Identification of state parameters for stealthy cyber-events in the power grid using PCA

One of the biggest efforts of securing cyberspace is the ability to secure the critical infrastructure power grid. This meshed network of geographically distributed control systems has recently been interlaced with network capable devices. This article manipulates power system state parameters to identify stealthy cyber-events based on the feature identification method principal component analysis. Principle component analysis is used to examine power system instances contained in an in-control set. To develop a set of observable power system instances the Newton-Rhapson method is used to solve the power flow equations. Cyber-events were created by changing the in-control instances of the power system parameters over a range of ±40%. If a changed instance, simulated malicious instance, was mapped via principle component analysis within close range of the average Hotelling's T
²
value that instance was deemed stealthy. Results indicate certain features have a higher likelihood of remaining in the stealthy regions.