Machine Learning-based Cyber-Physical Anomaly Detection in Wide Area Voltage Control Systems
Wide-area voltage control systems (WAVCS) are widely deployed in power grid to improve the voltage stability in transmission system using Flexible AC Transmission System (FACTS) devices. The WAVCS relies on wide-area measurement and control signals for closed-loop control of FACTS devices to improve the transient voltage stability in power grid in real-time. Since the WAVCS utilizes a cyber-layer communication during its normal operation, they are susceptible to cyber attacks from adversaries which can lead to a voltage collapse if the attacks go undetected and unmitigated. This paper proposes a supervised machine learning (ML)-based anomaly detection algorithm for detecting various stealthy cyber attacks in the context of WAVCS cybersecurity. In particular, a fuzzy logic-based wide-area controller, as proposed by the Bonneville Power Administration (BPA), is implemented on the Kundur’s four machine two-area system that is integrated with a static var compensator (SVC) to improve voltage profile on sensitive buses. Later, different types of data integrity attacks, including pulse and ramp attacks, are considered on the wide-area measurement and control signals to analyze the performance of the proposed anomaly detector. Our experimental evaluation shows a promising performance with a high true-positive rate (more than 99%) and low false-negative rate (less than 1%) while exhibiting a small prediction time.
- Research Organization:
- Hitachi Energy, Raleigh, NC; Iowa State University, Ames, IA
- Sponsoring Organization:
- USDOE Office of Cybersecurity, Energy Security, and Emergency Response (CESER)
- DOE Contract Number:
- OE0000897
- OSTI ID:
- 2223097
- Journal Information:
- 2022 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT), Conference: 2022 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT); New Orleans, LA, USA; 24-28 April 2022
- Country of Publication:
- United States
- Language:
- English
Similar Records
Cyber risk assessment and investment optimization using game theory and ML-based anomaly detection and mitigation for wide-area control in smart grids
Impact Analysis of Data Integrity Attacks on FACTS-based Wide-Area Voltage Control System