Distribution System State Estimation Based on Pseudo Measurement Modeling Using Convolutional Neural Network

In order to achieve the system observability, distribution system call for pseudo measurements and virtual measurements. This work presents a convolutional neural network (CNN) based method for pseudo measurement modeling. The input of the CNN is the historical load data and the outputs are pseudo measurements. The virtual measurements are taken as the linear constraints in the distribution system state estimation (DSSE) through improved equivalent current measurement transformation algorithm. Simulations are carried out to show the accuracy and numerical stability of the proposed method in DSSE.