New Robust Time Series Model for Estimation of Power System Inter-Area Oscillations

The dynamic stability of a power system is generally quantified in terms of the frequency and damping of its dominant electromechanical modes. Time series modeling of synchrophasor data for the estimation of such modes has become commonplace. However, sensitivity of these estimates to the time-series model choice continues to be a challenging topic. In this paper, a novel robust modification to a QR decomposition based adaptive algorithm for order recursive auto-regressive model estimation is introduced. This modular approach helps reduce the sensitivity of the estimates to model mismatch compared to the traditional algorithms by providing simultaneous multiple order solutions whilst lowering the computation costs. The robust modification helps in dealing with anomalies in the data. The results are validated on a model of the US western interconnect (WI). Some results using measured synchrophasor data from the US Western Interconnect are also presented.