<![CDATA[This research aims to develop an Internet of Things (IoT)-based predictive maintenance system for AC electric motors used in water treatment plants. The primary objective is to reduce unplanned downtime and enhance operational reliability by enabling proactive scheduling of maintenance activities. Research design adopts a research and development approach, beginning with a preliminary study, followed by system design, prototype implementation, data acquisition, and performance validation. The system integrates vibration, temperature, and rotation sensors with an Arduino/ESP32 microcontroller for real-time data collection. Data is transmitted via MQTT protocol to a cloud platform for storage and analysis. Machine learning algorithms, including Random Forest and Long Short-Term Memory (LSTM), are applied to classify equipment condition and detect anomalies. To address the limitation of failure data, Generative Adversarial Networks (GANs) are employed to generate synthetic training data, improving model robustness. Experimental results show that vibration levels reached 3.9 mm/s, temperature rose to 95 °C, and motor speed dropped to 1420 RPM, all of which signaled potential failure before actual breakdown. The LSTM model achieved an F1-score of 0.92, which increased to 0.95 when combined with GAN-based data augmentation, outperforming Random Forest. In conclusion, the proposed system demonstrates that integrating IoT with multi-sensor data and advanced machine learning enables early fault detection in AC motors. This approach offers a cost-effective and scalable solution for predictive maintenance, reducing downtime and extending equipment lifespan in water treatment operations.]]>
Copyrights © 2025
