Abstract
This research explores a new methodology to optimize a multivariate dry end milling process of the AISI 1045 steel. Once the use of cutting fluids in machining processes has been questioned, dry milling techniques are considered to be a way to the cleaner manufacturing in the context of the sustainable production. Four input parameters and six response variables were considered. The normal boundary intersection (NBI) is a multi-objective optimization method mainly developed to compensate the short comings attributed to the method of weighted sums. However, the NBI method tends to fail producing unreal results and non-convex frontiers if the multiple objective functions are correlated and with conflicting objectives. To deal with this constraint, this work presents a new multi-objective hybrid approach, called NBI-MMSE, that combine the NBI with multivariate mean square error (MMSE) functions. This approach utilizes a procedure that integrates the principal component analysis with the response surface methodology for problems with correlated multiple responses. Theoretical and experimental results indicate that the solution found by NBI-MMSE approach was characterized as a more appropriate optimal point in relation to one obtained with the traditional weighted sum. In this case, the process parameters optimization for end milling process without cutting fluids was able to achieve, at the same time, the maximum rate of removed material and minimum surface roughness, confirming the adequacy of the work's proposal. (C) 2016 Elsevier Ltd. All rights reserved.