An adaptive fuzzy logic power system stabilizer

In this dissertation, an adaptive fuzzy logic control algorithm has been developed for a Power System Stabilizer (PSS) to improve dynamic performance of the system. The proposed PSS deals with automating the parameter tuning and structure optimization in order to achieve the desired performance.; This approach combines the advantages of both Fuzzy Logic Control (FLC) and Artificial Neural Network (ANN) and avoids their drawbacks. The parameters of the controller, membership functions and inference rules are adjusted according to gradient decent learning algorithm.; Moreover, the mechanism of how the FLC can be trained in a closed-loop control system is investigated. In the first step, a desired controller is employed to generate the input-output data required for training. The FLC learns to copy the desired controller. This approach needs the existence of the desired controller. To overcome this problem, in the next step, a self-learning approach is utilized to train the FLC directly from the plant output. A genetic algorithm is also used to optimize the structure of FLC, preventing the learning algorithm from the overfitting problem.; Simulation studies and comparison between the proposed adaptive fuzzy PSS and the conventional PSS using a single-machine connected to an infinite bus are conducted. For verification, it has been applied to a multi-machine model of the power system.; A TMS320C30 Digital Signal Processor (DSP) and an ABB PHSC2 Programmable Logic Controller (PLC) were employed to develop a prototype real-time digital control environment and to implement adaptive fuzzy logic PSS.