Surrogate Model for Predicting Stresses and Strains of Buried PVC Pipes in Water Distribution Networks

This study presents the development of a surrogate model for evaluating the mechanical behavior of buried polyvinyl chloride (PVC) pipes, widely used in the water industry for conveying drinking water under pressure due to their durability, corrosion resistance, and cost-effectiveness. Through a series of experiments simulating buried water pipes subjected to internal pressures and external loads, a set of finite-element numerical models was calibrated against physical data. The calibrated models were then used to generate more than 7,500 synthetic scenarios, accounting for various pipe diameters, burial depths, soil and pipe properties, and load conditions. The results were employed to train an evolutionary polynomial regression model, providing closed-form equations for predicting axial and hoop stresses and strains around PVC pipes. The developed equations offer a highly efficient alternative to complex numerical modeling, calculating stresses and strains in seconds. This noninvasive approach allows utilities to assess pipe condition and predict failure risk, considering a range of factors such as soil conditions, pressure variations, and external loading changes. The model’s adaptability supports long-term asset management for buried pipelines.