TY - CHAP

T1 - Machine Learning Analysis in the Diagnostics of the Dynamics of Ball Bearing with Different Radial Internal Clearance

AU - Ambrożkiewicz, Bartłomiej

AU - Syta, Arkadiusz

AU - Gassner, Alexander

AU - Georgiadis, Anthimos

AU - Litak, Grzegorz

AU - Meier, Nicolas

N1 - Conference code: 10

PY - 2023

Y1 - 2023

N2 - Interpretation of acceleration time-series from rolling-element bearings is sometimes challenging if no prior knowledge of the system is available. The evaluation must adapt operational conditions or the actual value of operational parameters. In our analysis, we apply the machine learning methods and statistical indicators in the diagnosis of dynamical response in the self-aligning ball bearing with different radial internal clearance. Machine learning methods are applied for the quantification of the acceleration time-series with statistical indicators and their assignation to the specific state or clearance value. The results of the analysis allow recognizing the bearing’s condition and the clearance value based on experimental acceleration time-series. Additionally, confusion matrices are presented for showing the accuracy of proposed methods. The results of applied Machine Learning methods are on the level of around 80% in classifying the dynamical response to the specific radial clearance. The motivation of the research is to introduce it to on-site practice in the test rig.

AB - Interpretation of acceleration time-series from rolling-element bearings is sometimes challenging if no prior knowledge of the system is available. The evaluation must adapt operational conditions or the actual value of operational parameters. In our analysis, we apply the machine learning methods and statistical indicators in the diagnosis of dynamical response in the self-aligning ball bearing with different radial internal clearance. Machine learning methods are applied for the quantification of the acceleration time-series with statistical indicators and their assignation to the specific state or clearance value. The results of the analysis allow recognizing the bearing’s condition and the clearance value based on experimental acceleration time-series. Additionally, confusion matrices are presented for showing the accuracy of proposed methods. The results of applied Machine Learning methods are on the level of around 80% in classifying the dynamical response to the specific radial clearance. The motivation of the research is to introduce it to on-site practice in the test rig.

KW - Machine learning

KW - Radial internal clearance

KW - Rolling-element bearing

KW - Statistical indicators

KW - Informatics

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=85141786239&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/3bd7e3c7-0248-3701-9c92-4bd2af122847/

U2 - 10.1007/978-3-031-15758-5_61

DO - 10.1007/978-3-031-15758-5_61

M3 - Article in conference proceedings

AN - SCOPUS:85141786239

SN - 978-3-031-15757-8

T3 - Mechanisms and Machine Science

SP - 599

EP - 606

BT - Recent Trends in Wave Mechanics and Vibrations

A2 - Dimitrovová, Zuzana

A2 - Gonçalves, Rodrigo

A2 - Biswas, Paritosh

A2 - Silva, Tiago

PB - Springer Schweiz

CY - Cham

T2 - 10th International Conference on Wave Mechanics and Vibrations - WMVC 2022

Y2 - 4 July 2022 through 6 July 2022

ER -