Metal Culvert Renewal Prioritization Framework Using Fuzzy Inference System

Saskatchewan boasts an extensive road network, spanning over 250,000 km and featuring more than 26,500 culverts that facilitate the flow of water. Ensuring the timely replacement of these culverts is of paramount importance to prevent potential disasters and safeguard water quality. This research focuses on the prioritization of infrastructure renewal, with a particular emphasis on metal culverts. This study undertakes a rigorous evaluation of two statistical and computational methods—ordinal logistic regression and a fuzzy inference system—through an exhaustive examination of 1000 metal culverts situated along Saskatchewan’s highway network. Performance evaluation metrics, including area under the curve, percentage of correct predictions, confusion matrix analysis, accuracy, precision, recall, and the F1 score, consistently highlight the fuzzy inference system method as the superior and most efficacious approach for prioritizing culvert renewal within this crucial infrastructure context.