Design of a mixed H2/H∞ robust PID power system stabilizer with fuzzy adaptation and Simulated Annealing optimization

Abstract

In this paper, a mixed H2/Hinfin control theory and simulated annealing (SA) techniques in conjunction with adaptive neuro-fuzzy inference system (ANFIS) are combined to design two adaptively robust output feedback controllers. The first controller is a mixed H2/Hinfin that is solved using linear matrix inequalities (LMI) technique. The second one, robust PID, which is ideally practical for industry, whose optimum parameters are found using the optimization of mixed H2/Hinfin norms via SA. Canceling pole operation is used to perform the calculations of mixed H2/Hinfin norms. The former is characterized by a similar size as the plant that may be of higher order and thus creates difficulty in implementation in large systems. The latter is shown to be robust and more appealing from an implementation point of view since its size is lower. Two ANFISs are designed. The first ANFIS system (ANFISS), is used to predict the system parameters. The second ANFIS control (ANFISC) , is used to append the corresponding optimized control setting of the PID gains. The operating conditions represent the inputs of each ANFIS defined by the generator active power output, and terminal voltage. Both controllers are used as a power system stabilizer for a single-machine infinite bus system. The proposed controllers show robustness over a wide range of operating conditions and parameters change.