Fault Diagnosis System Of Hydroelectric Generator Sets And Study Of Data Mining Technology

With the progression of information technology and the rapid development of computer technology and network technology, large amounts of data collected (including statistical data, test data) by fault diagnosis system of hydroelectric generator sets can be easily stored in various databases. However, using the traditional method of data to analyze and deal with these data ,not only waste time but also difficult to effectively find the hidden knowledge in the data.In recent years, cross-penetration of artificial intelligence technology and database technology has made remarkable progress, on the basis of the development of artificial intelligence and database technology , people pay more attention to data mining technology. The basic conditions such as: massive data sets, advanced computers, sophisticated data mining methods are basically in place. Therefore, the application of data mining technology in the fault diagnosis system is feasible.In this article, the hydroelectric generator sets operating data which be stored in hydropower station monitoring system database is made as sample data. On the base of analysis data, using data mining technology to mine the relationship between the operating data and the condition of hydroelectric generator sets.And then, to establish the appropriate model.The model can well explain the operational state of hydroelectric generator sets, while the scientific laws which are extracted by mining model can also be used to predict the future trends of sets.This article uses the SQL Server2005 to build unit operation data warehouse,in order to pre-process the data in the process of data preparation,and to improved data quality; and then analyzed by decision tree algorithm,clustering algorithm, neural networks algorithm and timing?algorithm,use the decision tree algorithm and timing?algorithm to establish hydropower units data mining model; finally using the established data mining models to predict the state parameters. Through on-site data validation represents that the prediction accuracy is high.It shows that linking the data mining technology to fault diagnosis system of hydroelectric generator sets is feasible, and provide a effective solution method to acquire knowledge for fault diagnosis system of hydroelectric generator sets.