Localization of Rolling Element Faults Using Improved Binary Particle Swarm Optimization Algorithm for Feature Selection Task

The accurate localization of the rolling element failure is very important to ensure reliability rotating machinery. This paper proposes an efficient and anti-noise fault diagnosis model for elements. proposed composed feature extraction, selection classification. Feature extraction signal processing noise reduction. Signal carried out by local mean decomposition (LMD), reduction performed product function (PF) wavelet packet (WPD). Through steps reduction, high-frequency can be effectively removed, information hidden under extracted. To further improve effectiveness diagnostic model, improved binary particle swarm optimization (IBPSO) find most features from space. In IBPSO, cycling time-varying inertia weight introduced balance exploitation exploration capability escape solutions, crossover mutation operations are also capabilities, respectively. main contributions this research briefly described as follows: (1) process applied in remove establish a high-accuracy set. (2) algorithm has higher accuracy than other state-of-the-art algorithms. (3) strong environment, compared with terms classification accuracy. Experimental results show that maintain high environment. Therefore, it proved