Power system stability improvement with multivariable self-tuning control

A multivariable self-tuning adaptive control scheme is presented. This scheme is of a decentralized nature and is implemented locally for individual generating units. A discrete multivariable autoregressive-moving-average (ARMA) model is developed to represent a generating unit. The recursive-least-squares (RLS) estimation algorithm with variable-forgetting factor and the generalized-minimum-variance control technique are utilized to synthesize the local controllers. A dynamic goal-point-generating model is introduced to provide varying goal point for the local controller which leads the subsystem output to its equilibrium gradually. Extensive simulations are performed on the IEEE ten-machine test system. The results show that the proposed multivariable adaptive control scheme is effective in damping the severe oscillations after large disturbances as well as improving the system dynamics under small oscillations and is better than the conventional power system stabilizer (PSS) method. The controller demonstrates robustness and is compatible with the existing conventional controllers in multimachine systems