Voltage range estimation in low-voltage networks including unmeasured points through voltage distribution

As the penetration of distributed energy resources into the distribution network increases, the importance of voltage monitoring within the distribution network is growing. While increasing the number of measurement points is necessary, it is not economically viable due to the excessive number of nodes and the associated infrastructure costs. Therefore, this paper proposes a method for estimating the voltage range in the low-voltage (LV) network, including unmeasured points, by utilizing the available measurement data from the distribution network. A supervised learning model is developed to estimate the distribution of low-voltage network voltages based on measurements from the medium-voltage (MV) network. Given the input data from the medium voltage network, the model is trained to output the parameters of the voltage distribution and the probabilities of the minimum and maximum voltages in the low-voltage network. Using the output values, the estimated minimum and maximum voltages are calculated from the estimated distribution. Actual network data from Korea is utilized for the training and validation of the model. Simulation results demonstrate that the proposed method provides superior voltage range estimation in low-voltage networks with unmeasured points.