A new method to achieve dynamic heat input monitoring in robotic belt grinding of Inconel 718

Research question

How can heat entering a ground workpiece be tracked as belt wear and contact force change?

Method

Sound and force features are fused in a least-squares support-vector model tuned by Bayesian adaptive direct search. Heat-transfer calculations and infrared measurements provide validation of the predicted heat input.

Main findings

Experiments on robotic belt grinding of Inconel 718 tracked changing heat input within a grinding pass. A prediction took approximately 0.6 s on the reported computer, providing a monitoring method for subsequent thermal-control research.