This paper studies a method of compressor fault diagnosis based on Rough Sets theory , which is one of the latest tools in Data Mining area. Not like the usual methods that based on mechanical vibrancy, this method combines the Rough Sets theory with the Artificial Neural Network, which makes the best of the large characteristic running data from the DCS .In virtue of the ability Rough Sets has to decrease the amount of data and processes it with a reduction. It can reduce the amount of the training data and overcome BP's defect of slow training speed when process large data set, thus results in finding the rules and the diagnosis knowledge with fairly good performance.Rough Sets theory has made fast progress in recent years, it has outstanding ability in research of expressing, learning , concluding non-precise knowledge. It is based on practical large data sets, and deduce, find the knowledge and key of the classification systems.In this point, the Rough Sets theory is fit close with the facts of the compressor. That's why introduces this powerful tool. Applied with the...
|