Rolling bearing fault diagnosis by a novel fruit fly optimization algorithm optimized support vector machine

Based on the nonlinear and non-stationary characteristics of rotating machinery vibration, a FOA-SVM model is established by Fruit Fly Optimization Algorithm (FOA) and combining the Support Vector Machine (SVM) to realize the optimization of the SVM parameters. The mechanism of this model is imitating the foraging behavior of fruit flies. The smell concentration judgment value of the forage is used as the parameter to construct a proper fitness function in order to search the optimal SVM parameters. The FOA algorithm is proved to be convergence fast and accurately with global searching ability by optimizing the analog signal of rotating machinery fault. In order to improve the classification accuracy rate, built FOA-SVM model, and then to extract feature value for training and testing, so that it can recognize the fault rolling bearing and the degree of it. Analyze and diagnose actual signals, it prove the validity of the method, and the improved method had a good prospect for its application in rolling bearing diagnosis.