Enhanced Gear Damage Prediction Using Convolutional Neural Networks on Time-Series Sensor Data

Sohaib Arshad Mayo, Zhiqiang Cai

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

Abstract

Gearboxes are essential in various industries, serving as the core mechanism for power transmission in machines like wind turbines and vehicles. Traditional fault detection primarily uses vibration analysis, which, although effective, often fails to precisely quantify gear damage. This paper introduces a novel approach by integrating artificial intelligence (AI) with vibration data to enhance gearbox fault diagnosis. By employing machine learning algorithms, our model not only detects but also quantifies the severity of gear faults. This method involves collecting and analyzing vibration signals under different operational conditions to train a deep learning model that can accurately predict fault severity. This innovative approach, to the best of our knowledge, is the first to use AI to quantify gearbox faults, potentially revolutionizing maintenance strategies and improving industrial machinery reliability.

Original languageEnglish
Title of host publication15th Global Reliability and Prognostics and Health Management Conference, PHM-Beijing 2024
EditorsHuimin Wang, Steven Li
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350354010
DOIs
StatePublished - 2024
Event15th IEEE Global Reliability and Prognostics and Health Management Conference, PHM-Beijing 2024 - Beijing, China
Duration: 11 Oct 202413 Oct 2024

Publication series

Name15th Global Reliability and Prognostics and Health Management Conference, PHM-Beijing 2024

Conference

Conference15th IEEE Global Reliability and Prognostics and Health Management Conference, PHM-Beijing 2024
Country/TerritoryChina
CityBeijing
Period11/10/2413/10/24

Keywords

  • component
  • convolutional neural network (CNN)
  • Deep Learning (DL)
  • Machine Learning (ML)
  • Neural Networks (NN)
  • Vibration Analysis

Fingerprint

Dive into the research topics of 'Enhanced Gear Damage Prediction Using Convolutional Neural Networks on Time-Series Sensor Data'. Together they form a unique fingerprint.

Cite this