Advancing cybersecurity in smart power grid through artificial intelligence

Abstract

Our reliance on electricity and smart power grids for managing electricity distribution has made it a highly important and sensitive matter. Power supply disruptions can have wide-ranging implications and provoke significant social reactions. Therefore, the security of the smart power grid is of utmost importance. The backbone of this network is the communication networks, which connect different parts together and enable them to have a two-way relationship. The advanced features and extensive nature of this network make it more vulnerable to cyber-attacks. In this study, to detect intrusions in the power grid, we first normalized the data and then divided it into two categories: training and testing data. We used a Support Vector Machine (SVM) for training with the training data. After training the algorithm, we input the testing data to evaluate the model. We achieved an accuracy of 99.7%.