Predictive maintenance based on IIoT and machine learning aligned with industry 4.0: a case study in waste-water treatment plant

• Published in: Neural computing & applications Vol. 37; no. 24; pp. 20383 - 20407
• Main Authors: Abd Elhaleem, Sameh, Zanfal, Abdallah, Hamdy, Mohamed
• Format: Journal Article
• Language: English
• Published: London Springer London 01.08.2025; Springer Nature B.V
• Subjects: Accuracy; Algorithms; Artificial Intelligence; Asset management; Automation; Communication; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Costs; Data analysis; Data collection; Data Mining and Knowledge Discovery; Decision making; Downtime; Effectiveness; Failure; Fault diagnosis; Image Processing and Computer Vision; Industrial applications; Industrial Internet of Things; Industry 4.0; Machine learning; Maintenance management; Manufacturing; Mathematical programming; Methodology; Neural networks; Object linking & embedding; Original Article; Planning; Predictive maintenance; Preventive maintenance; Probability and Statistics in Computer Science; Process controls; Programmable logic controllers; Real time; Regression analysis; Reliability; Sensors; Structured Query Language-SQL; Vibration monitoring; Wastewater treatment; Water treatment plants; Machine learning (ML); Industry 4.0; Wastewater treatment plant; Industrial internet of things (IIoT); Predictive maintenance (Pdm)
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-025-11463-4

In industrial applications, maintenance strategies such as corrective, preventive, and predictive maintenance (Pdm) are essential for ensuring equipment reliability and availability. The Pdm which focuses on anticipating faults before they occur, has evolved from traditional threshold-based methods to advanced techniques powered by Industry 4.0 technologies. This paper presents an enhanced Pdm methodology by integrating the machine learning (ML) and the industrial internet of things (IIoT) to improve the fault prediction and minimize the unexpected failures. The proposed methodology utilizes real-time data collected by programmable logic controllers (PLCs) from field sensors which monitor key parameters such as temperature, vibration, and current. Historical data stored in a database server is processed using Node-Red for visualization and predictive analysis, enabling accurate failure predictions and remaining useful life estimation based on the health indicator value using regression analysis. The developed model generates automated alerts and reports via email and provides real-time insights into equipment health. These reports will encapsulate detailed information regarding the equipment healthiness and the estimated remaining useful life (RUL). To validate the effectiveness of this methodology, a case study was conducted on heavy-duty equipment in a wastewater treatment plant. The experimental results demonstrate precise failure point estimation with different regression methodologies and highlight the effectiveness of the proposed Pdm framework in improving industrial asset management, reducing downtime, and equipment reliability.