Prediction of wastewater treatment system based on deep learning

Abstract

IntroductionIn order to accurately model the IC reactor of the wastewater treatment system and efficiently control and adjust the water treatment process, this paper proposes a method to predict the operation effect of the IC reactor using an artificial neural network model. This paper takes the IC reactor section of a papermaking wastewater treatment plant as the research object, and predicts the COD value of its effluent through the neural network model established. The experimental results show that the simulation prediction value of BP neural network is basically consistent with the change trend of the actual value, and has a certain prediction ability. Among the 20 groups of sample data for simulation prediction, the prediction relative error value of 9 sample data pairs is less than 5%, that is, the prediction error of 45% sample data pairs is within 5%; The relative error value of 15 sample data pairs is less than 10%, that is, 75% of sample data pairs have a prediction error of less than 10%; The maximum relative error is 18.6%. Through the regression analysis of the real value and the predicted value, the correlation coefficient is 0.7431.ConclusionThe BP neural network can capture the non-linear mapping relationship between the selected input factors and the output, and can predict the COD value of the effluent of IC reactor in advance.