Fault diagnosis of rolling element bearings using fractional linear prediction and AI techniques

• DOI: 10.1007/s41872-018-0062-8
• Publication Date: 2018-08-29T11:15:16Z
• Authors: Aditya Sharma; Sharad Bhardwaj; Pavan Kumar Kankar
• Keywords: Rolling-element bearing

Rolling element bearings are one of the widely used and most critical components a rotating machinery. The performance of bearings is utmost important in applications such as power plants, automobiles, turbines, aerospace, materials handling and many more. In this paper, a new technique, fractional linear prediction, is presented for the fault diagnosis of bearings. For the examination of the proposed methodology, the vibration data of two different types of bearings are selected and analyzed. Three artificial intelligence techniques—rotation forest, support vector machine and artificial neural network—are used for the investigations. Comparison is also carried out among the artificial intelligence techniques to show their effectiveness towards fault diagnosis. Results indicate the superiority of rotation forest over support vector machine and artificial neural network.