A comparative study of JKR and DMT contact models for the DEM simulation of powder spreading in additive manufacturing

Powder spreading is the fundamental and most important process of powder bed fusion additive manufacturing. Powder particles experience cohesive forces due to their micron size, and these forces influence the quality of the layer. The dynamics of powder spreading is simulated using the discrete element method (DEM). DEM contact models with non-cohesive and cohesive interactions were used in past studies. This work compares two predominant cohesion contact models, the Johnson-Kendall-Roberts (JKR) and Derjaguin-Muller-Toporov (DMT). The influence of cohesion parameters and particle size on the spread layer quality is analysed. Additionally, mesoscopic analysis is carried out to gain insights into the behaviour of the spreading mechanism. The Tabor parameter (lambda T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda _{\textrm{T}}$$\end{document}) that establishes the suitability of a specific cohesion model is investigated in the context of powder spreading process. Both models predict similar packing fractions at lower lambda T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda _{\textrm{T}}$$\end{document}, whereas, at higher values of the lambda T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda _{\textrm{T}}$$\end{document}, the contact forces of the JKR and DMT models diverge, leading to differences in packing fractions and local particle configurations in the spread layer. The findings demonstrate that the JKR model is applicable across the entire range of Tabor parameter.