Research On Prediction Modeling Of Effluent Quality Of Sewage Treatment Plant Based On Long Short-term Memory Neural Network

In the sewage treatment process,water quality prediction is a very important step,that is,the effluent water quality parameter value is obtained through the relevant detection method,and the change of the sewage water quality is judged,so that the staff can timely judge whether the sewage effluent quality conforms to the water quality change.According to the standards set by the state,it is decided whether to discharge sewage in the next step.At present,water quality predictions are mostly carried out manually,using detection equipment to detect water quality parameters to further determine the sewage water quality.However,this method has certain shortcomings,such as time-consuming detection and cannot meet real-time requirements;some effluent water quality parameters cannot be directly measured by equipment.Therefore,how to achieve rapid and accurate measurement of important water quality parameters in the sewage treatment process,so as to facilitate the staff to grasp the change of sewage water quality,and realize the automatic monitoring of sewage water quality,is still one of the difficult problems to be solved.Some scholars have applied soft-sensing technology to the prediction of sewage water quality,by establishing a soft-sensing model to detect the effluent water quality parameters and have achieved good results.Among many soft sensor models,artificial neural network has good fitting performance,so it can be widely used in water quality prediction modeling.This paper establishes a sewage water quality prediction model based on the Long Short-Term Memory(LSTM)neural network,and realizes the prediction of chemical oxygen demand(COD),ammonia nitrogen(NH4N),total phosphorus(Total Phosphorus,TP)prediction of three sewage water quality parameters,so as to understand water quality changes.In order to make the proposed water quality prediction model be practically applied in sewage treatment plants and other places,this paper designs and develops a sewage water quality prediction system.The work done in this article mainly includes the following points:(1)Establish water quality prediction model based on LSTM neural network.As the wastewater data to be treated has time series characteristics,so in the prediction of water quality,the establishment of a prediction model that can handle time series can make the prediction more accurate and significant.Therefore,this paper establishes a water quality prediction model based on LSTM neural network.By inputting the wastewater data into the established LSTM neural network prediction model,the prediction results of the effluent water quality parameters can be obtained.In addition,this paper also calculates the RMSE(root mean square error)and MSE(mean absolute error)of the prediction model and compares them with other models to show the accuracy and real-time performance of this model.(2)Establishment of DBSCAN-LSTM water quality prediction model.The convergence speed of the neural network model is slow,and it is easy to fall into local minimization during the training process of the model,and the phenomenon of local optimization occurs,which cannot get a high accuracy in the global range,making the final prediction results not accurate enough.Therefore,this paper further proposes a joint DBSCAN-LSTM water quality prediction model based on the combination of LSTM neural network and DBSCAN algorithm.This joint model is mainly based on the LSTM neural network water quality prediction model,and the DBSCAN algorithm is introduced to reduce the influence of local minimization on the prediction results by clustering the wastewater data.Experiments show that,compared with a single model,the RMSE and MSE error values of the DBSCAN-LSTM combined model are smaller,and the prediction values are close to the true values,indicating that the prediction accuracy of the water quality prediction model is more accurate and the prediction effect is better.(3)In order to make the proposed water quality prediction model be practically applied in sewage treatment plants and other places,this paper also designed and developed the water quality prediction system.Taking into account the convenience of users,the system is divided into different modules,and the functions of each module are designed separately.This system mainly includes data import,water quality parameter prediction,data storage and other modules.This system is embedded with a water quality prediction model.After users log in to the system,they can get the results of water quality prediction through simple operations.The results of water quality prediction will be presented to users in the form of graphics,which is convenient for users to view and understand.