An automatic tuning method of a fuzzy logic controller for nuclear reactors

The design and evaluation by simulation of an automatically tuned fuzzy logic controller is presented. Typically, fuzzy logic controllers are designed based on an expert's knowledge of the process. However, this approach has its limitations in the fact that the controller is hard to optimize or tune to get the desired control action. A method to automate the tuning process using a simplified Kalman filter approach is presented for the fuzzy logic controller to track a suitable reference trajectory. Here, for purposes of illustration an optimal controller's response is used as a reference trajectory to determine automatically the rules for the fuzzy logic controller. To demonstrate the robustness of this design approach, a nonlinear six-delayed neutron group plant is controlled using a fuzzy logic controller that utilizes estimated reactor temperatures from a one-delayed neutron group observer. The fuzzy logic controller displayed good stability and performance robustness characteristics for a wide range of operation.