An Unscented Particle Filtering Approach to Decentralized Dynamic State Estimation for DFIG Wind Turbines in Multi-Area Power Systems

This paper introduces a novel application of a stochastic filtering algorithm-unscented particle filter (UPF)-to estimate the inaccessible state variables of doubly fed induction generator (DFIG) connected to a multi-area power system with local phasor measurement units (PMUs). This dynamic estimation implementation bears more advanced features than the particle filter (PF) method since it can not only track the dynamic states more accurately and smoothly when the power system experiences sudden disturbances, but also manage to resolve the particle degeneration problem that exists in the PF algorithm. Moreover, the proposed UPF-based dynamic state estimation method is achieved in a decentralized manner and only uses local PMU measurements of voltage and current. Through a comparison study where popular stochastic filtering methods, unscented Kalman filter (UKF) and PF, are employed to achieve the same estimation purpose, this paper shows the superiority of the UPF algorithm particularly designed for state estimation over the other two existing algorithms in terms of accuracy and error tolerance.