The Study Of Fans Fault Diagnosis On Power Plant Based On Combing Of Rough Sets Theory And BP Neural Network

This paper obtains the subsidization of Shanxi Province electric power company science and technology project, the primary mission is carring on the diagnosis to the fans in Shentou second thermal power plant, and achieves the the goal.Rough Sets Theory was introduced by scientists Pawlawk professor from Poland in the early 1980s, 20th century. The main idea of Rough Sets Theory , under the premise of maintaining the same classification capability, is derive the classification rules of concepts through knowledge reduction . It defines the knowledge as the partition capabilities to the domain from a new perspective, and introduces the mathematical equivalence relation to discuss, so as to provide data analysis, especially inaccurate, incomplete data analysis, with a new set of mathematical methods. At the same time, Rough Set Theory need not to provide any set of prior information except the data needed to deal with, which has become very important new methods and new theories in field of the intelligent information processing. During recent years, Rough Sets Theory have achieved remarkable results both in theoretical research and in the application of theory.Artificial Neural Networks is a kind of parallel information processing algorithms mathematical model which simulates animal neural network. This network relies on the complexity of the system by adjusting the internal connection between the large number of nodes relations, so as to achieve the purpose of processing information. Artificial neural networks have adaptive and self-learning abilities, according to a number of pre-reciprocal input - output data can be provided, which analyzes the potential laws and the based on these laws, ultimately outputs results according to new input data. This process of study and analysis is called as "training".Fault Diagnosis is integration subject developed from 60 or 70 years in the last century. With the development of Technology, Facility becomes more complex and difficult of fault diagnosis augment obviously. Therefore, we should research and explore practical fault diagnosis technology, not only for the need of modern product and facility automation, but also for the subject of maintenance and management of modern mechanism facility operating.In this paper, we start from introducing the history of Rough Set Theory, the basic concepts of Rough Set Theory , the models of Rough Set Theory, the history of Neural Network and the structure of BP Neural Network. Then, on the basis of above theories, we carry on the analysis to the fans structure in thermal power plant and point out the fault point possibly occurs, combines Rough Set and the BP Neural Network, carries on the real-time diagnosis to the fault points of fans, establishes the diagnostic program with the C language.