Title: Multi-objective optimisation of DMLS process parameters using evolutionary algorithm

Authors: C.D. Naiju; M. Adithan; P. Radhakrishnan

Addresses: School of Mechanical and Building Sciences, VIT University, Vellore 14, Tamil Nadu, India. ' Academic Staff College, VIT University, Vellore 14, Tamil Nadu, India. ' Institute of Advanced Studies, PSG Tech, Coimbatore, Tamil Nadu, India

Abstract: A multi-objective optimisation study is carried out for direct metal laser sintering process (DMLS). In order to optimise the individual process parameters and mechanical properties, analysis of variance (ANOVA) methods were used. Experiments were conducted by varying the process parameters using Taguchi's modified L8 orthogonal array method. Mechanical properties such as fatigue cycles to failure, tensile strength, compressive strength, hardness and wear rate of components have been ascertained and a mathematical model has been developed using multi-variable linear regression analysis. Genetic algorithm (GA) was used to carry out final multi-objective optimisation studies. Optimal process parameters for better mechanical properties were observed by initiating multi-objective optimisation using genetic algorithm. By selecting better process parameters, components could be manufactured in the near future with improved mechanical properties. Results obtained will be useful to manufacturing components for functional application so that DMLS technology can be brought to the mainstream manufacturing process.

Keywords: rapid prototyping; direct metal laser sintering; DMLS; optimisation techniques; Taguchi methods; algorithms; analysis of variance; ANOVA; process parameters; orthogonal arrays; mathematical modelling; genetic algorithms; mechanical properties; fatigue cycles; tensile strength; compressive strength; hardness; wear rate; rapid manufacturing.

DOI: 10.1504/IJRAPIDM.2012.046572

International Journal of Rapid Manufacturing, 2012 Vol.3 No.1, pp.23 - 35

Published online: 20 Dec 2014 *

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