Robust Contextual State Estimation with Limited Measurement Data

TL;DR

This paper introduces a robust, data-driven state estimation method for active distribution networks that effectively utilizes limited real-time and historical measurement data, improving accuracy and robustness in complex power systems.

Contribution

It presents a novel non-parametric estimator and its robust extension to enhance state estimation under measurement scarcity and uncertainty in distribution networks.

Findings

Improves state estimation accuracy with limited measurements.

Enhances robustness against measurement noise and data sparsity.

Demonstrates effectiveness through numerical experiments.

Abstract

The increasing integration of distributed energy resources (DERs) is transforming power systems into complex, decentralized networks, particularly at the distribution level, where active distribution networks (ADNs) introduce new challenges for monitoring and control. Accurate state estimation in these systems is complicated by limited and heterogeneous measurement infrastructures, where real-time data from supervisory control and data acquisition (SCADA) and micro-phasor measurement units ($\mu$-PMUs) is sparse, and slower data from Advanced Metering Infrastructure (AMI) is abundant but delayed. In this context, historical data becomes a key asset, yet its effective use requires robust methods that account for its relevance and variability. This paper proposes a novel data-driven methodology for state estimation under limited real-time observability. We first introduce a non-parametric estimator that leverages nearest neighbors to approximate the conditional expectation of delayed measurement residuals. Building on this, we develop a robustified extension that accounts for ambiguity in the estimated conditional distribution, mitigating the impact of sample sparsity and uncertainty. Numerical experiments demonstrate that the proposed approach brings substantial benefits in accuracy and robustness, particularly in scenarios with scarce and noisy measurements.