ABSTRACT
The emergence of Industry 4.0 has led to the development of modern production machines that are usually equipped with advanced sensors to collect data for further analysis. This study proposes a deep learning-based framework to perform equipment health monitoring (EHM) and further broadens its applicability through the integration of subject matter expert (SME) knowledge. A sliding window strategy was adopted to perform EHM in real time. Moreover, an autocorrelation function (ACF) and a partial autocorrelation function (PACF) were employed to determine the optimal window size based on the elbow method. An empirical study was conducted to demonstrate the effectiveness and practicality of the proposed framework. Furthermore, to provide better prediction results, an optimal combination of hyperparameters that minimized the loss function and further validated the window size obtained by the ACF and PACF was determined and used. The results showed that the proposed algorithm outperformed other representative machine learning models. Finally, a general framework was adopted to maintain equipment performance.
Acknowledgements
The author would like to acknowledge a very good collection of data from Hsiang-Po Tsai and the support from Hong-Yi Huang.
Disclosure statement
No potential conflict of interest was reported by the author(s).