Deep learning Algorithm Based Improved Voltage Switching Model for Efficient Power Stability in Power Grids

Assistant Professor, Department of Electrical, Kalinga University, Naya Raipur, Chhattisgarh, India .

^* Corresponding Author:ku.shaileshmadhavraodeshmukh@kalingauniversity.ac.in

Abstract

The problem of power stability in electrical systems has been well analyzed. There exist number of approaches to handle the issue which consider the factors like residual energy in various power grids of the system. However, they suffer to achieve higher performance in power stability. To handle this issue, an efficient deep learning based voltage switching model (DLVSM) is presented in this article. The proposed model adapts deep neural network towards the selection of electrical circuit from the list of serially connected grids. As the grids have their own residual energy and varying voltage production, the proposed model trains the network with different features like average voltage generation, average output voltage and residual energy. Using all these factors, the neurons are designed to measure the Optimized Power Stability Factor (OPSF) for various patterns of grid circuit. The network is designed with number of intermediate layers where each layer has set of neurons which estimates the Power stability factor (PSF) for specific grid unit. The output layer neurons estimates OPSF value for various sequences of grids. Based on the value of OPSF, the proposed model identifies a specific sequence and performs voltage switching to maintain the power stability in the electrical systems. The proposed model improves the performance of power stability and reduces voltage loss.