Rapid detection model of water pollution based on back propagation neural network

By carrying out the present situation analysis and risk prediction of high mineral groundwater pollution in water source protection areas, it can effectively prevent and control the deterioration of groundwater environment and protect the safety of groundwater resources in water source protection areas. It has certain guiding significance for the development and utilization of local water resources. In this paper, a fast detection model of water pollution based on back propagation neural network is proposed. Combined with the hydrogeological conditions of a water source protection area, a three-dimensional flow system for fast detection of water pollution is established, and the dynamic parameters of fast detection of water pollution are analyzed according to the groundwater flow model. The rationality of the fast detection model and parameter selection are verified by comparing the actual observation values with simulation calculations. In this paper, the back propagation neural network identification method is used to realize the classification and identification of water pollution rapid detection. under certain prediction conditions, Visual Modflow software is used to predict the migration of high-mineral groundwater pollutants in groundwater source protection area within 20 years, which provides relevant basis for the formulation of prevention and control measures of high-mineral groundwater pollution in this water source protection area. Combined with the characteristics of water source protection areas, ammonia nitrogen and chemical oxygen demand (COD) were selected as simulation factors to realize rapid detection of water pollution. The test results show that the accuracy of rapid detection of water pollution by this method is high, and the ability of prediction and quantitative detection of water pollution is improved.