Robust Hybrid Linear State Estimator Utilizing SCADA and PMU Measurements

Abstract

This paper intends to improve the accuracy of power system State Estimation (SE) by introducing a hybrid linear robust state estimator. To this end, automatic bad data rejection is accomplished through an M-estimator, i.e. a Schweppe-type estimator with Huber loss function. The method of Iteratively Reweighted Least Squares (IRLS) is used to maximize the likelihood function in the M-estimator. Leverage measurements are also treated by a simple yet effective formulation. To run the algorithm for real-world large-scale grids, cumbersome construction of the Jacobian matrix at each iteration is avoided. In addition, convergence to the local minima faced in the largescale Gauss-Newton algorithm is not a concern as the proposed formulation is linear with no approximation. As observability and redundancy considerations mandate SE to take advantage of traditional SCADA measurements along with available PMU measurements, the linearity of the proposed SE formulation is guaranteed regardless of whether PMU-only, SCADA-only or hybrid SCADA/PMU measurements are utilized. In this regard, covariance matrix for measurements weights is derived for both types of measurements. Thanks to the linear formulation and therefore swiftness of the proposed algorithm, SE could be run for different power systems with a few up to thousands of buses.

Development of electrical power transmission system linear hybrid state estimator based on circuit analysis techniques

“…The equivalent circuit models are coupled with circuit analysis to transform SCADA measurements into linear functions of voltage and current variables. The authors in [27] , [28] also propose a linear formulation for HSE where the state vector comprises complex voltage and a redundant state for the voltage phase angle. A Least Absolute Value (LAV) estimator is used in [27] for automatic rejection of bad data with simulations indicating the accuracy of LAV over WLS in the presence of bad data.…”

Normalized Deleted Residual Test for Identifying Interacting Bad Data in Power System State Estimation

“…where ǫ p 2 is defined by the right-hand side of (15). By (17), (18) we ensure that the pair of bad data (p, q) are included in S i2 .…”

“…Robust methods based on non-quadratic objective functions [9], [12]- [17] have also been suggested as alternatives to the LNRT in PSSE. Compared to the residual-based methods such as HTI and the LNRT, the advantage of the robust methods is that they reject outliers automatically, and therefore do not need a post-processing stage [1].…”

Optimal placement of Phasor Measurement Units in a Distribution Feeder under Reconfiguration

“…A robust linear state estimator from supervisory control and data acquisition (SCADA) along with PMU is discussed in [8]. Another state estimator for voltage security using PMU is discussed in [9].…”