Inspection is an important means of ensuring the safety of hydropower stations. On the basis of summarizing the manual inspection method of hydropower stations and abiding by the system description and relationships model

an inspection approach considering both states and performance is proposed by taking advantage of transverse correlation and longitudinal recalling of on-line monitoring systems. Within the integrated information platform of the smart hydropower station

the state data are firstly identified and corrected by validity check and redundant switch. The abnormalities of states are then detected by dynamic threshold analysis and trend analysis of similar impact factors. Meanwhile

the performance indices are calculated through macro tests or experts' empirical models

based on which the performance degradation and off-limits are automatically detected. Finally

the alarm information is sent out according to detection results of states and performance. In addition

the relevance of states is effectively guaranteed by data association and data synchronization

thus providing a support for fault diagnosis. The intelligent inspection system based on the approach has been successfully applied in Gezhouba Hydropower Station.