Study on SWMM Calibration and Optimizing the Layout of LID Based on Intellgent Algorithm: A Case in Campus

• Published in: Water resources management Vol. 39; no. 10; pp. 5025 - 5039
• Main Authors: Liu, Kehan, Li, Jiake, Gao, Jiayu, Chi, Xiaoqing, Jiang, Chunbo
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.08.2025; Springer Nature B.V
• Subjects: Algorithms; Atmospheric Sciences; Back propagation networks; Calibration; Civil Engineering; Drainage; Earth and Environmental Science; Earth Sciences; Environment; Extreme weather; Facilities management; Floods; Genetic algorithms; Geotechnical Engineering & Applied Earth Sciences; Hydrogeology; Hydrology/Water Resources; Layouts; Neural networks; Optimization algorithms; Optimization models; Overflow; Parameters; Ponding; Rain; Rainfall; Sorting algorithms; Storms; Stormwater; Stormwater management; Urbanization; Water management; Water resources; BP Neural Network; SWMM; Layout Optimization; NSGA-III Algorithm; Overflow
• ISSN: 0920-4741; 1573-1650
• DOI: 10.1007/s11269-025-04187-6

Rapid urbanization has significantly increased impervious surfaces and exacerbated extreme rainfall events, leading to urban overflow and water ponding issues. To address these challenges, Low Impact Development (LID) facilities have been increasingly implemented as effective solutions. Storm Water Management Model (SWMM) model parameter calibration and LID facility layout optimization has emerged as a research focus. This paper establishes a SWMM model for a campus in Xi’an, China. Employing a coupled Back Propagation (BP) neural network for model parameter calibration. The SWMM model is coupled with the Non-dominated Sorting Genetic Algorithm III (NSGA-III) to construct an optimization model. The optimal LID layout scenario is selected using the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) comprehensive evaluation method. Results demonstrate that the BP neural network-coupled calibration approach is both efficient and feasible. The proposed optimization model enhances the convergence of optimization results, and the TOPSIS method successfully identified the optimal solution. This study offers valuable insights for model parameter calibration and facility optimization.