Abstract

Predictive maintenance (PDM) is becoming increasingly important across industries, as accurate fault detection and timely failure prediction are essential for minimizing downtime, reducing operational costs, and optimizing machine performance, ultimately leading to more sustainable and efficient maintenance systems. Advance PdM enables precise analysis, forecasting failures, and optimizing maintenance schedules and plays a key role using artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL) techniques. This review paper examines the current limitations and opportunities associated with deploying AI for PDM. It presents key methods and strategies to overcome existing challenges and highlights emerging opportunities, such as the integration of AI with the Internet of Things (IoT) and edge computing, which enhance real-time decision-making and system scalability. By synthesizing recent advances and identifying research gaps, this study aims to guide future developments in leveraging AI for more effective and sustainable machine maintenance systems.

Authors

Protik Barua¹, Rajnita Barua², Mumit Hassan³, Imran Hossain⁴
World University of Bangladesh, Bangladesh^1,4, Chittagong University of Engineering and Technology, Bangladesh^2,3

Keywords

Predictive Maintenance (PdM), Machine Learning (ML), Deep Learning (DL), Artificial Intelligence (AI), IoT