Applying Artificial Intelligence To Distribution Network State Estimation

The paper focuses on researching and building a program to estimate the state of the distribution network with two different methods: the weighted least squares (WLS) and the artificial neural network (ANN) using the Python programming language. Firstly, both methods are used to compute the standard IEEE 118-bus system. The state estimation results of voltage and phase angle are compared with those calculated by Newton-Raphson method in steady-state power flow problem. Furthermore, the paper presents the computed results applied to the real distribution network to demonstrate the effectiveness of state estimation problem for power systems. Datasets are generated by altering loads n times at buses in distribution networks. Each time loads are modified, power flow calculations are performed to record powers, voltages, and phase angles. Finally, the advantages and disadvantages of each method are evaluated to facilitate practical applications.