved prediction models should be considered, particularly advanced numerical models which are becoming increasingly feasible.