Statistical and fuzzy-logic approaches in on-line surface roughness recognition systems for end-milling operations

摘要

This paper describes new approaches for on-line surface-roughness recognition (OSRR) systems using vibration signals and cutting conditions for predicting surface roughness (Ra) while end milling is taking place. The parameters considered in thispaper are spindle speed, feed rate, depth of cut, and the cutting vibration between tool and workpiece, which are measured by an accelerometer. The analyses of the data and the building model are carried out using a multiple-regression analysis and fuzzylogic system. Experimental results show that feed rate is the most significant independent variable to predict Ra (surface roughness), and that vibration data contribute to increase R{sup}2 and improve the prediction ability in the multiple-regressionmodel. Surface roughness can also be predicted with 91% accuracy by OSRR using multiple-regression analysis and with 95% accuracy by using fuzzy logic system.