Research On Anomaly Detection Method Of Urban And Rural Water Supply Based On Machine Learning

Water supply security concerns the national economy and the people's livelihood.With the rapid development of the social economy,the problem of water pollution has become more serious.In order to make early warning before water pollution occurs and predict water quality anomalies,it is necessary to study the early warning system for water pollution and conduct water quality anomaly judgment and analysis.Water quality anomaly detection technology is an important part of the water quality pollution warning system.Therefore,the study of high-performance water quality anomaly detection technology is of great significance to water quality assurance of the water security department.The main work and research content of this paper are as follows:1.Through the literature review and the current status of water quality supervision,the existing water quality standards and the definition of water quality anomalies were studied,the general process of water quality anomaly detection was summarized,and several commonly used water quality anomaly detection technologies were introduced.2.Study the water quality anomaly detection method based on the time series model,achieve high-precision tracking and prediction of water quality background signal through autoregressive model and vector autoregressive model,calculate the prediction residual,compare with the set threshold,judge whether the water quality is abnormal,and finally The ROC curve was used to evaluate the differences in the effect of water quality anomaly detection between the autoregressive model and the vector autoregressive model.At the same time,by selecting the best tangent line in the upper left corner of the ROC curve to obtain the optimal threshold point,the optimal threshold is calculated to complete the water quality anomaly detection process.The experimental results show that the water quality anomaly detection method based on vector autoregression model has higher detection rate and lower false alarm rate.3.To study a water quality anomaly detection method based on long-term and short-term memory networks,construct a long-term and short-term memory network model,integrate a large number of water quality data,use part of the data for parameter training and regression calculation,and achieve water quality anomaly detection function on this basis The water quality anomaly detection results are given.4.Combine the research on water quality anomaly detection methods,simulate the dynamic changes of water quality parameters based on multiple water quality events,study the diffusion process of water pollution,and analyze the actual water quality anomaly detection performance.