Author(s)

Kaiyuan Zheng¹, Ying Zhang^1,2,3*

¹Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China.
²Kyoto University-Tsinghua University Cooperative Research and Education Center for Environmental Technology, Kyoto, Japan.
³State Environmental Protection Key Laboratory of Microorganism Application and Risk Control, School of Environment, Tsinghua University, Beijing, China.

Aiming at working out more accurate pollutant degradation coefficient of the deep tunnel system, this work puts forward a novel optimized algorithm to calibrate such coefficient and compare it with the ordinary fitting method. This algorithm incorporates the outlier filtration mechanism and the gradient descent mechanism to improve its performance, and the calibration result is substituted into storm water management model (SWMM) source codes to validate its effectiveness between simulated and observed data. COD, NH₃-N, TN and TP are chosen as pollutant indicators of the observed data, and the RMSE, MSE and ME are selected as indicators to present the efficiency. The results show that the outlier filtration mechanism obtains better performance than fitting method, with the gradient descent mechanism nearly reduces 92.42% of the iterative amounts and improves 55 times of the computation efficiency than the ordinary iterative method, such algorithm is expected to function better with substantial observed data.

SWMM, Pollutant Degradation Coefficient, Deep Tunnel System, Optimized Algorithm

Zheng, K. Y. and Zhang, Y. (2024) A Novel Optimization Algorithm for Calibrating Pollutant Degradation Coefficient in Deep Tunnel Based on Storm Water Management Model. Journal of Geoscience and Environment Protection, 12, 207-217. doi: 10.4236/gep.2024.1212013.