Machine learning based Prediction and Analysis of Incipient Faults in Power Systems
Electric utilities currently rely on a reactive approach to grid management, towards the disturbances and faults that have occurred. This approach can lead to severe grid faults and high costs associated with downtime, energy loss, damaged equipment, and unscheduled maintenance.
IVAs 100-lista 2023
Chalmers tekniska högskola
Ebrahim Balouji, Karl Bäckström, Tomas Mckelvey
In this research, we have developed an artificial intelligence (AI) based solution that predicts incipient faults within weeks of horizon using only voltage and current signals recorded by existing devices.
The Proposed solution enables utilities to proactively identify potential issues and take preventive measures. By implementing this online, grid-wide solution, we have observed a significant reduction in yearly downtime by 60% and a 20-30% decrease in operation and maintenance (O&M) costs. This predictive maintenance system allows utilities to maintain a more stable and reliable grid while minimizing the financial and operational impact of unexpected faults and disturbances.