Distributed fault diagnosis of rotor-bearing system using dimensional analysis and experimental methods

• Published in: Measurement : journal of the International Measurement Confederation Vol. 166; p. 108239
• Main Authors: Jadhav, Prakash M., Kumbhar, Surajkumar G., Desavale, R.G., Patil, Shubham B.
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 15.12.2020; Elsevier Science Ltd
• Subjects: Bearings; Catastrophic failure analysis; Defects; Dimensional analysis; Distributed defects; Experimental methods; Fault diagnosis; Finite element analysis; Localized defects; Machinery; Machinery condition monitoring; Machining; Maintenance costs; Mathematical models; Regression analysis; Roller bearings; Rotor-bearing systems; Signatures; Vibration; Vibration analysis; Distributed defects; Regression analysis; Localized defects; Dimensional analysis; Vibration analysis
• ISSN: 0263-2241; 1873-412X
• DOI: 10.1016/j.measurement.2020.108239

•A need to model bearings for detecting distributed defects.•Dimension theory with Pi-theorem saves modeling time & efforts.•Dimensional analysis accurately detects distributed faults.•Dimensional model and regression analysis simplify the bearing diagnosis.•Distributed defects produce different vibrations than localized defects. Distributed defects in rolling element bearing arise due to the progression of small localized pits. If not monitored on time, this leads to premature or may turn into catastrophic failure of a rotor system and thus increases the cost of maintenance of machinery in industries. This research work introduced a novel method for the diagnosis of rolling element bearings to detect distributed defects using Dimensional Analysis (DA). The rolling element bearing system is modeled using DA with Buckingham’s theorem to study the vibration characteristics theoretically. Significant geometrical and response parameters are considered for modeling the rotor-bearing system. The in-house experimental test rig is designed and fabricated to study the vibration characteristics of distributed defects experimentally. Geometrical imperfections are created intentionally on the bearings surfaces using electro-discharge machining and vibration signatures generated are studied under different operating conditions. Experimental results are compared with analytical models developed using the DA. A comparison is made between distributed and localized defects and vibrations signatures from both are correlated with healthy bearings. It is noticed that there is a good correlation between theoretical and experimental results. The present study contributes as a simple and effective diagnosis of rolling element bearings for distributed defects and useful for condition monitoring of industrial machinery.