Utilizing Machine Learning for Intrusion Detection in Smart Grid Systems

The Smart grid delivers efficient and dependable power through advanced data analysis and communication technology. The reliability of these smart grids holds paramount importance, given that any crucial issue within the system can impact many devices connected through the communication network. Sadly, cyberattacks can put the reliability of smart grids at risk. That's why it's really important always to keep an eye on their cybersecurity to make sure they stay safe and reliable. In our approach, we came up with a way to use machine learning to detect these cyberattacks in smart grid systems. We focused on three particular types of attacks called “false data injection,” “Relay setting change,” and “Remote tripping command injection,” which is a big problem in smart grids. To do this, we took a smart grid attack dataset collected by Mississippi State University and Oak Ridge National Laboratory. Which is a combination of three datasets with 37 power system event scenarios. Then, we used our machine learning approach to figure out when and how these attacks were happening to stop them.