Power system static state estimation using a least winsorized square robust estimator

• A new robust least winsorized square state estimator is introduced by applying it to the power system static state estimation problem.
• The power system state estimation problem is solved as an optimisation problem using the jDE-self adaptive differential evolution algorithm.
• The proposed method provides accurate state estimates even in the presence of bad data measurements and during ill-conditioned system.
• The proposed method improves the accuracy, robustness, and reliability of the state estimates under different scenarios, namely normal condition, bad data condition, different operating points, different measurement condition, under ill-conditioned system, and presence of false data injection.

State estimation is the heart of the energy management system, primarily used for control and monitoring of electrical power systems. Commonly used conventional weighted least square estimator is not reliable, particularly in the presence of bad data. Subsequently, bad data processing techniques have been developed to detect, identify, and eliminate the bad data present in the measurement set. In the present paper, a robust least winsorized square (LWS) estimator is proposed for the power system static state estimation. One of the main advantages of this estimator is that it has an inbuilt outlier rejection property and is less sensitive to bad data (outliers) measurements. The power system state estimation problem has been solved as an optimisation problem using the jDE-self adaptive differential evolution algorithm. The proposed approach has been implemented and tested on standard IEEE test systems. The effectiveness of the LWS estimator has been demonstrated under different operating conditions, namely, normal operating condition, bad data condition, different operating points, different measurement condition, under the ill-conditioned system, and the presence of false data injection. The performance of the proposed approach has been compared with the conventional and the evolutionary based state estimation techniques. Based on the various performance indices, the results thus obtained show that the proposed technique has better accuracy, robustness, and reliability compared to the results obtained using conventional and evolutionary based state estimation techniques.