Water Quality Prediction Model Based on DSAPSO Optimized Radial Basis Function Neural Network

Sewage treatment is a nonlinear and complex biochemical process, which needs reasonable monitoring and control to make it run stably and efficiently. However, online monitoring of relevant water quality parameters is currently difficult, time-consuming and costly. In order to improve the situation and improve the effluent quality, a DSAPSO (Dual Search Adaptive Particle Swarm Optimization) algorithm was proposed to optimize the soft measurement model of effluent quality of RBFNN (Radial Basis Function Neural Network). By comparing the improved algorithm proposed with the standard PSO (Particle Swarm Optimization) algorithm and the APSO (Adaptive Particle Swarm Optimization) algorithm under four test functions, the results show that the improved algorithm has better optimization accuracy and robustness. Finally, the prediction results of effluent COD (Chemical Oxygen Demand) content of DSAPSO-RBF model were compared with those of RBF model, PSO-RBF model and APSO-RBF model. The MSE (Mean Square Error) and RMSE (Root Mean Square Error) of DSAPSO-RBF model were proposed to decrease by 0.72%-4.23% and 4.39%-15.45% respectively. MAE (Mean Absolute Error) decreased by 3.46%-15.04%, and R
²
(Model Correlation Index) increased by 4.26%-25.14%, which proved that the model built in this paper has better prediction accuracy and can provide a new idea for the measurement of key parameters of sewage treatment.