Gearbox Failure Diagnosis Using a Multisensor Data-Fusion Machine-Learning-Based Approach

• Published in: Entropy (Basel, Switzerland) Switzerland), 2021-06, Vol.23 (6), p.697
• Main Authors: Habbouche, Houssem, Benkedjouh, Tarak, Amirat, Yassine, Benbouzid, Mohamed
• Format: Article
• Language: English
• Subjects: Accuracy; Artificial intelligence; Artificial neural networks; Classification; Comparative studies; convolutional neural network; Data integration; Diagnosis; Downtime; Electric power; Engineering Sciences; Expert systems; Failure; Failure detection; Fault diagnosis; Feature extraction; gearbox failure; Gearboxes; Internet of Things; linear predictive coefficients; long short-term memory; Machine learning; mel-frequency cepstral coefficients; Multisensor fusion; Neural networks; Shutdowns; Signal processing; Support vector machines; Useful life
• ISSN: 1099-4300; 1099-4300
• DOI: 10.3390/e23060697

Failure detection and diagnosis are of crucial importance for the reliable and safe operation of industrial equipment and systems, while gearbox failures are one of the main factors leading to long-term downtime. Condition-based maintenance addresses this issue using several expert systems for early failure diagnosis to avoid unplanned shutdowns. In this context, this paper provides a comparative study of two machine-learning-based approaches for gearbox failure diagnosis. The first uses linear predictive coefficients for signal processing and long short-term memory for learning, while the second is based on mel-frequency cepstral coefficients for signal processing, a convolutional neural network for feature extraction, and long short-term memory for classification. This comparative study proposes an improved predictive method using the early fusion technique of multisource sensing data. Using an experimental dataset, the proposals were tested, and their effectiveness was evaluated considering predictions based on statistical metrics.