Application of a Robust Decision-Making Rule for Comprehensive Assessment of Laser Cutting Conditions and Performance

Abstract

Laser cutting parameters synergistically affect, although in different quantitative and qualitative manners, multiple process performances, such as the resulting cut quality characteristics, material removal rate, cutting time, and costs, and the determination of the most appropriate laser cutting conditions for a given application is of prime importance. Given the existence of multiple mutually opposite performances, assessment and laser cutting conditions and performance can be considered a multiple-criteria decision-making (MCDM) problem. In order to overcome the possible inconsistency of rankings determined by different MCDM methods while solving the same decision-making problem, the present study promotes a novel methodology for the assessment and selection of laser cutting conditions by developing a robust decision-making rule (RDMR) that combines different decision-making rules from six MCDM methods and Taguchi’s principles of robust design. In order to illustrate the application of the proposed methodology, CO2 laser cutting in a stainless-steel experiment, based on the use of the Box–Behnken design, was conducted. On the basis of the experimental results, a comprehensive laser cutting MCDM model was developed with seven criteria related to cut quality (i.e., kerf geometry and cut surface), productivity, variable costs, and environmental aspects. It was observed that there was no laser cutting condition that could be considered as the best regime with respect to the different laser cutting process performances. Kendall’s and Spearman’s rank correlation coefficients indicated a certain level of disagreement among the resulting rankings of the laser cutting conditions produced by the considered MCDM methods, whereas the application of the proposed RDMR ensured the highest level of ranking consistency. Some possibilities for modeling of RDMR and its further use for the assessment of arbitrarily chosen laser cutting conditions and the use of the derived model to perform sensitivity analysis for determining the most influential laser cutting parameters are also discussed and addressed. It was observed that laser cutting parameters in different laser cutting conditions may have a variable effect on the resulting overall process performances. The comparison of the obtained results and the results determined by classical desirability-based multi-objective optimization revealed that there exists substantial agreement between the most preferable and least preferable laser cutting conditions, thus justifying the applied methodology.