ISSN 2384-5058
Abstract
The combination of digital twin (DT) technology with artificial intelligence (AI) has been recognised as a major change in the way a predictive maintenance system works in Industry 4.0 manufacturing environments. This research presents a hybrid DT architecture that uses real-time IoT sensor data and physics-based deep learning to forecast equipment failures even before they occur. The system proposed makes it possible to less frequent interruptions of the production process, to save the time of maintenance, and to adjust production changes automatically. The results of the verification process, which was simulation and a CNC machining case study, confirm the effectiveness of the approach in achieving higher predictive accuracy, lowering the frequency of maintenance tasks, and increasing operational efficiency.
Keywords: Digital Twin, Predictive Maintenance, Industry 4.0, AI, Physics-Informed Neural Networks, RUL