Malfunction Detection on Production Line Using Machine Learning: Case Study in Wood Industry

The concept of the Internet of Things, especially in the last decade, has created the opportunity to place sensors in every event and location that can be tracked to collect data via these sensors. Collecting data from sensors is not a stand-alone solution. After the problem of data collection and storage of large amounts of data collected has been overcome, it has been made easier by performing analytical operations with this data. The machine learning algorithms and methods used in the robotics sector are used in different fields to make productions, process and machine groupings by making various estimations for the industry with more complex algorithms or clustering operations with the collected data. Within the scope of this project, it is aimed to monitor the condition of the machines on the production line with the data collected from the machines used in the production process and to make fault detection on the machines by using the machine learning methods for the maintenance and repairs of the machines before they break down, produce faulty products and stop the production line. In this study, anomaly detection methods which are proposed in the literature were performed to data which was collected by sensors. Also, the artificial neural network was applied to the dataset. The results show us these technics can be used in the manufacturing sector for fault detection.