Data Mining Techniques for Predictive Maintenance in Manufacturing Industries a Comprehensive Review

¹ CEO MahaaAi Group of Companies and International Labs, Dallas Texas, USA
² Associate Professor, Department of CSIT, Koneru Lakshmaiah Education Foundation, Vaddeswaram, Guntur District, Andhra Pradesh, India
³ Professor, Department of Mechanical Engineering, J.J.College of Engineering and Technology, Tiruchirappalli, Tamil Nadu, India
⁴ Department of Computer Science and Engineering, MLR Institute of Technology, Hyderabad, Telangana, India
⁵ Associate Professor & HOD, Department of Electronics and Communication Engineering, Study World College of Engineering, Coimbatore, Tamil Nadu, India
⁶ Assistant Professor, Department of ECE, New Prince Shri Bhavani College of Engineering and Technology Chennai, Tamil Nadu, India

Narender.chinthamu@gmail.com
dr.aashishsinha@gmail.com
mathis09051970@yahoo.co.in
dev.bolleddu@gmail.com
kannadhasan.ece@gmail.com
suganya.ece@newprinceshribhavani.com

Abstract

Predictive maintenance (PdM) is one of the major methods used in modern manufacturing to realize downtime minimization, lower the cost of maintenance and maximize machine service life by analyzing the collected data using data mining methodologies. However existing works mainly focus on conventional ML models without provide systems design real world applications systems and do not include any dimension related to network security dimension, cost and benefit analyzing dimension utility dimension and light weight A.I model for edge computing. In this paper, we contribute with a systematic literature review of state-of-the-art data-mining techniques for predictive maintenance with emphasis on hybrid AI frameworks, deep learning and online data processing approaches, as well as, privacy-aware methods. We contribute by providing a number of real-world industrial use case which differentiate us from previous researched; we discuss details of cybersecurity issues in IoT-enabled PdM; and we discuss use of XAI (Explainable AI) to build interpretable models. Moreover, this survey introduces marginal AI applications in edge computing, predictive maintenance frameworks with scalability, and AI-powered anomaly identification for enhancing predictions in industrial-scale production. It also covers a review of predictive maintenance methodologies in addition to a future research agenda, highlighting emerging patterns such as digital twins, Industry 5.0, and reinforcement learning in predictive maintenance. The current study aims to bridge critical gaps in the literature and support valuable direction for researchers, industry practitioners and policymakers for effective predictive maintenance strategies and task performance.