Condition Based Maintenance System of Hydroelectric Generating Unit

A three-level system for condition based maintenance (CBM) of hydroelectric generating unit (HGU) is proposed in this paper. The aim is to overcome the projected problems like monitoring items not covering most critical equipments, information inconsistency or redundancy, and insufficient information exchanging and sharing. A systematic audit to HGU is performed, covering from acquired data and parameters to equipment degradation mechanism and experience knowledge. This is the first step to obtain the unit-relevant information and gauge how reliability-improvement funds can be spent most effectively. The architecture of this CBM system is presented from the low field level to the high enterprise one. Key implementation strategies are described in details, involving hiberarchy storage method for data management, condition monitoring based on equipment criticality and fault modes, combined neural network diagnosis and prediction, and hybrid reasoning techniques based maintenance decision making. Eventually an application in a medium hydropower plant is described. The results show that such a system is feasible and practical, and it's promising to be used in other industrial fields.