Abstract
Internet of Things, in this technological era contributes for many applications particularly when we deal this technology in industry, we name it as Industrial Internet of Things (IIOT). In this paper one such application of Industrial Internet of Things, (i.e. by using Internet of things method) for predicting the failure of the machine with the audio signal generated from the machine. Predictive maintenance is a type of maintenance of the machine where the failure of the machine is predicted well in advance, so that the consecutive occurrence of the failure is eradicated by introducing appropriate maintenance measure. Generally, the traditional way of predictive maintenance is based on the statistical tool that is employed so far in many industries. Now a day the recent development of IOT applied for predictive maintenance of the machine can replace the traditional way in orders to increase the life time or operating hours of the machine in production line in a better way. The novelty of this study is by implementing the sensors to trap the audio signal of the machine operation. Now this signal is further fed to edge impulse software for anomaly detection using k means clustering using Edge impulse software. This way of predictive maintenance using IOT would be in line with real time data. The analysis could be further expanded to identify and analyses the type of machine error based on the application of machine learning algorithm.
Keywords: Audio Signal, Edge Impulse, Industrial Internet of Things, k means Anomaly Detection, Predictive Maintenance.