Application of Deep Learning Algorithm and Platforms in Hydropower Units

The conventional online monitoring system for hydroelectric turbines usually uses high-speed A/D data acquisition cards to collect multi-channel data in real-time, transmit it to the central control unit, and employ analysis tools installed in that unit for signal processing. Advancements in computer technology have greatly enhanced CPU computing and storage capacities, allowing the data analysis software to include extra signal processing functions. For the online monitoring system of hydropower units, the design and development of the entire software monitoring and fault diagnosis system will be raised to an unprecedented level, affecting the efficiency and actual contribution of the hydropower plant monitoring and fault diagnosis system. This investigation not only enhances and refines the signal processing capabilities but also develops an intelligent diagnosis module that introduces seven functional modules based on the LabVIEW platform. These modules have been validated using both simulated and real data from Shenzhen Pumped-storage Power Station.