Machine Learning Framework for Tool Condition Monitoring in Milling

In the smart manufacturing era, the dynamics of monitoring and maintenance of the machines are changed. After the 4th industrial revolution, artificial intelligence and machine learning techniques are proven to be beneficial for carrying out predictive maintenance of machines. Internet of Things (IoT) along with the Cyber-Physical Systems (CPS) has made it possible to conduct a data-driven prognosis of a system. Predictive maintenance techniques have been developed in order to monitor an in-service machine for estimating when maintenance should be performed. Big data analysis and machine learning techniques enable the detection of the current health state and the remaining useful life of the equipment. Above mentioned developments have played an efficient role in increasing production efficiency by minimizing downtime during the manufacturing processes. This study describes the application of a tool condition monitoring (TCM) framework to a real-time milling data-set, with the aim of classifying the tool condition (worn, unworn) of the milling tool during a running process. The application of the framework can help in optimizing the maintenance operations and preventing the breakdown of the equipment.