Tool wear prediction in end milling of Ti-6Al-4V through Kalman filter based fusion of texture features and cutting forces

The current trend of Industry 4.0 development requires an integration of the process sensing technologies with the cloud computing network to design the cyber-physical systems which can seamlessly transfer data between the connected devices. This will pave a way for real-time process monitoring as well as the control of the process by the use of decision-making algorithms. The current paper explores the opportunities with this regard in the area of tool wear prediction in end milling of Ti-6A1-4V alloy at various operating conditions. A series of slot milling passes were made at various parameter combinations of feed, speed, and depth of cut until the flank wear on tool crosses the failure criterion. The cutting force data acquired in the process with the dynamometer and the texture features from the image of the milled surface are used to build a model for predicting the flank wear through the Kalman filter approach. The fusion model built using the Kalman filter methodology achieves a good accuracy in predicting the flank wear on the tool. The model is highly accurate in predicting the wear as the tool approaches the failure threshold. Thus the model can enable the decision control module to trigger a tool change signal and improve the overall productivity of the process. (C) 2018 The Authors. Published by Elsevier B.V.