Effects of Measurements and Pseudomeasurements Correlation in Distribution System State Estimation

Abstract

Distribution system state estimation (DSSE) is one of the key elements of the monitoring activity of an active distribution network, and is the basis for every control and management application. The DSSE relies on real measurements collected by the distributed measurement system and on other available information, mainly obtained from historical data that help in obtaining observability. This prior information is necessary to derive the so called pseudomeasurements. Accurate input data are fundamental for an accurate estimation, as well as knowledge on possible correlation in the measured and pseudomeasured data. A degree of correlation can exist in the measured data, due to measurement devices, and among power consumptions or generations of some particular nodes. This paper presents an exhaustive analysis on the influence of correlations on the quality of the estimation. The importance of including correlation in the weighted least square estimation approach is discussed using both traditional and synchronized measurements. Results obtained on a 95-bus distribution network are presented and discussed

Performance Assessment of Linear State Estimators Using Synchrophasor Measurements

“…4) As demonstrated in [43], neglecting PMU correlations (both in amplitude and phase) in the estimator model does not lead to a significant decrease of the SE quality. 5) Voltage and current phasors are measured separately, per phase, i.e., we do not assume to use any 3-ph multifunction meter [44].…”

“…At this point, it is important to mention that, in principle, as analytically described in [43] and [44], R k can be a full matrix, since it is generally inappropriate to consider the measurement errors independent unless this hypothesis is properly justified. In our case, based on the observations that are mentioned in the following, we can neglect the correlations between different PMU measurements.…”

“…1) Measurement values given by different devices can be reasonably considered independent [43]. 2) Measurements come only from PMUs, i.e., we do not use any conventional measurement (e.g., power flows and power injections) or pseudo-measurements.…”

“…3) The measurements of voltage and current amplitudes in the same PMU can be usually considered uncorrelated [43]. 4) As demonstrated in [43], neglecting PMU correlations (both in amplitude and phase) in the estimator model does not lead to a significant decrease of the SE quality.…”

Undetectable Timing-Attack on Linear State-Estimation by Using Rank-1 Approximation

“…• measurement values obtained by different devices can be reasonably considered independent (e.g., [28]); • we only use PMUs, i.e., no typical measurements are used (e.g., power flows, power injections) or historical data; • voltage and current amplitude measurements taken by the same PMU can usually be considered uncorrelated [28]; • in [28] it is confirmed that neglecting PMU correlations (both in amplitude and phase) in the estimator model, does not lead to a significant decrease of the SE quality; • based on the nomenclature and definitions given in [29], we use only independent Gaussian-distributed measured data and not processed dependent measurements; • A unique sensor per measured quantity (i.e., voltage / current) is used and the cross-talk interference is negligible. In view of the above, R is diagonal and can be expressed as…”