A Deep Convolutional Neural Network Technique for Power Losses Estimation in Distribution Grid

Abstract

Estimating and analyzing the abrupt changes in the power grid is a difficult task. These changes are nonlinear, uncertain and of higher orders which lead to certain losses in power grid. Stability and reliability of power grid network is affected by excessive power losses. A fast, accurate and stable algorithm is needed to estimate the actual condition of power grid in order to minimize the losses. This paper proposed a real-time technique which estimates the actual power losses in distribution system. Deep convolutional neural network power loss estimation (DCNN-PLE) method is proposed for calculating distribution network losses on the cost of less computation and complexity. The proposed method achieves high accuracy and less execution time as compared to other traditional methods. Root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) are used as evaluation metrics. The experimental and simulation results show that the proposed method achieves better accuracy for 69bus distribution system in estimating the network losses.