Self-adaptive fault diagnosis of roller bearings using infrared thermal images

Fault diagnosis of roller bearings in rotating machinery is of great significance to identify latent abnormalities and failures in industrial plants. This paper presents a new self-adaptive fault diagnosis system for different conditions of roller bearings using InfraRed Thermography (IRT). In the first stage of the proposed system, 2-Dimensional Discrete Wavelet Transform (2D-DWT) and Shannon entropy are applied respectively to decompose images and seek for the desired decomposition level of the approximation coefficients. After that, the histograms of selected coefficients are used as an input of the feature space selection method by using Genetic Algorithm (GA) and Nearest Neighbor (NN), for the purpose of selecting two salient features that can achieve the highest classification accuracy. The results have demonstrated that the proposed scheme can be employed effectively as an intelligent system for bearing fault diagnosis in rotating machinery.