AI-Enabled Preventive Maintenance for Industry 5.0: Advancing Equipment Reliability, Operational Resilience, and Sustainable Manufacturing

Abstract

Industry 5.0 represents a new dimension in the manufacturing industry with its focus on human-centeredness, sustainability, and resilience. The core of this change is the transformation of asset management from a reactive or purely scheduled approach to one based on artificial intelligence – proactive asset management. The purpose of this paper is to explore AIoT-based predictive and preventive maintenance within the Industry 5.0 framework and introduce the concept of Resilience-Based Maintenance (RBM). The authors propose a layered architecture that incorporates the use of DL models - namely, CNNs and Transformers – along with the high-fidelity Digital Twins within an Edge-Cloud cooperative computing environment. This enables processing high-frequency sensor data (vibration, thermal, and audio) at the edge with immediate detection of anomalies, while making long-term RUL estimation and fleet analytics in the cloud. Furthermore, the article investigates the impact of AI-based maintenance on the reliability of machinery and sustainability rates. The framework is to curtail industrial energy consumption, diminish material waste, and cut down carbon emissions by adaptive scheduling of maintenance as per energy-efficiency anomalies, and decreasing mechanical friction. In addition, the study looks into the important socio-technical challenges that hinder implementation, including the usage of Explainable AI (XAI) through SHAP frameworks to boost operators’ trust, how to resolve data shortage issues through physics-informed neural networks (PINNs), and the promotion of human-in-the-loop collaboration to improve cognitive ergonomics in manufacturing.