Damping Inter-Area Oscillations With Large-Scale PV Plant by Modified Multiple-Model Adaptive Control Strategy

Studies suggest that inter-area low-frequency oscillations can be damped with supplementary damping control of large-scale photovoltaic (PV) plants. Multiple-model adaptive control (MMAC) is applied as the damping strategy to deal with uncertain variations of the postdisturbance operating conditions. To reduce the scales of the model bank and the damper bank, K-means clustering algorithm is used taking inter-area modes as features for operating condition clustering, and a common damper is designed for each cluster. Based on the deviation between the output dynamic responses of the actual system and models, Bayesian approach is employed to calculate the probability of each model representing the actual system in real time, followed by updating weights of each damper's output. The weighted average of the individual damper's outputs is attached to the reactive power control of the PV plant. A new continuous form of probability calculation with an inertial link is proposed to smooth weight fluctuation caused by discrete calculation in conventional MMAC and ease stress of power regulation of the PV plant. The nonlinear simulation results demonstrate that the proposed control strategy is able to sufficiently damp the inter-area modes of interest in unexpected operating conditions without any prior knowledge about the postdisturbance state.