Optimal Design of Stormwater Collection Networks Considering Hydraulic Performance and BMPs

Abstract

Stormwater collection networks are assumed to be a critical infrastructure for cities. Performance issues in such systems could cause major problems, such as flooding. Moreover, having a stormwater collection network is not enough to avoid flooding. Mostly, flooding occurs because of decrease in network performance over the time. City managers normally try to improve the network performance by applying best management practices (BMPs) to the existing network. This approach is known to be very time consuming and expensive for municipalities. In this paper, simultaneous application of BMPs with network design is evaluated. Size of the channel and BMPs area and location are designed, and optimized using the NSGA-II multi-objective genetic algorithm. Results for case study of the Tehran stormwater network showed that by designing the network with no BMP (simple design), the maximum reliability that can be achieved is about 99%. To apply BMPs on the existing network (FPMB design) using Nash bargaining method, the best answer of simple design Pareto front is selected to apply BMPs on it. By applying BMPS, reliability reaches to 100%. Finally for the third design, BMPs and network are designed simultaneously (BMP design) and Pareto front is obtained. It can be observed that all the points from the Pareto front when the network and BMPs are designed simultaneously are considered non-dominant comparing to the two previous designs. Therefore, it can be concluded that designing the network and best management practices simultaneously results in an increase in reliability and a decrease in costs.