Fuzzy scheduling control of a power plant

Current control strategies for fossil fuel power units are multiloop schemes based on conventional PID control algorithms, that have demonstrated their effectiveness to regulate power output under random load disturbances. Normally, the parameters of the controller are tuned at a predefined operating point (i.e., base load), and left fixed thereafter. However, wide-range load-tracking operation challenges this approach. The performance of the unit may decrease due to the nonlinear dynamics of the process, that change with the operating point. Gain scheduling and multimode control have been pointed out as two viable control alternatives to deal with this problem in industrial processes. This paper introduces a fuzzy PID controller that merges both techniques in a single mechanism that preserves their advantages. The amalgamation is carried out by a multi-input-single-output Sugeno-type fuzzy inference system. This controller is used to extend the application range of the control system in a power plant without changing the essential philosophy of the already existing control structure, achieving smooth wide-range load tracking operation.