Research Projects

The global adoption of green renewable energy, primarily driven by Inverter-Based Resources (IBRs) like solar PV and wind, is reshaping power networks but also introducing serious vulnerabilities. Unlike conventional synchronous generators, IBRs inject low-magnitude, non-synchronous fault currents, impairing the reliability of traditional protection relays and increasing the risk of undetected or mismanaged transmission line faults. These challenges are compounded by the growing threat of cyberattacks, particularly False Data Injection (FDI) attacks, which manipulate relay data and can cause protective systems to malfunction, potentially triggering widespread outages. Real-world events such as the 31 interstate faults in India between 2022–2023, the 2024 Victoria storm that affected 500,000 customers, and high-profile cyber incidents in Ukraine (2015) and Israel (2020) highlight the urgent need for more cyber-resilient protection solutions. Existing technologies—including traveling-wave relays, rule-based and fuzzy logic methods, adaptive impedance relays, and signal processing techniques like Discrete Wavelet Transform (DWT) and Hilbert Huang Transform (HHT)—face limitations under high IBR penetration due to weak or distorted fault signatures, bidirectional flows, non-sinusoidal waveforms, and sensitivity to parameter selection. Synchrophasor-based methods offer some promise but struggle with latency, cost, and communication dependencies. Additionally, current cybersecurity defenses often fail to keep pace with evolving attack strategies, as many remain rule-based and inflexible. Recognizing these gaps, the proposed project aims to develop a unified, efficient data-driven AI, cyber-resilient protection framework tailored to the unique characteristics of IBR-dominated networks. This solution integrates advanced machine learning and signal-based feature extraction to enable accurate fault detection, classification, localization, and direction estimation, even under weak or distorted conditions, while embedding anomaly detection mechanisms to counter cyber-physical threats. The novelty of this approach lies in combining analytical precision with cybersecurity resilience in a single adaptive system that dynamically responds to both physical and cyber challenges. The work plan involves six tasks: preliminary AI techniques implementation, data set collection, choosing relevant IEEE standard benchmark test systems, identifying relevant physics-aware AI or ML models, training and testing of relevant identified models for accurate line fault detection, classification, direction identification and location estimation in power networks with IBRs, and cyber-resilient AI methods to append above methodologies with robust anomaly detection against data, replay, and data lag attacks. The proposed methodologies will be validated through rigorous hardware-in-the-loop (HIL) testing and prototype development, targeting technology readiness level 6 (TRL 6), with future scaling supported by industry collaboration. The anticipated impact includes improved grid reliability, reduced outages, enhanced cyber defense, and strengthened protection frameworks essential for supporting India’s 500 GW renewable target by 2030 and global ambitions for clean, stable, and secure power systems. By addressing both isolated and intertwined challenges of fault protection and cybersecurity, this work advances the state of the art beyond fragmented solutions and paves the way for next-generation protection systems in renewable-rich grids.