Distributed Acoustic-Vibration System for Industrial Predictive Maintenance

Pumped-storage hydropower plants play a critical role in the energy transition by providing flexible grid storage; yet, their turbines are subject to significant mechanical stress due to frequent switching between generation and pumping modes. Unplanned equipment failures in such facilities carry substantial economic consequences, motivating the need for automated fault detection and localization systems. While acoustic anomaly detection has been established as a viable approach for industrial condition monitoring, its application to hydropower environments is complicated by strong background noise and variable operating conditions. Furthermore, existing systems address only whether an anomaly has occurred without identifying where within the machine room it originates.