Quantitative methods for inferring process-structure-property relations in machining

Abstract

Metal machining subjects material to large strains (e > 1) at extremely high rates (rate > 1000 s-1) while simultaneously imposing heating due to dissipated plastic work. This complex thermomechanical loading drives microstructure evolution and consequently property evolution. Oxygen Free High Conductivity Copper (OFHC Cu) is studied as an example FCC system. Collected chips from machining experiments, subjected to differing process conditions, are imaged to reveal the underlying microstructure. The structure of these machined chips are quantified using robust descriptors using both scanning electron microscopy and orientation imaging microscopy. Chip properties are inferred from a combination of spherical nanoindentation experiments and a novel Bayesian nanoindentation inverse solution. The process-structure-property linkages are established for OFHC Cu using a coupled nonparametric regression model and a robust method for inverting the process-structure-property linkages is developed and illustrated. Furthermore, a process-property model calibration methodology is presented for the establishment of machining constitutive model parameters directly from machining forces.