Optimal Design Based on Surrogate Reliability Measures for Water Distribution Systems

In this paper, we propose two surrogate reliability indices for a water distribution system (WDS), which are the pipe resilience index (PRI) and the modified pipe resilience index (MPRI), respectively. The proposed indices, as well as the other three existing indices, are applied to two WDSs to optimize the layout. In the optimization model, two objective functions of pipe construction cost and reliability measure were considered with the control variables of pipe diameters. The optimization model was solved by the widely applied nondominated sorting genetic algorithm II (NSGA-II). The results indicated that the two indices proposed are suitable for WDS with various sizes. The MPRI performs better in optimizations of large-scale networks, while the PRI performs better in small-scale networks. In addition, the MPRI values are smaller than the PRI values with the same cost, which indicated that the effect of the pipe diameters' uniformity is greater than the pipe diameter itself.