Design and Implementation of an IoT-Enabled Predictive Maintenance System for Smart Manufacturing
Year: 2026
Type: article
Abstract: The accelerated development of the Industry 4.0 has reinvented traditional manufacturing systems into incredibly connected, informationally guided smart production landscapes in which efficiency in operations and dependability of equipment are paramount. The Internet of Things (IoT) has brought predictive maintenance (PdM), which can remove the shortcomings of reactive and preventive maintenance methods by providing the opportunity to monitor and make decisions based on the collected data in real-time. In this research, the proposed design and implementation of an integrated IoT-based predictive maintenance architecture designed to be used in smart manufacturing usage is presented. The suggested system is built with a combination of multi-sensing data collection, edge computing to support low-latency preprocessing, cloud storage and analytics, and the development of advanced machine learning models to predict fault. Sensors on the industry constantly record the vibration, temperature and electrical readings, which are computed at the edge layer to minimise the use of bandwidth and guarantee a quick response to anomaly. Hybrid edge-cloud architecture enables model training that is scalable as well as system optimization. One of the models tested outperformed the rest with a prediction error of 94.6 and this shows that a Long Short-Term Memory (LSTM) network has the ability to capture a time-dependence in time-series industrial data. The benefits of experimental validation in a simulated environment of a smart factory have demonstrated a decrease in unplanned downtime by 35% and a decrease in the costs of maintenance by 28% as compared to the traditional approaches. The proposed architecture provides the security of data transmission, modular scalability, flexibility between heterogeneous industrial resources. In general, the paper establishes the robust and effective framework of implementing IoT, edge intelligence, and deep learning in improving reliability and productivity and low cost of smart manufacturing systems in the next generation.
Author: Robbi Rahim
Institution: Sekolah Tinggi Ilmu Manajemen Sukma
Language: English
Subfield: Industrial and Manufacturing Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7220119679