Pseudo-measurements modeling using neural network and Fourier decomposition for distribution state estimation

This work aims to propose a pseudo-measurement modeling method for Distribution State Estimation (DSE) application embedded in a Distribution Management System (DMS). The entire system is already installed on the distribution MV network of Sanremo, in the North of Italy, within the Smartgen research project. The acquisition architecture consists of a SCADA system, which allows the data exchange from meters installed in the MV-LV substations. In order to satisfy the system observability conditions and to perform the State Estimation (SE) algorithm, real-time measures need to be integrate with the pseudo-measures of the non-monitored substations. The paper investigates a load modeling technique, based on Artificial Neural Network (ANN) and Fourier decomposition, that allow the generation of pseudo-measurements starting from the historical database of the monitored substations.