Physics-Informed UHF Sensing for Partial Discharge Condition Monitoring in High-Voltage Insulation

Abstract:  

Modern electric power systems are the backbone of almost all other infrastructure. Hospitals, water, industry, and transportation constitute essential services supported by the electric grid, to name a few. The reliability of these critical systems depends on many capital assets, such as transformers, switchgear, cables, and rotating machines. These capital assets rarely fail without warning due to their high reliability and robustness. Nonetheless, their insulation gradually degrades over time. Partial discharge (PD), a localized electrical discharge that produces nanosecond current pulses that radiate electromagnetic energy in the ultra-high frequency (UHF) range, is responsible for many recorded failures in such devices. UHF sensing is among the most attractive PD detection and diagnostic methods because it is highly immune to low-frequency noise, can be non-intrusive, and is widely used in online monitoring. The biggest challenge associated with UHF is that there is no fixed relationship between PD magnitude and recorded UHF signal strength. When such signals propagate within high-voltage devices, the path they follow is dispersive, resonant, and geometry-dependent. This is why UHF sensors cannot be calibrated in terms of apparent charge the way conventional methods are. In this seminar, I present a research program that treats that “path” as the object of study. This work stems from my full-wave electromagnetic modeling of different gas-insulated switchgear systems, a patented UHF sensor for PD detection, and over four years of experience in developing RF systems (front ends) and control for over-moded cavities. The end goal is adaptive RF/UHF sensing, guided by physics-informed inference, that grows into closed-loop smart RF platforms and sensors for monitoring high-voltage devices against PD activity.

Bio:

Ahmad Darwish is an assistant professor of electrical engineering at the University of Alabama. Prior to that, he served as a senior research associate at Purdue University and as CTO of Lyowave Inc., which he co-founded. His research interests focus on RF systems architecture development, wideband antennas and sensing front ends, RF PCB design and packaging, and high-power microwave chain integration with protection and measurement validation. His application areas include smart RF sensing for condition monitoring, particularly partial discharge diagnostics, and microwave-heating platform development for industrial process control, including microwave-assisted freeze-drying. Darwish has authored over 25 peer-reviewed journal articles and conference papers, has a book in press, and is an inventor on one issued patent and three provisional patent applications. He received his Ph.D. in electrical engineering from Texas A&M University (TAMU), College Station, Texas, in 2021, with a focus on UHF sensing for partial discharge diagnostics.