Ball bearing defect models: A study of simulated and experimental fault signatures. (21st July 2017)

• Record Type: Journal Article
• Authors: Mishra, C.; Samantaray, A.K.; Chakraborty, G.
• Is Part Of: Journal of sound and vibration. Volume 400(2017)
• Page Start: 86
• Page End: 112
• Publication Date: 2017-07-21
• Subjects: Rolling element bearing -- Simulated vibration signature -- Fault diagnosis -- Bond graph -- Multi-body dynamics; Sound -- Periodicals; Vibration -- Periodicals; Son -- Périodiques; Vibration -- Périodiques; Sound; Vibration; Periodicals; Electronic journals; 620.205
• DOI: 10.1016/j.jsv.2017.04.010
• Languages: English
• ISSNs: 0022-460X

Abstract

Abstract: Numerical model based virtual prototype of a system can serve as a tool to generate huge amount of data which replace the dependence on expensive and often difficult to conduct experiments. However, the model must be accurate enough to substitute the experiments. The abstraction level and details considered during model development depend on the purpose for which simulated data should be generated. This article concerns development of simulation models for deep groove ball bearings which are used in a variety of rotating machinery. The purpose of the model is to generate vibration signatures which usually contain features of bearing defects. Three different models with increasing level-of-complexity are considered: a bearing kinematics based planar motion block diagram model developed in MATLAB Simulink which does not explicitly consider cage and traction dynamics, a planar motion model with cage, traction and contact dynamics developed using multi-energy domain bond graph formalism in SYMBOLS software, and a detailed spatial multi-body dynamics model with complex contact and traction mechanics developed using ADAMS software. Experiments are conducted using Spectra Quest machine fault simulator with different prefabricated faulted bearings. The frequency domain characteristics of simulated and experimental vibration signals for different bearing faults are compared and conclusions are drawn regarding usefulness of the developed models. Highlights: Three models of faulty ball bearings are used to generate vibration signatures. The first is a simple kinematics based model developed in block diagram form. The second is a bond graph model which includes cage and traction dynamics. The third is a complex multi-body dynamics model developed using ADAMS software. Frequency domain results of the third model yield best match with the experiments.