Title: Multivariate analysis of AISI-52100 steel machining: a combined finite element artificial intelligence approach

Authors: Anastasios Tzotzis; Nikolaos Efkolidis; César García-Hernández; Panagiotis Kyratsis

Addresses: Department of Product and Systems Design Engineering, University of Western Macedonia, Kila Kozani, 50100, Greece ' Department of Product and Systems Design Engineering, University of Western Macedonia, Kila Kozani, 50100, Greece ' Department of Design and Manufacturing Engineering, University of Zaragoza, Zaragoza, 50018, Spain ' Department of Product and Systems Design Engineering, University of Western Macedonia, Kila Kozani, 50100, Greece

Abstract: The present study focuses on the analysis of AISI-52100 steel hard-turning with standardised square inserts. The process is being studied in terms of the resultant cutting force and the cutting power under a wide range of four key machining parameters: the cutting speed, the feed rate, the depth of cut and the tool nose radius. First of all, an updated finite element method (FEM) model has been used to generate a dataset, which in turn was used to train an artificial neural network (ANN), minimising this way the required experimental work and the utilisation of high amounts of computing resources. The developed networks were evaluated with regard to their reliability, revealing increased levels of accuracy. The mean absolute percentage error (MAPE) was calculated equal to 8.1% for the force prediction network and 11.2% for the power prediction network, respectively. Furthermore, the multivariate interaction was evaluated and visualised.

Keywords: AISI-52100 turning; cutting forces; cutting power; 3D FEM; ANN; artificial neural network; DEFORM-3D; artificial intelligence.

DOI: 10.1504/IJMMS.2024.143012

International Journal of Mechatronics and Manufacturing Systems, 2024 Vol.17 No.2, pp.99 - 116

Received: 03 Jan 2024
Accepted: 23 Mar 2024

Published online: 02 Dec 2024 *

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