The study of the temporal and spatial evolution of pollutants in modern water bodies and river water quality management largely depend on river water quality models. These models are usually constructed based on detailed hydrodynamic models but cannot satisfy the high precision and efficiency requirement of model simulations in high-resolution terrain. To solve this problem, we developed a high-resolution comprehensive water quality model based on the graphics processing unit (GPU) acceleration techniques in which hydrodynamics and accompanying contaminants were perfectly coupled. The performance and functionality of the model were improved by using the Compute Unified Device Architecture (CUDA) parallel computing architecture and robust model algorithms. A simulation function of the transport attenuation and interaction among multicomponent pollutants was stablished by solving mass conservation, momentum and chemical concentration equations of the flow process. The simulation results of the pure pollutant reaction process and the transport and reaction process were validated against an analytical solution and the results of a two-dimensional steady water quality model by using the Nash-Sutcliffe efficiency (NSE) coefficient. The coefficients were 0.996 and 0.984, respectively. The water quality of the Xidagou River in Yinchuan was evaluated and analyzed by using the validated high-resolution comprehensive water quality model, and the numerical results of two-dimensional pollutant transport were explored. On the adopted RTX3070d computer, only 0.12 h were needed to simulate the hydrodynamic and water quality evolution process involving 350.000 cells with 5-h model settings. The model achieved a high operation speed and high calculation accuracy and could effectively simulate, predict and evaluate various complex water environment problems. This study provides a reliable tool for further research on the water environment and water ecology and the treatment of black and odorous water bodies.
Read full abstract