Enhancing Industrial Efficiency: Early Detection of Fan Failures with IoT Technology

The Internet of Things is an innovative system that enables the integration of technological devices into daily life, enabling them to exchange data with each other and with other systems with the help of machine learning. Fan systems play a critical role in industrial facilities. They support the efficient operation of facilities by providing functions such as ventilation and cooling. Failures in these systems can cause operational interruptions and cost increases. By quickly detecting faults, production can be sustained and costs can be reduced. In this study, acceleration values collected from the fan motor with Arduino Nano33 BLE Sense device were analyzed and classification was performed with Decision Tree, Gradient Boosting Machine and Convolutional Neural Networks. According to the results obtained, the best classification accuracy was obtained in the Decision Tree and Gradient Boosting Machine methods. The three methods were also compared in terms of latency and size. In terms of latency, the best performance belongs to the Decision Tree and Gradient Boosting Machine methods. In terms of size, the Decision Tree method has the smallest size.