ESTIMATION OF CUTTING FORCE MODEL COEFFICIENTS TO TRACK WEAR IN MILLING USING BAYESIAN ANALYSIS

摘要

Our previous research has demonstrated the feasibility of monitoring tool wear during milling by continuously updating the coefficients of a cutting force model. The method requires a robust method for on-line model coefficient estimation. Estimation using Least Square Regression (LSR) is easily implemented but requires that the data points come from different cutting conditions which is not always possible. In this paper, a method for coefficient estimation and wear tracking based on Bayesian updating is described. The Bayesian method has the advantage that it can be used when the cutting conditions are constant.