The primary standard for precision machining involves the production of parts that adhere to the necessary surface quality requirements for optimal tribological performance. The surface assessment in accordance with ISO 13115 (DIN 4776) is a crucial aspect of the methodologies used to assess and validate these requirements. This study seeks to merge a methodology that combines Taguchi’s \(L_{16}\) design of experiments (DOE), analysis of variance (ANOVA), artificial neural networks (ANN), and the multi-objective dragonfly algorithm (MODA) to ascertain the influence of cutting parameters ( \(V_{c}\) , \(f\) , \(ap\) , \(r_{\varepsilon }\) , and \(X_{r}\) ) on the bearing area curve parameters ( \(R_{pk}\) , \(R_{k}\) , \(R_{vk}\) , \(Mr1\) , and \(Mr2\) ). Turning operations are conducted using medium-carbon steel C45 and uncoated carbide inserts. Unlike previous studies that have treated BAC parameters, ANN, or RSM individually, the novelty of this work lies in the integration of Taguchi’s \(L_{16}\) with ANN modeling and MODA optimization. ANOVA results indicate that feed rate ( \(f\) ) and cutting speed ( \(V_{c}\) ) are typically the most significant variables. Moreover, the models developed account for a substantial portion of the variability observed in the experimental data, with \(R^{2}\) values exceeding 95% for all responses. Ultimately, the MODA algorithm identifies the optimal conditions as follows: maximum \(V_{c}\) (420 m/min), minimum \(f\) (0.031 mm/rev), maximum \(ap\) (0.4 mm), minimum \(r_{\varepsilon }\) (0.4 mm), and minimum \(X_{r}\) (45°). This shows that the simultaneous improvement of the five responses led to \(R_{pk}\) = 1.27 µm, \(R_{k}\) = 1.54 µm, \(R_{vk}\) = 1.09 µm, \(Mr1\) = 5.68%, and \(Mr2\) = 91.74%.
Graphical Abstract
