Abstract
The noise and outliers offer difficulty in monitoring the health of the bearings. In the present study, a new bearing fault detection technique is suggested to target these outliers. The proposed technique is based on the comparison between the entropy and statistical features extracted from modes functions obtained after decomposing acquired signals through variational mode decomposition (VMD). The study is derived from the work presented at the 8th International Conference on Signal Processing and Integrated Networks (SPIN), 2021. The five entropy-statistical features were extracted from the modes and were considered for fault’s classification. Different fusion matrixes of entropy values and statistical features were obtained. These matrixes derived for different experimental conditions were then used for the classification of different bearing faults. The classification was based on two states-of-art classifiers: support vector machine and artificial neural network. The effectiveness of entropy-based feature and statistical feature was compared. The experimental investigations showed that the entropy-based features were effective over statistical features in detecting various bearings.
Acknowledgements
The authors are thankful to the Case Western Reserve University, Cleveland, Ohio for keeping the open access to the bearing data.
DISCLOSURE STATEMENT
No potential conflict of interest was reported by the author(s).
Additional information
Notes on contributors
Abhishek Anand Singh
Abhishek Anand Singh received his BE degree in mechanical engineering from Chhattisgarh Swami Vivekanand Technical University, Chhattisgarh, India in 2019 and MTech degree in design engineering from Dr B R Ambedkar National Institute of Technology, Jalandhar in 2021. His research interest includes the fault diagnosis of rolling bearings signal processing. Email: [email protected]
C. I. Harikrishnan
C I Harikrishnan received his BTech degree in mechanical engineering from Govt College of Engineering, Kannur, India in 2018 and MTech degree in design engineering from the National Institute of Technology, Jalandhar, India in 2021. He is currently working in technical sales department of Panchsheel Filters, Pune, India. His areas of interest are vibration analysis, condition monitoring, artificial intelligence and machine learning. Email: [email protected]
S. K. Tiwari
S K Tiwari received his BE degree in mechanical engineering from Dr Babasaheb Ambedkar Marathwada University, formerly Marathwada University, Aurangabad, Maharashtra, India, in 1992 and MTech degree in design engineering and PhD in mechanical engineering from National Institute of Technology, Jalandhar, India in 2008 and 2015, respectively. He is currently associate professor in the Department of Mechanical Engineering, National Institute of Technology, Jalandhar, India. His areas of interest are friction stir welding, machine fault diagnosis, artificial intelligence, machine learning and signal processing. Email: [email protected]
Snehsheel Sharma
Snehsheel Sharma received his BTech degree in mechanical engineering and MTech in production engineering from Punjab Technical University, Punjab, India, in 2003 and 2007, respectively. He is currently pursuing PhD in the Department of Mechanical Engineering, National Institute of Technology, Jalandhar, India. His areas of interest are machine fault diagnosis, artificial intelligence, machine learning and signal processing. Corresponding author. Email: [email protected]