Title: Productivity improvement in hard turning of AISI 4340 with response surface methodology and machine learning

Authors: A. Ginting; R.A. Sidabutar; F.D. Pranata; K. Syam; S. Situmorang; T. Fikriawan

Addresses: Laboratory of Machining Processes, Department of Mechanical Engineering, Faculty of Engineering, Universitas Sumatera Utara, Jalan Almamater, Kampus USU, Medan 20155, Indonesia ' Department of Mechanical Engineering, Faculty of Engineering, Universitas Darma Agung, Jalan Dr. T.D. Pardede No. 21, Medan 20153, Indonesia ' Laboratory of Machining Processes, Department of Mechanical Engineering, Faculty of Engineering, Universitas Sumatera Utara, Jalan Almamater, Kampus USU, Medan 20155, Indonesia ' Laboratory of Machining Processes, Department of Mechanical Engineering, Faculty of Engineering, Universitas Sumatera Utara, Jalan Almamater, Kampus USU, Medan 20155, Indonesia ' Laboratory of Machining Processes, Department of Mechanical Engineering, Faculty of Engineering, Universitas Sumatera Utara, Jalan Almamater, Kampus USU, Medan 20155, Indonesia ' Laboratory of Machining Processes, Department of Mechanical Engineering, Faculty of Engineering, Universitas Sumatera Utara, Jalan Almamater, Kampus USU, Medan 20155, Indonesia

Abstract: This study aims to optimise the hard turning of AISI 4340 steel to improve productivity using response surface methodology (RSM) and machine learning (ML) techniques. The novelty lies in integrating these methods to enhance material removal rate (MRR) while maintaining surface roughness (Ra) quality. Experiments were conducted with an uncoated carbide tool under dry and minimum quantity lubrication (MQL) conditions, varying cutting speed, feed, and depth of cut. RSM identified feed as the most significant factor affecting Ra, while ML, specifically linear regression (LR), predicted optimal cutting conditions. Key findings include achieving an optimum MRR of 5.2 cm³/min under dry and 7.2 cm³/min under MQL conditions, with Ra within the acceptable range (1.6 μm-3.2 μm). Validation confirmed the model's accuracy, demonstrating high agreement between predicted and experimental Ra values. This integrated approach offers a robust solution for optimising hard-turning processes in industrial applications.

Keywords: surface roughness; Ra; material removal rate; MRR; linear regression; LR; minimum quantity lubrication; MQL.

DOI: 10.1504/IJMMM.2025.145073

International Journal of Machining and Machinability of Materials, 2025 Vol.27 No.1, pp.85 - 103

Received: 21 Apr 2024
Accepted: 04 Aug 2024
Published online: 18 Mar 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article