Data-Driven Approach for OEE Enhancement in the Manufacturing Industry

Shravani Shalgar , Tanmay Landge , Prerana Jagtap

IoT, Predictive Maintenance, Industrial Automation, Machine Learning.

This paper proposes a predictive maintenance system designed for industrial machinery, integrating IoT technology, cloud computing, and machine learning. The system uses IoT-enabled ESP32 microcontrollers to collect real-time operational data such as positional deviations, production counts, and cycle time. Machines without built-in IoT capabilities are retrofitted with sensors to ensure uniform monitoring. Collected data is transmitted securely to an OpenStack-based cloud infrastructure via an API gateway, enabling scalable data storage and processing. The preprocessing pipeline cleans and normalizes the data, ensuring it is suitable for advanced predictive analytics. The Auto-Regressive Integrated Moving Average (ARIMA) model is employed for time-series forecasting to detect anomalies and predict machine failures. These predictions, along with real-time machine metrics are visualized on a web-based dashboard. Operators can monitor key performance indicators (KPIs) such as production, quality, and machine health, facilitating informed decision-making and minimizing unplanned downtime. Feedback loops enhance the model's accuracy and system performance over time. The proposed architecture demonstrates scalability, robustness, and effectiveness in reducing operational inefficiencies and increasing equipment lifespan. By combining IoT and cloud-based predictive analytics, the system contributes to Industry 4.0's vision of smart, interconnected factories. Experimental results confirm the feasibility and practicality of the system in industrial applications

Volume 10 Issue 1, January-2025

Pages : a311-a316