Polyvinyl chloride (PVC) pipes are used extensively in water infrastructure due to their lightweight, low cost, and ease of jointing. Failure occurrences in PVC pipes are attributed to poorly manufactured pipes, bad installation, excessive operating conditions, or third-party damage. Data-driven modeling techniques have been widely used in simulating and solving water infrastructure problems, specifically when the collected data are limited. The objective of this paper is to develop a data-driven modeling system that utilizes computational approaches and provides the analytical underpinnings to predict future PVC pipe failures. The system is based on data from El Pedregal City in Peru, simulation and regression algorithms, and supported by easy-to-perceive schematic representations. The hydraulic pressure and flow rate data are streamlined and fed to the regression machine. Subsequent to successive simulation iterations and various polynomial functions, the best fit model is selected. The efficacy of the model is investigated via different performance metrics, in tandem with a residual analysis scheme. The validation results revealed the robustness of the model with mean absolute error (MAE) of 0.35. The proposed model is a predictive tool that can be used by utility managers as a proactive measure against future pipeline bursts or leaks.

聚氯乙烯（PVC）管道轻巧，成本低且易于连接，因此广泛用于水基础设施中。PVC管道发生故障的原因是管道制造不良，安装不当，过度的操作条件或第三方损坏。数据驱动的建模技术已广泛用于模拟和解决水基础设施问题，特别是在收集的数据有限的情况下。本文的目的是开发一种数据驱动的建模系统，该系统利用计算方法并提供分析基础来预测未来的PVC管道故障。该系统基于秘鲁El Pedregal市的数据，模拟和回归算法，并以易于理解的示意图表示为支持。精简了液压和流量数据，并将其输入到回归机中。在进行连续的仿真迭代和各种多项式函数之后，选择最佳拟合模型。结合残差分析方案，通过不同的性能指标来研究模型的有效性。验证结果显示模型的鲁棒性为0.35，平均绝对误差（MAE）。所提出的模型是一种预测工具，公用事业管理者可以将其用作预防未来管道爆裂或泄漏的主动措施。验证结果表明模型的鲁棒性为0.35，平均绝对误差（MAE）。所提出的模型是一种预测工具，公用事业管理者可以将其用作预防未来管道爆裂或泄漏的主动措施。验证结果表明模型的鲁棒性为0.35，平均绝对误差（MAE）。所提出的模型是一种预测工具，公用事业管理者可以将其用作预防未来管道爆裂或泄漏的主动措施。