Research On Fault Diagnosis Of Nuclear Power Equipment Based On Support Vector Machine

A new learning machine—support vector machine(SVM) is one of the statistical learning theories, which also based on the statistical learning theory. Now, SVM is viewed as the most convincing tool in solving the classification and regression problems and the research focus in the filed of machine learning after neural network. On the basis of structural risk minimization and VC dimension, it seeking the optimum tradeoff between model complexities and learning abilities under the limited samples, to achieve the best generalization. SVM is regard as a good development to traditional classifier, showed many unique advantages in solving small samples, nonlinear, high dimension and other machine learning problems.The main contents of this thesis are as follows:(1) In this thesis, SVM is introduced the fault diagnosis of nuclear power equipment, and verified the classifying performance of the SVM by experiment.(2) In this thesis, methods about how to extract features from nuclear power equipment has been intensively studied and a new feature extraction method is Successfully applied to mechanical fault diagnoses based on the thought of short-time Fourier Transform.(3) The selection of the SVM-kernel with suitable form and parameters has become a key-point both in theoretical research and application consideration. In combining with RBF kernel function by improving the"exhaustive method"to find the optimal parameter combination and verify the classifying performance before and after optimization by experiment.(4) This thesis also combines the Bagging algorithm of ensemble learning, formed Bagging-SVM, and because of the SVM has good generalization ability, if further improve the performance by ensemble learning, will be a better solution for an application problem, and verify the recognition effect of the Bagging -SVM by experiment.(5) With the traits of the nuclear power equipment, build the SVM response model; design the fault diagnosis system for nuclear power equipment system based on SVM and combine with the software of fault diagnosis which was developed by us to test the system to verify the feasibility of the whole fault diagnosis system.This research can promote the progress of on-line monitoring and technology of fault diagnosis; have important theoretical and practical value in guaranteeing the nuclear power equipment's safe and reliable operations and increasing the combat effectiveness of the national defense.