Separated Artificial Neural Network Based Distribution System State Estimation

Distribution system state estimation is becoming an increasingly important feature in the modern power grid, but the lack of real measurements makes its implementation particularly difficult. A solution for this problem could be the implementation of Artificial Neural Networks for direct distribution system state estimation. In our work, we created an ANN-based distribution system state estimation which can function based on sparse data, the accuracy of the ANN-based state estimation can be further improved utilizing a separated distribution system state estimation (S-ANN), which estimates the state of the network separately according to measurement type based on the same input data. With this method, the root mean square error and standard deviation of the state estimation can be significantly decreased.