A CNN-BiLSTM Bearing Fault Diagnosis Model Based on an Improved Grey Wolf Optimization Algorithm.

Saved in:

Bibliographic Details
Title:	A CNN-BiLSTM Bearing Fault Diagnosis Model Based on an Improved Grey Wolf Optimization Algorithm.
Authors:	Deng, Xitong¹ dxt@stu.ustl.edu.cn, Wang, Yukun¹ wyk410@ustl.edu.cn, Meng, Qingyao² mengqingyao@qdec.edu.cn
Source:	Engineering Letters. May2026, Vol. 34 Issue 5, p1874-1891. 18p.
Subjects:	Fault diagnosis, Grey Wolf Optimizer algorithm, Hilbert-Huang transform, Convolutional neural networks, Long short-term memory, Mechanical vibration research, Deep learning
Abstract:	As industrial machinery continues to progress, ensuring the reliable identification of bearing abnormalities has become a key topic in mechanical engineering. To address the need for fast and precise fault assessment, this work develops a CNN-BiLSTM-based diagnostic approach enhanced by an Improved Grey Wolf Optimization strategy (HSGWO). In the proposed framework, the HSGWO-tuned ICEEMDAN-PE method is first applied to condense and preprocess vibration measurements. A combined CNN and BiLSTM network is then constructed to perform fault classification and condition forecasting, taking advantage of the CNN's strong feature-extraction capability and the BiLSTM's effectiveness in modeling temporal dependencies. Moreover, HSGWO is used to automatically search for suitable network hyperparameters--such as hidden-layer size, learning rate, and L2 penalty--allowing the model to better capture complex signal patterns and improving its robustness. Experimental findings indicate that the designed model consistently surpasses traditional machine-learning algorithms and several advanced deep-learning baselines, showing notable gains in accuracy, precision, recall, and F1 score. [ABSTRACT FROM AUTHOR]
	Copyright of Engineering Letters is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	As industrial machinery continues to progress, ensuring the reliable identification of bearing abnormalities has become a key topic in mechanical engineering. To address the need for fast and precise fault assessment, this work develops a CNN-BiLSTM-based diagnostic approach enhanced by an Improved Grey Wolf Optimization strategy (HSGWO). In the proposed framework, the HSGWO-tuned ICEEMDAN-PE method is first applied to condense and preprocess vibration measurements. A combined CNN and BiLSTM network is then constructed to perform fault classification and condition forecasting, taking advantage of the CNN's strong feature-extraction capability and the BiLSTM's effectiveness in modeling temporal dependencies. Moreover, HSGWO is used to automatically search for suitable network hyperparameters--such as hidden-layer size, learning rate, and L2 penalty--allowing the model to better capture complex signal patterns and improving its robustness. Experimental findings indicate that the designed model consistently surpasses traditional machine-learning algorithms and several advanced deep-learning baselines, showing notable gains in accuracy, precision, recall, and F1 score. [ABSTRACT FROM AUTHOR]
ISSN:	1816093X