A Robust State Estimation Method for Unsymmetrical Three-Phase Power Systems

In the current literature, it is well established in most cases, that the same weights or weight factors (reciprocal squares of the variances) are used in the conventional weighted least squares (WLS) power system state estimation algorithm. However, this has caused some significant deviation in the solution when compared to the true or actual solutions, due to significant errors. Furthermore, several research studies from the literature had often analyzed the state estimation procedure in single-phase. Nevertheless, this paper proposes a new 3-phase re-weighted nonlinear regression method, that uses the Welsch weight function as the updating weight factor incorporated into the weighted least squares normal equation formulation, to minimize the significant deviations caused by the measurement errors. The proposed method is in 3-phase and is investigated on a 13-node IEEE test feeder in MATLAB, with 0.01% error added to the measurements, and compared using the same conditions with the 3-phase conventional state estimation (CSE) and a 3-phase load flow actual standard solution. First, it is shown that the proposed method outperforms the 3-phase conventional state estimation method in terms of feasibility when the error feasibility performance index is computed. Finally, the differences in term of estimation accuracy, for the proposed method is highlighted using two performance accuracy indexes (Voltage and Angle) for validation and are shown on the bar chart.