Pareto-Based Methodology for the Calibration and Uncertainty Analysis of Gated Culvert Flows

Abstract

An improved flow rating equation for full-barrel flows through a gated culvert was developed by calibrating a flow model to measured flow data with the Pareto-based approach to inverse parameter estimation. The computed Pareto front was used to identify the appropriate minimum values of both a measurement objective function that depicted differences between measured and computed flows and a Tikhonov regularization objective function that reflected deviations from preferred parameter values. Agreement between measured and computed values was generally good. Additionally, Pareto-based techniques were used to estimate the 95% confidence intervals of several predicted flow rates associated with a set of measured flows that were not included in the calibration dataset. Overall, this study revealed that Pareto-based parameter estimation techniques are useful for obtaining the best flow rating equation that can be achieved given the available flow data along with the engineer’s knowledge of the flow parameters for the gated culvert.