Cyber Security Threats are emerging as one of the major concerns in a smart grid mainly because of the extensive use of heterogeneous communication technologies. As the fields of Detection and Prediction of these attacks have been explored in momentum making a Smart Grid more Resilient to these attacks is still challenging. Cyber Resilience which can be briefly defined as the ability of the grid to recover from a cyber-attack, is of utmost concern since it can be considered the self-defense mechanism of a smart grid. Machine Learning can be implemented in Cyber Resilience Techniques to make it more efficient and find optimal strategies to help the system recover. The Machine Learning Algorithm learns from historical data of the system. Owing to the restrictions imposed by the security and privacy policies of the utility systems, the availability of historical data poses one of the greatest challenges in the implementation of such techniques. So the model has to be developed with significant consideration of privacy preservation and data security of the systems. Therefore, the proposed project aims at designing a Cyber Resilient system using Machine Learning techniques on forth with a Privacy Preservation algorithm that assures the data security of the system. This research has a collaboration with University of Trento, Italy under student mobility scheme for bi-lateral agreement, on the development of a digital twin for smart grid and another collaboration for a project with West Virginia University, USA.
Research Progress
The first phase of the research is based on the detection of cyber attacks in a smart grid environment using machine learning algorithms. Implementation of an Autoencoder- based False Data Injection Attacks detection in smart grids has been successful, with a detection rate of 98%. The algorithm has also been validated through a comparative analysis to prove the superiority of the proposed method.
Collaborations with Universities / Industry Partnerships
Dr. Sarika Khushalani Solanki
Associate Professor
Lane Department of Computer Science and Electrical Engineering
West Virginia University
Title: Development of a Digital Twin of Smart Grid Cyber-Physical System
Dr. Fabrizio Granelli
Full Professor, University of Trento Dept. of Information Engineering and Computer Science
Via Sommarive 9 38123, T, Italy
