A study on multi-objective optimization in non-circular helical pocket milling of Al 6061-T6 using integrated approach response surface method (RSM) with genetic algorithm (GA)

• Research novelty is a non-circular helical pocket milling. • Used RSM to develop a model for each response variable. • ANOVA showed the most influential input parameter in machining Al 6061-T6. • K-means grouped the RSM-MOGA optimal solutions, and TOPSIS chose the best solution. • Confirmatory experiments verified the accuracy of RSM-MOGA, especially for Al 6061-T6. Helical pocket milling is an advanced CNC machining technique that achieves high accuracy, improved surface finish, extended tool life, and reduced burr formation. The present research investigates the influence of machining parameters on response variables and aims to balance conflicting responses in non-circular helical pocket milling of aluminum alloy 6061-T6 under dry-cutting conditions, as well as to group Pareto-optimal solutions into clusters. The key input parameters are spindle speed, feed rate, and helix pitch, while the quality variables include cutting zone temperature, material removal rate, dimensional error, and surface roughness. These were measured using an IR laser digital thermometer (Nicety ST380A), machining time from a Siemens CNC controller panel, a coordinate measuring machine (Mitutoyo Crysta-Plus M443), and a surface roughness tester (SURFOCOM 2900SD3), respectively. Twenty experimental runs were designed using Central Composite Design (CCD) based on the Response Surface Method (RSM). To address trade-offs among conflicting responses, RSM was integrated with a Multi-Objective Genetic Algorithm (MOGA). Analysis of Variance (ANOVA) revealed that helix pitch was the most significant factor, contributing 39.96 %, followed by feed rate (22.30 %) and spindle speed (13.74 %). The optimum parameters were obtained as s = 3393 rpm, f = 703.59 mm/min, and p = 0.2233 mm, based on the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). Pareto-optimal solutions were grouped into five clusters using K-means clustering, enabling machinists and programmers to select parameter sets according to specific requirements. The confirmatory experiments produced results close to the predictions, with percentage errors below 6 %. Future work should examine the influence of arc feed, defined as the feed rate during curved tool paths, the use of coated carbide tools (TiN, TiAlN, AlTiN), and machining of superalloys (Inconel 718, Pyromet 860, Ultimet) to broaden the applicability of non-circular helical pocket milling.