Based On RBF Fuzzy Neural Network Control Of DO In Sewage Treatment Process

For sewage treatment system in the actual runtime are always facing complicatedentered water, large wave, often overload situation in different working days, It requestsdifferent concentration of the dissolved oxygen, a fuzzy RBF neural network controlleris designed and it is used to control concentration of dissolved oxygen in activatedsludge wastewater treatment system of different parameters. In order to achievestandard, stability, high efficiency, low consumption, intelligent modeling and controlmethod research of the sewage treatment process has important theoretical and realisticsignificance.First briefly introduce the common control variables in Sewage treatment processand automatic control level of sewage treatment plant and the research status ofintelligent control in wastewater treatment process, the key to understand the methodand principle of the sewage treatment process and static and dynamic models, byconnecting with the kinetics of microbial growth and organic matter degradationprocess, the object of study in this paper is established.And then put forward the control scheme, two kinds of fuzzy neural network andRBF neural network are introduced in this paper, according to the four parts of fuzzycontrol, the corresponding four layers of the fuzzy RBF neural network is designed, anduse the K-clustering algorithm and the algorithm of gradient descent method forlearning algorithm of fuzzy RBF neural network controller.Finally, by the MATLAB simulation curve comparison and analysis, it can beconcluded that fuzzy control rules of this method can be improved, and it candynamically adjust membership function on line. Compared with the traditional fuzzycontrol method，this control method has better control effect, the control method can make the concentration of dissolved oxygen accurate reached the requirements ofexpectation, and anti-interference ability and robust performance of the method arestrong.