Deep Neural Network Driven Electric Spring for Voltage Regulation

In this paper, a Deep Neural Network (DNN) is proposed to perform robust voltage regulation using Electric Spring (ES). This work focuses on both the design and implementational details of a Neural Network that has been used to drive ES under severe loading conditions of the power distribution system. ES has been previously used to perform voltage regulation; however, the robustness added due to the well-trained DNN is the essence of this work. The data set for training DNN parameters have been obtained using offline dry runs of a typical distribution network. Later, the trained model is operated under unseen test cases. It has been shown that DNN based ES outperforms the previous implementations of ES due to a smaller number of sensors and fewer dependencies on-grid variables.