Fault Diagnosis Of Rolling Bearing Based On Optimized Support Vector Machine

Bearings are one of the key components and parts of rotating machinery but easily lead to wear and tear parts. Operating state of bearings often directly affect the performance of the whole machine. Fault diagnosis on bearings is a subject of great concern of academic and engineering. At present, there is an urgent need to establish bearing monitoring system in rotating machinery, through the off-line or on-line monitoring bring abou real-time monitoring for bearing health conditions and integrated diagnosis system of fault reason and serious level, thereby reasonably arrange the maintenance project and repair time. This kind of condition monitoring device can not only find the health problems on failure bearings themself, but also can detect the health of the whole machine or other parts. It has become to an important content of promoting the modern equipment management. Therefore, to study fault diagnosis on bearings is very important.This paper gives a comprehensive overview on the existing method of intelligent fault diagnosis, points out the advantages and disadvantages of various methods, and introduces the basic structure and types of rolling bearing on rotating machinery and their common failure form, also briefly introduces the vibration mechanism and current fault diagnosis methods on rolling bearings.This paper introduces the principle of support vector machine(SVM) and several commonly used kernel function, analyse the principle of the artificial bee colony algorithm, and propose optimization method for support vector machine based on artificial bee colony algorithm. In order to prove the effectiveness of artificial bee colony algorithm, the first work of this paper is to choose three representative UCI standard data sets for training and experimental verification. The experimental results show that optimized support vector machine by the artificial bee colony algorithm can overcome the local optimal solution, and obtain higher classification accuracy, but also reduce running time in the small number classification. It make theoretical preparation for intelligent diagnosis of fault bearing below.Based on the above results, the second work of this paper is to use optimized support vector machine by the artificial bee colony algorithm for intelligent diagnosis of fault bearing. First this paper describes the characteristics of vibration signal of rolling bearings, the method for signal acquisition and its analysis, and then put forward the diagnosis model of fault bearing based on support vector machine, and the parameters for diagnosis model are optimized. Finally, this paper takes the 50 sets of data for nonnal rolling bearing and fault rolling bearing because of corrosive pitting in their rolling body, inner ring, or outer ring, to carry out intelligent diagnosis test. Optimized support vector machine by artificial bee colony algorithm is compared with genetic algorithm, ant colony algorithm, the standard particle swarm optimization in performance. The results show that optimized support vector machine by the artificial bee colony algorithm can overcome the local optimal solution, obtain higher classification accuracy, which prove that there is great valuable application of engineering for optimized support vector machine by the artificial bee colony algorithm.