Statistical Analysis of Variations in Model Predicted Cutting Forces for End Milling

This paper quantifies statistical variations in model predicted machining forces while cutting aluminum, carbon steel, stainless steel and titanium. An accurate estimate of the variability is essential for use in process planning to determine appropriate factors of safety when setting cutting conditions that are both safe and efficient. A linear regression is performed to estimate the coefficients of a tangential cutting force model. Density ellipses are used to define the confidence limits of the coefficients under varying spindle speeds and radial immersions. The locus of coefficients at the 95% confidence level is then used in a mechanistic force model to quantify the variability in the cutting force predictions. Forces calculated by the model were within 20% of the nominal predicted values for cases for aluminum, steel and stainless steel. The results were much worse for titanium because of a smaller sample size. Experimentally measured forces were within the error bounds predicted by the simulation.Copyright © 2009 by ASME