Machine Learning Application in Prediction of Scour Around Bridge Piers: A Comprehensive Review

Abstract

Scour is one of the most difficult challenges faced by hydraulic engineers, which refers to the erosion of sediments that surrounds hydraulic structures. In the past, scour prediction has generally relied on physical models as well as empirical formulae. However, these methods may not satisfactorily account for the complex nature of scour processes. Hence, this paper aims to provide a concise overview of the latest advancements in the field of scour prediction, particularly focusing on the use of machine learning (ML) techniques. The review begins by examining the basic ideas and methodologies of various machine learning algorithms which are commonly employed, it then looks into the key factors that affect scour processes, including flow velocity, sediment characteristics, and bed morphology. The paper provides an in-depth assessment of the advantages and drawbacks of current machine learning models used for estimating scour, taking into account various issues such as the availability of data, models understandability, and their capacity to adapt in changing environmental conditions. This study will be a helpful resource for researchers, practitioners, and decision-makers in the field of hydraulic engineering. It provides insights into the evolving field of ML applications for predicting scour and sets the stage for the advancement of more precise and versatile scour prediction models.