International Journal of Advanced Research in Computer and Communication Engineering

Abstract: Industrial machines experience gradual degradation due to continuous operation, mechanical wear, and varying load conditions. Unplanned breakdowns result in production losses, increased maintenance costs, and reduced equipment lifespan. To address these challenges, this work presents an IoT-based predictive maintenance system that continuously monitors machine health parameters and performs real-time analysis using the ThingSpeak cloud platform. An ESP32 microcontroller is integrated with sensors such as an ADXL345 three-axis accelerometer for vibration measurement, a temperature sensor, and a current sensor to capture critical machine health indicators. The acquired data is transmitted to ThingSpeak through Wi-Fi, where MATLAB Analytics is used to extract features such as vibration RMS, spectral peak frequencies, temperature trends, and load variations. These features are further analyzed to detect anomalies, estimate machine degradation, and predict possible failure conditions. Threshold-based logic and machine learning algorithms are implemented on the cloud to classify machine states into healthy, warning, and fault categories. The system also triggers alerts using ThingHTTP and webhooks, enabling immediate maintenance actions. Experimental results show that the proposed solution provides accurate early-warning detection, reduces downtime, and offers a scalable, low-cost architecture suitable for industrial automation environments. The research demonstrates that IoT-enabled predictive maintenance significantly improves reliability, enhances operational efficiency, and supports data-driven decision-making in industrial machine monitoring.

Keywords: Predictive Maintenance, Industrial Machines, IoT-Based Monitoring, ThingSpeak Cloud, ESP32 Microcontroller, ADXL345 Accelerometer, Vibration Analysis, MATLAB Analytics;

Downloads: | DOI: 10.17148/IJARCCE.2025.141174