Constrained optimization model of the volume of initial rainwater storage tank based on ANN and PSO

Environ Sci Pollut Res Int. 2020 Jun;27(17):21057-21070. doi: 10.1007/s11356-020-08630-6. Epub 2020 Apr 7.

Abstract

Rainfall runoff pollution is one of the main causes of water quality deterioration in urban water system. Setting up initial rainwater storage tank could be one of the rapid and effective methods to control runoff pollution. In order to speed up the water environment management processes, the Chinese government has adopted the Public-Private-Partnership (PPP) mode in the water environment treatment to deal with the shortage of funds. Ensuring water quality and controlling water environment management cost are key to PPP projects. Therefore, factors such as pollutant accumulation characteristics of the catchment, land space availability, sewage treatment plants capacity, and river water management cost should be considered during the design of the initial rainwater storage tank on the premise of ensuring water quality. The empirical design method can hardly meet these requirements simultaneously. Under the background of PPP water environment treatment project, a constrained economic optimization model of the initial rainwater storage tank was presented in this paper. The relationship between the total cost of the water environment management and the interception rate of the initial rainwater storage tanks was established by means of Artificial Neural Network (ANN), while the penalty function was used to transform the constrained optimization problem into an unconstrained optimization problem. The interception rate of the initial rainwater storage tanks was then optimized by means of Particle Swarm Optimization (PSO), and the designed volume of the storage tanks was calculated according to the relationship between the interception rate of the storage tank and the cumulative runoff of the related catchment. Finally, a case study of a PPP demonstration project in a plain city in China was conducted. The results demonstrated that compared with the specification method, the total volume of the initial rainwater storage tank increased by 38.7%, the interception rate increased by 68.4%, and the total cost of river water treatment decreased by 5.7% under the constraints of land space availability and sewage treatment capacity. In addition, the optimized method proposed in this paper could reflect the pollutant accumulation characteristics of the catchment. It not only reduce the total cost of the water environment management but also effectively reduce the impact of non-point source pollution on urban water system, and could be more widely used in other areas and PPP projects.

Keywords: ANN; Constrained optimization model; PPP; PSO; Plain city; The initial rainwater storage tank.

MeSH terms

  • China
  • Cities
  • Motor Vehicles
  • Neural Networks, Computer
  • Rain*
  • Water Supply*