Research Of Methods For The Unit Coordination Control System Based On Fuzzy-Neural Network

The design of the unit coordinated control system (UCCS) is of great importance in the fossil fuel power plant. In this thesis two methods based on Fuzzy and Neural Network ( NN ) are researched for the UCCS. No.1. a new scheme based on Fuzzy-rule-tuning and self-adaptive-single- neutrons is presented in the third part of the thesis. In this method two self- adaptive-single-neutrons are adopted to control the boiler ioop and the turbine loop respectively, meanwhile two sets of Fuzzy rules are used to tune their gains K1,K2 on line according to the current system errors. The coupling effects are decreased by the weights of the self-adaptiveO-single-neutrons. No.2 . another new scheme based-on-BP-Neural-Network-tuning and FuzzyP+ID is given in the fourth part of the thesis. In this method two FuzzyP+ID controllers are adopted to control the two ioops respectively, at the same time two BP Neural Networks are used to tune their proportional integral and differential coefficients on line according to the current system errors. The coupling effects are decreased by using a traditional decoupling constant numerical matrix. After theoretical analysis, a great amount of simulation experiments are made, manifesting their satisfactory performances in many aspects with comparison to the traditional methods based on PID, such as better robust property-. better disturbance-resistant and easy designing etc.. Research of the UCCS will be of realistic significance and of great value in the theory This thesis stems from the application research project of 揚rojects of Youth Technology Promotion Funded by the National Electrical Power Company? ?The Application Research of Fuzzy Optimization and Control Theory in the industrial process control...