A Numerical Study of Uncertainty in Stability and Surface Location Error in High-Speed Milling

Abstract

High-speed milling offers an efficient tool for developing cost effective manufacturing processes with acceptable dimensional accuracy. Realization of these benefits depends on an appropriate selection of preferred operating conditions. In a previous study, optimization was used to find these conditions for two objectives: material removal rate (MRR) and surface location error (SLE), with a Pareto front or tradeoff curve found for the two competing objectives. However, confidence in the optimization results depends on the uncertainty in the input parameters to the milling model (time finite element analysis was applied here for simultaneous prediction of stability and surface location error). In this paper the uncertainty of these input parameters such as cutting force coefficients, tool modal parameters, and cutting parameters is evaluated. The sensitivity of the maximum stable axial depth, blim, to each input parameter at each spindle speed is determined. This enables identification of parameters with high contribution to stability lobe uncertainty. Two methods are used to calculate uncertainty: 1) Monte Carlo simulation; and 2) numerical derivatives of the system eigenvalues. Once the uncertainty in axial depth is calculated, its effect is observed in the MRR and SLE uncertainties. This allows robust optimization that takes into consideration both performance and uncertainty.