Robust State Estimation Method Based on Mahalanobis Distance Under Non-Gauss Noise

Abstract

Under non-Gauss noise condition, the performance of traditional state estimation methods based on Gauss measurement noise will be greatly reduced. In order to solve this problem, a robust state estimation method based on Mahalanobis distance under non-Gauss noise is proposed in this paper. First of all, based on the Mahalanobis distance, the calculation method of optimal buffer length for PMU measurements is used, which can unify the SCADA measurements with PMU measurements in the same snapshot. Then, Based on the two-stage processing method, in the first stage, the SCADA measurements are used for filtering by using maximum likelihood estimator to obtain the estimated values, and then the estimated values in the first-stage are combined with PMU measurements as the second-stage measurements for filtering, and finally the final results are obtained. Based on the IEEE-39 buses system and IEEE-118 buses system, under Gaussian noise and non-Gaussian noise, the AEE results of proposed method are very small, and which are all within 10−3, numerical tests under different simulation conditions verified the robustness and effectiveness.

Keywords

• Non-Gauss noise
• optimal buffer length
• two-stage
• maximum likelihood estimator
• state estimation