A multiobjective optimization model for machining quality in the AISI 12L14 steel turning process using fuzzy multivariate mean square error

• Published in: Precision engineering 2019-03, Vol.56, p.303-320
• Main Authors: Daroz Gaudêncio, Juliana Helena, de Almeida, Fabrício Alves, Turrioni, João Batista, da Costa Quinino, Roberto, Balestrassi, Pedro Paulo, de Paiva, Anderson Paulo
• Format: Article
• Language: English
• Subjects: Fuzzy decision maker; Normal boundary intersection; Principal components analysis; Robust parameter design
• ISSN: 0141-6359; 1873-2372
• DOI: 10.1016/j.precisioneng.2019.01.001

Organizations focus on determining optimal operating conditions to ensure quality; however, industrial processes exhibit a high degree of variability and the use of robust estimators is a suitable alternative to model experimental data. As a case study, the surface roughness (Ra) of an AISI 12L14 steel turning process is optimized to find a centrality measure close to its target with minimum dispersion and thus improve the quality of the machined surface by choosing the best values of the associated parameters. The main contribution of this research is the proposal of a multiobjective optimization method that uses principal components analysis to minimize the redundancy of objective functions in terms of multivariate mean square error, thus making optimization of the process possible with a better explanation of all centrality and dispersion estimators proposed herein. The method uses a fuzzy decision maker to show the surface roughness' optimum result with the most efficient production taken into consideration. To prove its efficiency, confirmation runs were conducted. At a confidence level of 95%, the optimal value falls within the multivariate confidence intervals only for Model B, in which the estimators' median and median absolute deviation are considered, thus affirming which pair of estimators achieves the most robust parameter design solution. Through the proposed research, the developed model can be used in industries for determining machining parameters to attain high quality with minimum power consumption and hence maximum productivity. •Process optimization can ensure product quality for organizations focusing on determining optimal operating conditions.•Robust estimators are a suitable alternative for modeling experimental data.•Process optimization can offer a better explanation of all proposed estimators.•Machining parameters can be chosen to attain quality with maximum productivity.