Multisensory Data-Based Fault Diagnosis of Induction Motors Using 1D and 2D Convolutional Neural Networks

Samuel Ayankoso, Yinghang He, Fengshou Gu, Andrew Ball

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Induction motors are widely employed across various industries, which necessitates the development of an effective fault diagnosis system to prevent unplanned downtime. Lately, there has been a notable emphasis on deep learning techniques, primarily attributed to their capacity for directly extracting features from unprocessed data. In this study, we collected multisensory signals (vibration, current, voltage, and speed) from an induction motor test stand under diverse operating conditions, including healthy, outer bearing fault, and shaft misalignment. Seven combinations of the multisensory data were employed to train 1D and 2D Convolutional Neural Network (CNN) architectures for motor fault classification. A comparative analysis was undertaken to evaluate the diagnostic accuracy, number of parameters, and computational time of the 1D and 2D CNN models. The results reveal that the best performance was achieved when the 1D and 2D CNN models were trained solely on the motor’s vibration signals. Additionally, the 2D CNN model slightly outperformed the 1D CNN in terms of its overall validation accuracy under different multisensory signal combinations.

Original languageEnglish
Title of host publicationProceedings of the UNIfied Conference of DAMAS, IncoME and TEPEN Conferences (UNIfied 2023) - Volume 2
EditorsAndrew D. Ball, Huajiang Ouyang, Jyoti K. Sinha, Zuolu Wang
PublisherSpringer, Cham
Number of pages11
ISBN (Electronic)9783031494215
ISBN (Print)9783031494208, 9783031494239
Publication statusPublished - 29 May 2024
EventThe UNIfied Conference of DAMAS, InCoME and TEPEN Conferences - Huddersfield, United Kingdom, Huddersfield, United Kingdom
Duration: 29 Aug 20231 Sep 2023

Publication series

NameMechanisms and Machine Science
Volume152 MMS
ISSN (Print)2211-0984
ISSN (Electronic)2211-0992


ConferenceThe UNIfied Conference of DAMAS, InCoME and TEPEN Conferences
Abbreviated titleUNIfied 2023
Country/TerritoryUnited Kingdom
Internet address

Cite this