Direct energy deposition metamodeling using a meshless method. Issue 3 (13th July 2020)

• Record Type: Journal Article
• Authors: Abbes, Boussad; Anedaf, Tahar; Abbes, Fazilay; Li, Yuming
• Is Part Of: Engineering computations. Volume 38:Issue 3(2021)
• Page Start: 1226
• Page End: 1240
• Publication Date: 2020-07-13
• Subjects: Additive manufacturing -- Titanium alloy -- Meshless method -- Direct energy deposition -- Finite point method -- Surrogate model; Computer-aided engineering -- Periodicals; Computer graphics -- Periodicals; 620.00285
• DOI: 10.1108/EC-10-2019-0447
• Languages: English
• ISSNs: 0264-4401

Abstract

Abstract : Purpose: Direct energy deposition (DED) is an additive manufacturing process that allows to produce metal parts with complex shapes. DED process depends on several parameters, including laser power, deposition rate and powder feeding rate. It is important to control the manufacturing process to study the influence of the operating parameters on the final characteristics of these parts and to optimize them. Computational modeling helps engineers to address these challenges. This paper aims to establish a framework for the development, verification and application of meshless methods and surrogate models to the DED process. Design/methodology/approach: Finite pointset method (FPM) is used to solve conservation equations involved in the DED process. A surrogate model is then established for the DED process using design of experiments with powder feeding rate, laser power and scanning speed as input parameters. The surrogate model is constructed using neutral networks (NN) approximations for the prediction of maximum temperature, clad angle and dilution. Findings: The simulations of thin wall built of Ti-6Al-4V titanium alloy clearly demonstrated that FPM simulation is successful in predicting temperature distribution for different process conditions and compare favorably with experimental results from the literature. A methodology has been developed for obtaining a surrogate model for DED process. Originality/value: This methodology shows how to achieve realistic simulations of DED process and how to construct a surrogate model for further use in optimization loop.