Application of fuzzy C-means method and classification model of optimized K-nearest neighbor for fault diagnosis of bearing

Abstract

To make accurate judgments of bearing main fault types using relatively small amounts of fault data, a fault diagnostic method was proposed in this paper. This method combined fuzzy C-means method (FCM) and optimized K-nearest neighbor model to achieve the aim. Firstly, the local mean decomposition and its Shannon entropy are used to analyze the collected bearing vibration signals, so the characteristic values fault features are extracted. Then, FCM was used to pre-classify fault training samples, so as to classify the typical features basically. Then, the K-nearest neighbor model optimized by the particle swarm optimization model was established and used to achieve the bearing fault diagnosis finally. Experimental results have proven that the proposed method achieves perfect results in diagnosing bearing’ faults.