Applying the response surface methodology, first-order empirical models for face milling of low carbon steel were graduated. The models relate some response measures such as surface roughness, power consumption and tool vibration to the cutting conditions, i.e.cutting speed , feed rate and depth of cut. The adequacy of the graduated models was tested using analysis of variance. The models were used to develop constant-response curves and optimization curves. The optimization curves are based on minimization of a cost function describing power consumption and cutting time of the process. The constant-response and optimization curves demonstrate how performance can be impro ved with proper selection of cutting conditions.
IntroductionMost organizations use experience to acquire their machinability data, which may be quite satisfactory for conventional machining technology. However, as more emphasis is placed on automation, it becomes necessary to develop models and data bases that more accurately describe the performance of machining processes to ensure optimum production from the expensive equipment involved.The milling process is important for producing flat as well as curved machined surfaces. Many attempts have been made to model and establish machinability data for the different milling operations. A non-linear regression model relating surface roughness to the cutting conditions has been developed in an experimental investigation (Wong and Middleton 1984). In another experimental work, tool-life tests were carried out for face milling using response surface methodology, and in this way general guidelines for assigning cutting conditions were established (Pandey and Mehta 1984). Experimentally based research has been carried out to determine the optimum working conditions for milling cast iron and to establish the influence of cutting conditions on specific power (Sletkov et al. 1985). It was shown that specific power decreases rapidly when the chip thickness increases from 0·02 to 0·2 mm.Diei and Dornfield (1987 a) carried out a study on the nature of the acoustic emission signal generated during face milling of steel. They found good correlation between the signal and the severity of the contact conditions during chip formation and at tool entry and exit. The same authors applied acoustic emission to the on-line sensing of tool wear in face milling (Diei and Dornfield 1987 b). It was found that both acoustic emission and cutting forces have parameters that correlate closely with flank wear.In the present study a systematic approach was used to model and establish machinability data for face milling operations. The approach described herein should enable optimum results to be achieved more quickly than previously.
