Enhancing Cyber Security in Power Sector using Machine Learning

Abstract

Nowadays, our lives have become very much dependent on the power systems, whether it is in home or in offices or anywhere. Any failure in the power systems can bring our lives to a halt. To ensure no power fault, a continuous and remote monitoring, control and automation are needed. The implementation of constraints increases the efficiency of the power systems. But, to put monitoring, control and automation into practice we need network, and with this come the threat of cyber-attacks. With more open standard-based communication network, the automated power systems have become the target of the cyber-attacks. By exploiting the cyber components in networks, critical cyber components can be manipulated. Intruders can tamper the communication links by injecting false or modified data. To come up with security measures against these attacks, vulnerabilities of the power systems are being assessed to analyze the impacts of the cyber-attacks. Several techniques have been implemented so far to make the power systems less prone to threats. In this paper, technology like Machine Learning is used as anomaly discriminator and to provide security to the power system against the cyber threats.