Research On Water Quality Simulation And Evaluation Based On HHO-SVM Model

In recent years,water pollution incidents have occurred frequently in China,and reasonable simulation and evaluation of water quality is especially important to protect the water environment and maintain ecological balance.The introduction of water quality simulation and evaluation can effectively assess the pollution status of water bodies and provide a scientific basis for solving the prevention and control of water pollution accidents,so it is important to study new water quality simulation and evaluation methods for water environmental protection and management.In this paper,after analyzing the current status of water quality data from2018-2021 for the lower reaches of the Duilong River,a tributary of the Niulan River,a water quality simulation and evaluation model based on the combination of Harris Hawks Optimization（HHO）and Support Vector Machine（SVM）is proposed according to the water quality characteristics of the region.The research results of this paper are as follows:（1）The spatial and temporal changes of the water quality of the lower reaches of the Duilong River were analyzed based on the water quality monitoring data of the three cross-sections from January 2018 to October 2021;the trends of the water quality factors of the three cross-sections of the lower reaches of the Duilong River were analyzed using the mathematical model based on the seasonal kendall test;the Water Pollution Index（WPI）method based on the lower reaches of the Duilong River was water quality evaluation analysis,found that the overall water quality of each section in 2018-2021 in class III standards;the correlation analysis between the water quality factors,found that the COD and BOD₅showed a significant medium positive correlation.（2）To address the problem of subjectivity in the selection of penalty factor（C）and radial basis kernel function parameters（γ）in support vector machines,this paper proposes to use the Harris Hawk Optimization to find the best for these two parameters and establish a HHO-SVM water quality simulation and evaluation model.The Harris Hawk Optimization is used to obtain the optimal value of the function by simulating the hunting behavior of the Harris Hawk flock,which is characterized by simple operation,few parameters to adjust and excellent global search capability.（3）The established HHO-SVM model was applied to simulate and evaluate the water quality of the lower reaches of the Duilong River.For COD simulation,the MAE of HHO-SVM model is 0.4836,which is lower than that of GA-SVM（0.5159）and SVM（0.6643）.For BOD₅simulation,the MAE of HHO-SVM model is 0.0507,which is lower than that of GA-SVM（0.0825）and SVM（0.1160）.For TP simulation,the MAE of the HHO-SVM model is 0.0116,which is lower than that of GA-SVM（0.0131）and SVM（0.0166）.The overall simulation results show that the accuracy of HHO-SVM model is significantly higher than GA-SVM and SVM,which proves the superiority of Harris Hawk Optimization in selecting the optimal parameters;the water quality evaluation in the lower reaches of the Duilong River can find that the water quality evaluation results of HHO-SVM model are basically consistent with WPI,which indicates that the water quality evaluation results of HHO-SVM are more consistent with the actual situation and can well reflect the The water quality condition of the river.Therefore,it is convinced that the HHO-SVM model can be better applied to water quality simulation and evaluation,and can provide a new and effective method for water quality simulation and evaluation work.