Fault diagnosis of multi-point cutting tool while machining of MMCs through vibration signal using decision tree algorithm technique

Applications of machine learning algorithms are increasing due to advances in an automated machining environment for improving manufacturing system fault diagnosis. Therefore, monitoring of tool conditions is becomes an imperative. A faulty tool generates during the machining, very rough surface texture, inaccurate dimensions and geometry and leads to non-economical machining. Thus, it is vital to monitor the condition of cutting tool during the metal cutting operation in order to control over superior quality and economic aspects of manufacturing process. In this research work, experiments were carried out to categorize according to the tool conditions into healthy, blunt and loosely blunt during machining of MMCs at high speed on milling machine. Continuously, while performing experiments, vibrational signals were noted with help of accelerometer, with and without worn-out/blunt cutting tool for fault diagnosis using machine learning techniques for evaluating tool condition on real time. The statistical data were determined from the vibration signal and a set of prominent features of that data were chosen as an input to the algorithm in order to find the fault of the tool. The results obtained in this paper will be helpful to observe the tool condition in live time.