The complexity of the equipments in power plant has increased the difficulty for operation management and fault diagonosis. With the development of the automation and imformation technology, more and more real-time data is sent to databases by DAS and large amounts of data are accumulated. Abundant knowledge exists in historical data and it hard to find and summarize in a traditional way due to mass and strong coupling of the eletricial history data. Data mining is a nontrival process of identifying implicit, valid, novel, potentially useful and ultimately understandable patterns in data. This paper introduces the decision tree, rough set and association rules into the fault analysis in power plant. The data mining technology is used to classify and discover the fault features in power plant equipments. The experiment results show that it is effective and the reliability of diagonosis can be improved greatly. |