Energy Consumption Prediction And Sensitivity Analysis Based On Optimized RBF Neural Network

With the rapid construction and development of urban sewage treatment plant,energy conservation and consumption reduction of sewage treatment plant have become the key problem urgently needed to be solved.The sewage treatment process of sewage treatment plant has the characteristics of high complexity,uncertainty and nonlinearity.Therefore,it predicts its energy consumption,feedbacks and adjusts the sewage treatment process parameters to achieve the purpose of energy conservation and consumption reduction,which has become an important goal for the optimization and transformation of the sewage treatment process.In order to solve the problem that the prediction accuracy of traditional energy consumption prediction methods is generally low and is greatly affected by data fluctuations,this paper has carried out sensitivity analysis and reasoning research on energy consumption prediction and energy consumption prediction model based on clustering algorithm-error back propagation neural network(K-means RBF).The content and research results are as follows:(1)Data pretreatment and data analysis based on principal component analysis(PCA).In view of the problems that will affect the prediction accuracy of the model,such as data loss and noise,first of all,the correlation between the original water quality data collected from the sewage treatment plant and water quality factors is analyzed;then,the original water quality data is analyzed by PCA method and the main component is extracted;finally,in order to fully consider each number According to the weight in the overall sample,the score value of the principal component is further calculated.(2)Research on energy consumption prediction model based on RBF neural network.The pre-processing data obtained from PCA analysis is used as the input of the RBF energy consumption prediction model to predict the energy consumption of the sewage treatment plant.The results show that the RBF model has a certain gap between the change trend of the predicted value of the energy consumption of the sewage plant and the actual value.When the RBF model predicts the energy consumption of the sewage treatment plant,the prediction error at each point is relatively unstable,and the error value is large.The average absolute percentage error MAPE is 13.68%,and the fit degree R~2 is 0.3458,the TIC coefficient is 0.0849,so there is a lot of room for improvement for the prediction effect of this model.In addition,after using PCA preprocessing data,the RBF model can not only simplify the establishment process of the model and shorten the calculation time of the model,but also improve the prediction accuracy of the model to a certain extent.(3)K-means optimizes the establishment of the energy consumption prediction model of the RBF neural network.In order to reduce the overfitting of the RBF network and the number of center points,on the basis of the RBF energy consumption prediction model,the K-means clustering algorithm is used to find the RBF clustering center,improve the RBF neural network algorithm,and build the K-means RBF neural network model.Taking the energy consumption of the sewage plant as the research object to verify the effectiveness of the K-means RBF neural network energy consumption prediction model,and compare and analyze it with the RBF neural network energy consumption prediction model.The results show that the average absolute percentage error of the K-means RBF neural network model MAPE is 5.12%,the fit degree R~2 is 0.8905,and the TIC coefficient is 0.0413.Compared with the RBF model,the K-means RBF model has greatly improved the evaluation indicators of the energy consumption prediction of the sewage plant,and the fit has been significantly improved.It can be seen that using the K-means clustering algorithm to optimize the RBF model can significantly improve the prediction accuracy,and the proposed K-means RBF energy consumption prediction model has good feasibility and applicability.(4)Sensitivity analysis based on energy consumption prediction model and energy saving and consumption reduction in sewage plants.Considering the correlation between the energy consumption of the sewage treatment plant and the amount of treated water and the water quality of the inlet water,this paper carries out a sensitivity analysis study based on the energy consumption prediction model of K-means RBF neural network.By analyzing the factors affecting the water quality and energy consumption of the sewage treatment plant,the sensitivity coefficient between the water quality parameters and energy consumption of the sewage treatment plant is calculated,and then the conditions of feedback reasoning are obtained,and the process parameters and equipment of the sewage treatment plant are adjusted to achieve the purpose of energy conservation and consumption reduction.According to the energy consumption prediction data obtained from the K-means RBF neural network energy consumption prediction model,sensitivity analysis is carried out to verify that the model is practical.