Application of Self Organizing Map for Intelligent Machine Fault Diagnostics Based on Infrared Thermography Images

• Published in: 2011 Sixth International Conference on Bio-Inspired Computing: Theories and Applications pp. 123 - 128
• Main Authors: Widodo, A., Satrijo, D., Huda, M., Gang-Min Lim, Bo-Suk Yang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2011
• Subjects: Cameras; Clustering algorithms; Condition monitoring; Data acquisition; Feature extraction; infrared thermography; machine fault diagnostics; Neurons; pattern recognition; Rolling bearings; self organizing map
• ISBN: 1457710927; 9781457710926
• DOI: 10.1109/BIC-TA.2011.15

This paper concerns with implementation of self organizing map (SOM) for intelligent machine fault diagnostics. The present study employs infrared images acquired by thermography camera as data base of machine diagnostics system. Image processing is carried out using thresholding for image segmentation and clustering by means of k-means algorithm. Feature extraction of images is conducted by calculating area, perimeter and central moment of region of interest (ROI). All data of this work was acquired by capturing the images of rolling element bearings from rotating machine fault simulator (MFS). The simulator is able to experiment a normal and seeded fault conditions such as outer and inner race defects of rolling element bearing, unbalance, misalignment and looseness. Pattern recognition technique is then employed to diagnose the machine conditions by mapping the image features through SOM. The result shows that SOM based infrared thermography image can perform intelligent machine fault diagnostics with plausible accuracy.