Rough Sets In Fault Diagnosis

Rough set is a good tool for the math problem. The advantage of Rough set is the ability to acquire the knowledge. Today artificial neural network develops very fast. It has a lot of advantages, such as self-learning and self-organization. But it also has a problem. It is difficult to make a decision from it.Rough set theory can make some improvements. Based on the theory of rough set, this dissertation studies rough set-artificial neural network. A newly reduced algorithm based on rough set theory is proposed for fault diagnosis when the power network failed and then caused possibly the protections to malfunction and defects in communication. Furthermore, the reduced result can be synthesized into a tabulated expert decision library. With fuzzy sets and probability applied to the rules of rough sets, the confidence levels of each rule and relevant equipment are taken into consideration and computed. Another algorithm is also proposed to analyze synthetically the confidence level of a diagnostic conclusion in accordance to the number of rules for a certain decision making process and the confidence level of each rule , which is also to be put into application to power network fault diagnosis. In the example given as fault diagnosis, The example proves this method which makes fault diagnosis exact and the speed of the proceeding is fast.