Effect of temperature and substrate geometry on single aluminium weld bead geometry deposited by Wire Arc Additive Manufacturing: Proposition of an experimental procedure

Limousin, Maxime; Manokruang, Supasit; Vignat, Frédéric; Museau, Matthieu; Grandvallet, Christelle; Béraud, Nicolas

doi:10.1016/j.cirpj.2023.06.010

Cited by 4 publications

(1 citation statement)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dellarre [18] shows that the image post-processing that produces the thermal metric is a key step in improving measurement accuracy. Moreover, WAAM can easily produce thin walls, and managing the thickness of the manufactured bead is a key challenge to produce nearnet shape parts [19]. This article focuses on thin wall manufacturing and thickness as a quality indicator.…”

Section: Introductionmentioning

confidence: 99%

A method to qualify image post-processing for thin wall thickness prediction from NIR camera image of aluminum WAAM process

Béraud,

Lombard,

Dellarre

et al. 2024

Preprint

View full text Add to dashboard Cite

Wire Arc Additive Manufacturing (WAAM) is a metal arc welding additive process that allows the production of large parts. Managing the quality of produced parts is a key challenge to the adoption of this technology in the industry. A particular case of this is the production of aluminum thin walls and the management of their thicknesses. Literature supports that the use of a near-infrared camera is a good way to monitor the meltpool shape to predict the manufactured thickness. Nevertheless, the post-processing applied to the image has a big influence on the accuracy of the measure. That is why, this article proposes a method to qualify image post-processing for thin wall thickness prediction from near-infrared camera images of aluminum WAAM process. A dataset from the literature is used to evaluate the accuracy of different post-processes. First, a method is proposed to evaluate the accuracy of a given post-process. Secondly, two types of post-processing are presented: post-processing based on conventional image processing and post-processing based on neural networks. Their accuracy is evaluated with the proposed method. The article concludes that the proposed method is adapted to qualify post-processes and that the proposed post-processing based on neural networks gives significant results in terms of accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%