ADAPTIVE FAULT PREDICTION AND MAINTENANCE IN PRODUCTION LINES USING DEEP LEARNING

Abstract

In the era of Industry 4.0 and intelligent manufacturing, accurately predicting and maintaining production line faults is crucial in manufacturing. This study introduces a novel deep learning-based adaptive fault prediction and maintenance strategy, overcoming limitations of traditional statistical and machine learning methods in prediction accuracy and adaptability in complex environments. A new prediction model is developed, incorporating a wide convolutional feature extraction module, a customized gating module, and a multi-layered progressive extraction module. The model's process and parameters, including fault stage division using Wasserstein distance and optimization with L2 regularization and neuron dropout, are detailed. An adaptive maintenance strategy for predictive fault detection is established, enhancing precision in fault prediction and developing more effective maintenance approaches, ultimately boosting production efficiency and reducing operational costs. (Received in August 2023, accepted in October 2023. This paper was with the author 5 weeks for 1 revision.).