Font Size: a A A

Research On The Intelligent Fault Diagnosis Technology Based Of Support Vector Machine

Posted on:2008-05-19Degree:MasterType:Thesis
Country:ChinaCandidate:G Q YuFull Text:PDF
GTID:2178360242470835Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Support vector machine (SVM) and the neural network are both currently hot subject in the area of machine learning technology. The difference between them is that the former is based on the structural risk minimization principle and the latter is based on the experiential risk minimization principle. Both of them suit to fault diagnosis very much. However, the latter must have many sample data for fault diagnosis, and it have some other disadvantages including training at the slow speed and confirming its structure difficultly and getting into the part tiny value easily and the weak generalization ability. The former not only can solve the problem of the small sample, but also have the advantages of the global optimization and the better generalization ability. In view of their advantages and disadvantages, the application of the neural network based on SVM is studied to equipment fault diagnosis in the paper from the aspect of the theory and method of SVM combined the theory of the neural network, and examples are used to make a study of the validity of the application in fault diagnosis.At the beginning of the dissertation, some important concepts about statistical learning theory are introduced, such as the structural risk minimization principle and VC dimension; Then the concept and solution process of SVM are introduced. SVM is a kind of raised optimized question from that and its solution has the characteristic of overall optimum and it also have stronger generalization ability. The radial basis function neural network has the advantages of the optimal approach and the global optimization. SVM is originally introduced under the condition of linear and divisible and two kinds and is developed to an effectual means to solve the problems of nonlinear and many kinds of pattern recognition. The SVM that one kind is opposite to others is mainly introduced to solve those problems in the paper. The radial basis function neural network is also introduced. Both of them are used to diagnose an example in the paper. Some mended diagnostic methods are presented by combining their advantages and disadvantages in the paper .These methods are as follows: an approach of the radial basis function neural network combined the character transformation based on SVM for fault diagnosis, an approach of the radial basis function neural network optimization based on SVM, an approach of the radial basis function neural network optimized based on SVM combined the character transformation based on SVM. They are all used to diagnose an example.The diagnostic results indicate that the approach of the radial basis function neural network combined the character transformation based on SVM for fault diagnosis not only simplifies the dimension of the sample data, but also improves the training speed of the radial basis function neural network and obtains a more accurate diagnosis; The approach of the radial basis function neural network optimization based on SVM confirms and simplifies the structure of the network and improves the training speed of the radial basis function neural network ,and obtains a more accurate diagnosis; The approach of the radial basis function neural network optimized based on SVM combined the character transformation based on SVM not only simplifies the dimension of the sample data, but also confirms and simplifies the structure of the network and improves the training speed of the radial basis function neural network ,and obtains a more accurate diagnosis.
Keywords/Search Tags:Support vector machine, Radial basis function neural network, Character transform, Fault diagnosis
PDF Full Text Request
Related items