Integrating machine learning and IoT for real-time predictive maintenance in industrial ecosystems: A case study analysis

Improving operational efficiency, optimizing resource utilization, and reducing downtime can all be achieved by applying  Artificial Intelligence (AI)-driven Predictive Maintenance (PdM) in manufacturing. This study explores the application of AI in PdM across various manufacturing sectors, highlighting the transition from traditional methods to smart solutions. Current PdM approaches often rely on time-based or manual inspections, which can be costly and inefficient. These methods suffer from high false-positive rates, inaccurate fault detection, and an inability to predict complex system interactions in Real-Time (RT). As a result, unplanned downtime and maintenance activities remain a significant challenge in manufacturing operations. The proposed study introduces the Smart Manufacturing using AI (SM-AI) framework to address these issues. SM-AI integrates Machine Learning (ML) algorithms, RT, Data Analytics (DA), and Internet of Things (IoT) sensors to provide an intelligent, data-driven PdM solution. CPS architecture enables RT fault detection, optimizes scheduling, and reduces unnecessary interventions by accurately predicting system failures before they occur. The proposed method is applied across multiple manufacturing case studies, demonstrating its ability to enhance predictive accuracy and reduce operational disruptions. SM-AI integrates predictive models with operational technology, ensuring seamless communication across machines and processes. The findings reveal that the SM-AI framework significantly improves maintenance scheduling accuracy, reduces false alarms, and minimizes downtime. Implementing AI-driven PdM provides tangible cost savings, better asset utilization, and higher production efficiency, highlighting the capability of AI integration in future manufacturing ecosystems.

کلیدواژه ها:

Smart Manufacturing ، AI ، Machine Learning ، PdM ، Ecosystems ، CPS