Impact of Input Data Correlation on Distribution System State Estimation

The State Estimation (SE) 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. Distribution networks present their own features, requiring specific estimation methodologies. The most widespread algorithms for Distribution System State Estimation (DSSE) are based on a weighted least squares (WLS) minimization of the residuals of the measurements and usually compute the state in terms ofeither node voltages or branch currents. The DSSE relies on real measurements collected by the distributed measurement system and on other available information, mainly obtained from historical data, that help obtaining observability. The prior information (the so called pseudo-measurements) derived from expected power profiles of both loads and generators is crucial for the accuracy of the estimation process and has to be fully exploited. A degree of correlation can often be assumed among power consumptions or generations of some particular nodes. The influence of the correlation on the quality of the estimation is investigated in this paper using a branch current DSSE algorithm. The importance of including correlation in the WLS is discussed using both traditional and synchronized measurements. Results obtained on a 95-bus distribution network are presented and analyzed.