Robust state estimation based on orthogonal methods and maximum correntropy criterion

Abstract

This paper presents an orthogonal implementation for power system state estimators based on the Maximum Correntropy Criterion (MCC). The proposed approach leads to a numerically robust estimator which exhibits self-healing properties, in the sense that gross errors in analog measurements are automatically rejected. As a consequence, robust estimates are produced without the need of running the state estimator again after bad data identification and removal. Numerical robustness is achieved by means of a specialized orthogonal algorithm based on fast Givens Rotations, which is able to handle the dynamic measurement weighting mechanism implied by the Parzen window concept associated to MCC. Results for a 3-bus test system are presented to properly illustrate the Correntropy principles, and several case studies conducted on the IEEE 30-bus and 57-bus benchmark systems are used to validate the proposed methodology. © 2017 IEEE.