Researches On The Rough Sets And Fault Diagnosis

Soft computing includes Artificial Neural Networks, Fuzzy Logic, Evolutionary Algorithms, Rough Sets (RS) Theory, etc. As a new soft computing, Rough Sets can analyze and handle imprecise, inconsistent and incomplete data efficiently. Rough Set is a powerful tool for analyzing data and is tolerant for faults. Hence, Rough Set can overcome shortages of other soft computing in some aspects. So using the advantages of RS, the paper does some pilot study on the fault detection and diagnosis. The majority of the work is listed as follows:1. Discretization study based on the Rough Sets.Rough Sets can only deal with the discretization data, which is the choke point for its application. The discretization result directly has an influence on the later deal with Rough Sets method. The ill-conditioned problem may exist after discretization of continuous features, a new optimal discretization method based on the Artificial Fish-swarm Algorithm was proposed in this paper. It turned the problem of divide points making to the problem of searching the optimal value with the Artificial Fish-swarm Algorithm. By using visual divide points, the proposed method makes area's combination more efficiently.2. Rough Sets Neural NetworksRough Sets simulate the abstract ability, while the artificial neural network (ANN) has the reasoning ability. This paper combines the two methods, that is using the Rough Sets to mine the useful information from the sample data, then the ANN is used for training. Besides, the Rough Sets based adaptive neuro-fuzzy inference system (ANFIS) is proposed, different from the general ANFIS, the rule is mined by the Rough Sets, which reduce some unnecessary connections. The FDD system of Rough sets Neural Networks is explodered in Matlab software.3. FDD based on Support Vector MachineFault diagnosis is the pattern recognition problem. The support vector machine (SVM) theory is proposed to solve the classification problem. In this paper, the comparison of the classification ability is done between the ANN and SVM, then present the new fault diagnosis method based the Rough Sets and the Support Vector Machine.Conclude the Rough Sets' core algorithms and SVM's solve algorithms, the GUI of SVM classifier is explodered.