Comparative Performance of Neuro-Fuzzy PSS Architectures with Adaptive Input Link Weights and Nonlinear Functions

Based on a Neuro-Fuzzy Controller (NFC) architecture, two approaches are presented for the design of a Power System Stabilizer (PSS) with adaptive input scaling. In the first approach, input link weights (ILWs) are introduced and the NFC is made adaptive by the online modification of the ILWs and the consequent parameters (CPs) through the gradient descent method. In the second approach, nonlinear functions (NLFs) are used in the first layer of the NFC and both NLFs and CPs are modified online by using a hybrid adaptation process. Comparison studies on a one machine-infinite bus system and a multimachine power system show the ability of the proposed PSSs to improve the system dynamic performance.