Adaptive PSS controller based on operating condition dependent ARX model

Power system stabilizers (PSSs) have been proven to be very effective and economic devices in enhancing stability of power systems. However, the conventional PSSs (CPSSs) do not provide satisfactory performance over a wide range of operating conditions because they are fixed-parameter linear controllers. To solve that problem, adaptive PSSs (APSSs) have attracted much attention in recent years.;Two types of APSSs are specially designed for the OC-dependent ARX model in this dissertation. The first type of APSS applies the pole-shift (PS) algorithm, an adaptive control strategy based on linear dynamic models, while the second type of APSS applies a new adaptive control scheme based on nonlinear dynamic models and for more general application, named as recurrent adaptive control (RAC). An advantage of the RAC control is that it can remove system oscillations introduced by the model reference adaptive control (MRAC) scheme.;Effectiveness of both APSSs is demonstrated by simulation studies on the single-machine infinite bus system, simulation studies on the multi-machine system and experimental studies on a scaled physical power system located at the Power System Research Laboratory at the University of Calgary.;Most APSSs are model-based. The widely used model in APSSs is the auto-regressive with exogenous inputs (ARX) model. In this dissertation, an operating-condition-dependent (OC-dependent) ARX model is developed and realized by local model networks (LMN). The proposed model has the capability of fast learning and can work for various operating conditions without updating its parameters. Effectiveness of the model is verified by simulation studies.