Automated Vibration Analysis in Loaded Condition of Hydro Turbine Generator Sets

Vibration analysis can reveal rich health condition information of hydro turbine generator sets. And it is still performed by the experts manually through tests. Besides, the accuracy of the analysis results is affected due to lack of experts with rich experience and domain knowledge. And manual tests increase downtime of the units. To address these issues, an automated analysis approach in the loaded condition based on the operation condition association rules (OCARs) is presented. It simulates the analysis process of the experts and makes the process automated, like automated operation condition classification, automated operation performance indicators calculation, automated health condition assessment and automated fault analysis. And it has advantages in automated vibration performance degradation detection, fault analysis and operation laws discovery in normal operation of the units instead of tests. The approach has been applied successfully in the vibration analysis of the 22 units in Gezhouba Hydropower Plant. The application shows that it has good prospects.