Prediction of Aeration Quantity of Biochemical Tank Based on Ensemble Learning Algorithm

Abstract

This paper presents an intelligent prediction method for aeration capacity of biochemical tank for sewage treatment. Firstly, the data collected in the field is processed from the actual sewage treatment plant and the data set is obtained through correlation analysis. Secondly, after optimizing the model parameters, RF, GBDT, LGB and LR models are established respectively to obtain the forecasting capabilities of each model. Furthermore, the fusion of the Stacking model is introduced by using RF, GBDT and LGB as the first layer and LR as the second layer. Experimental results show that the optimized model can better predict the aeration required by the biochemical tank according to the real-time incoming and outbound water quality and quantity data, so as to ensure that the urban sewage treatment plant can save energy and reduce consumption to a certain extent and maintain the sustainable development of carbon neutrality.