Methodology for soft-sensor design and in-process surface conditioning in turning of aluminum alloys

Abstract

Surface properties significantly influence the performance of machined parts. However, they cannot be measured directly during machining. For surface conditioning based on a soft sensor, it is necessary to monitor process characteristics like temperatures and forces, which influence the surface state. Soft-sensor development in machining requires a robust methodology, which is adaptable to different materials and machining processes. In addition, a measurement system that combines hardware sensors to measure temperatures and process forces has to be implemented into the machine tool. In the present study, a suitable methodology is proposed and tested using a tool-workpiece thermocouple and a dynamometer to determine the thermomechanical workpiece load during turning of the aluminum alloys EN AW-2017 and EN AW-7075. Experimental investigations are performed according to a D-optimal statistical design of experiments. For this, the machining parameters cutting speed, feed, depth of cut, as well as the flank wear land width are varied on four levels. Subsequent measurements of residual stresses and the surface roughness are used to correlate the surface state with input parameters and their resulting thermomechanical workpiece load by multiple regression based on analysis of variance (ANOVA). It is found that the methodology is applicable and allows for the prediction of surface states. The developed soft sensors enable an in-process control of machining parameters, which enables a robust prediction and targeted conditioning of the addressed surface properties during machining.