Real-time State Estimation in a System Partially Observed by PMUs: A Coherency Data Mining Based Approach

This paper presents a real-time state estimator (SE) in a system that is partially observed by phasor measurement units (PMUs). The algorithm involves two stages. The first stage runs at SCADA speed and is based on a static state estimator (SSE) with measurements from remote terminal units (RTUs) and PMUs. The second stage runs at PMU speed and is based on a linear state estimator (LSE) that uses only phasor measurements. In order to compensate the lack of measurements in stage two and ensure the observability of the entire system a novel methodology that generates voltage dynamic pseudo-measurements is proposed. This approach is based on the concept of coherency of an electric power system (EPS) and defines, firstly, the PMU location that allows observing all the coherent areas and, secondly, a classifier that forecasts coherency in real time with the aim of calculating dynamic pseudo-measurements. The state estimation algorithm together with the proposed methodology have been evaluated on the New England system under several operating scenarios. Results show the ability of the methodology for generating dynamic pseudo-measurement to operate with accuracy in real time conditions. As a result, the SE is able to accurately estimate the state in real time in the presence of both slow and fast transient phenomena.