A~2O Process COD Modeling Based On The Improved Differential Evolution Algorithm

As people concern the problems of environment rise ceaselessly, wastewater treatment problem has been concerned by more and more people. A2O technology in wastewater treatment process is one of the main second-level process method, has good ability to remove nitrogen and phosphorus, and has wide application in our country. Wastewater treatment process is a big lag, nonlinear intermittent process. In this process, using artificial neural networks modeling is an effective method.This paper first uses the classic BP neural network and RBF neural network to model and simulate A2O system, to validate the feasibility of modeling A2O system using artificial neural networks. Then, paper uses DE algorithm and DE-ABC algorithm to improve BP network, and model the A2O system. After comparing the results of different model, DE-ABC algorithm has better performance than other algorithm can be found. Therefore, this article draws the conclusion, DE-ABC improved algorithm neural network compared with DE algorithm and BP algorithm is more excellent in A2O wastewater treatment modeling. So, the DE-ABC algorithm provide a reference value with the new method to model the A2O wastewater treatment system.