Clustered Surface Water Parameter Prediction Modeling And Diagnosis

Water quality forecasting is an important tool for managing water resources and controlling pollution.It can accurately predict the trend of pollutant concentrations in water over time so that appropriate measures can be taken to ensure compliance with national standards and regulations.Traditional water quality monitoring methods using manual testing equipment has the disadvantages of long detection time and the inability to monitor in real time.Therefore,new monitoring technologies,such as remote sensing and automated monitoring systems,need to be developed to achieve real-time monitoring and prediction of water quality.In this paper,a combined Empirical Mode Decomposition(EMD)-Singular Value Decomposition(SVD)-Long Short Term Memory(LSTM)surface water quality prediction model was established to achieve the prediction of In order to make the proposed water quality prediction model can be used to predict the value of ammonia nitrogen(NH)parameters in surface water in the future.In order to make the proposed water quality prediction model can be applied in practice,this paper designs and develops the surface water quality prediction diagnosis system.The work done in this paper includes the following main points:(1)Establish a water quality prediction model based on LSTM neural network.Specifically,this paper will monitor the water quality data processed and input into the established LSTM neural network prediction model,you can get the prediction results of water quality parameters in the future moment.In order to assess the accuracy of the model prediction,this paper calculated the prediction model RMSE and MSE and other indicators,and will be compared with other models.The experimental results show that the water quality prediction model based on LSTM neural network has higher accuracy and real-time,and can better predict the trend of water quality parameters in the future moment.(2)In order to further improve the accuracy of the LSTM prediction model,reduce the impact of interference information in the original water quality series data on the prediction results,the use of empirical modal decomposition-singular value decomposition of water quality series data processing method will be decomposed into a number of relatively smooth intrinsic modal function and a residual term and then select the singular value reconstruction decomposition of the components obtained,proposed based on EMD-SVD-LSTM algorithm based on the combined water quality parameters prediction model.And the improved whale optimization algorithm is used to optimize the model parameters of the LSTM model and the parameters in the reconstruction process of singular value decomposition.The improved whale optimization algorithm improves the convergence speed and solution accuracy of the algorithm compared with the traditional whale optimization algorithm,and the experiment proves that the improved whale optimization algorithm improves the training speed and prediction accuracy of the LSTM model.(3)Designed and developed the water quality prediction system.Water quality prediction system mainly includes the following modules: water quality data collection module,responsible for the collection of water quality data in the monitoring station;data communication module,responsible for the transmission of the collected data to the water quality prediction model for processing;water quality parameter prediction and diagnosis module,the use of trained water quality prediction model,to achieve the prediction of water quality parameters in the future period of time and diagnosis.