Multi Sensing on Bearing Faults Detection with Internet of Things (IoT) based

Bearing is a machine part that has a function to keep the shaft always rotating or moving linearly to the axis of the shaft and its path. Bearings are often found in automotive equipment and home appliances, one of them is the bearing that has found in a single-phase induction motor (water pump). But, until now the largest percentage of induction motor faults occurs in bearings. Therefore, an accurate system of bearing faults detection is the key to protecting an induction motor from such any faults. In this study, we proposed bearing faults detection on a single-phase induction motor with water loads and based on Internet of Things (IoT). This system used multi-sensors, i.e. a temperature sensor, a current sensor, and a vibration sensor. Some processes in this bearing faults detection system are feature extraction process using Empirical Mode Decomposition (EMD) and pattern recognition process using Backpropagation Neural Network (BNN). Then the results from pattern recognition is displayed through the Internet of Things (IoT) system. The results of this project show that EMD can decompose the vibration signal and BNN is able to classify signals with 100% accuracy of current signals and 98% for vibration signals.