A Fuzzy Neural Network Based On Differential Algorithm And Its Application In Wastewater Treatment Plants

Activated sludge process is a common used method in the modern urban wastewater treatment plants. Because of the dramatic change of water quality and the complexity of the growth of microorganism, making the process has many variables, strong coupling, nonlinear and hysteresis characteristics, which are the main reason leading to the modeling and control level relatively backward. In order to improve wastewater treatment of effluent water quality, ensuring treatment system was stable and efficient operation, do intelligent modeling and intelligent control method research has important theoretical and practical significance, and it provides effective reference to other strong coupling, nonlinear complex system modeling and control method. Fuzzy neural network combines fuzzy reasoning knowledge ability of expression with neural network self-learning ability, it is widely used in the adaptive control, nonlinear system identification, pattern recognition and other modern industrial fields. Fuzzy neural network design includes network structure identification and parameter identification. BP algorithm is often used in the parameter optimization, but BP algorithm is easy to fall into local minimum value. In the view of the deficiency of the existing methods, this paper puts forward DEBP algorithm which can optimize fuzzy neural network structure parameters better.Based on analysis of the existing research results in the foundation, the fuzzy neural network which is optimized by the DEBP algorithm is applied in the urban wastewater treatment. The main studies are listed as follows:1. Summarizes the fuzzy neural network, puts forward the DEBP algorithm and analyzes the operation process of the algorithm. Verified the validity of DEBP by standard test functions. Through the second order system with the time delay model simulation, and the results show that the fuzzy neural network is effective.2. Analysis the activated sludge model mechanism deeply, using the fuzzy neural network to establish wastewater treatment effluent BOD water quality prediction model, the simulation results show that the validity of the model, and through the data analysis, it is proved that this model has good performances.3. Designs a fuzzy neural network controller, using this controller to control the DO in the wastewater treatment. The simulation shows the effectiveness of the controller and its good control performances.This paper has certain reference significance to the intelligent modeling and intelligent control methods in our country wastewater treatment.