Evaluating Parameter Uncertainty of Empirical Equations for Assessing Local Scour around Bridge Piers

Abstract

This thesis aims to assist engineers and practitioners in deciding on the correct mitigation and adaptation measures for bridge scour by understanding the accuracy of scour depth estimations. Local scour on bridge piers is a major cause of bridge damage and failure worldwide, posing significant risks to infrastructure and human life. Recognizing the dangers associated with
physical inspections by divers, this study evaluates empirical and semi-empirical equations for determining scour depth using a comprehensive approach. A database from the US Geological Survey (USGS) comprising 936 field measurements and 569 laboratory measurements is used in
this work. The database includes dimensions of the pier, water flow velocity, water flow depth, angle at which the water hits the pier, soil bed gradation, and measurement apparatus used in the field to collect the measurements. The study methodology was in three phases. Phase one
compared the measured scour depths from the database with those calculated using eight empirical and semi-empirical equations. Scatter plots with boundary lines at ±30%, ±50%, and factors of 1.3 and 1.5 were used to assess accuracy. The database is categorized by scour depth magnitude, bed gradation, and measurement apparatus. Similar analyses were performed for laboratory measurements to compare field and laboratory performance. Phase two involved one-at-a-time sensitivity analysis to determine the uncertainty within the eight scour depth equations. Phase three employed the Global Sensitivity Analysis (GSA) method called PAWN, using the Sensitivity Analysis for Everyone (SAFE) toolbox. PAWN is a moment-independent GSA method that uses the CDF of the output to consider the entire distribution of the model output. PAWN
compares the distribution of outputs when an input variable is fixed at different values and ranks the inputs for each equation from most to least influential, using bootstrapping to ensure analysis robustness. To determine the best-fit distribution for each input parameter, the Akaike Information Criterion (AIC) was used. The thesis concludes by discussing appropriate mitigation and adaptation measures based on the findings, providing significant insights into uncertainties
associated with scour depth estimation, thereby enhancing decision-making in bridge design and maintenance.

Date of Award	30 Sept 2025
Original language	English
Awarding Institution	University of Bristol
Supervisor	Paul J Vardanega (Supervisor), Theo Tryfonas (Supervisor) & Maria Pregnolato (Supervisor)