Scholarlink

<p>Heat accumulation is a critical factor in additive manufacturing (AM) of metals, as it directly impacts quality metrics such as dimensional accuracy, microstructure, and surface roughness. Identifying potential heat accumulation zones before printing is essential to control and mitigate undesirable defects. In Laser Powder Bed Fusion (LPBF) technology, the heat applied by the laser to melt the powder layers dissipates primarily through the part to the printing base. Any interruption in this heat dissipation process can lead to the formation of heat accumulation zones. Geometric factors such as overhangs, printing area, and part height significantly influence heat accumulation. While current approaches, particularly those based on finite element methods (FEM), provide accurate assessments of heat accumulation, they are computationally intensive and often unsuitable for iterative design processes. This research introduces a graph neural network-based tool trained on FEM simulation data, designed to identify potential heat accumulation zones in LPBF parts using only geometric information. The tool achieves an average computational speedup of 4.44<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\times \)</EquationSource> </InlineEquation> compared to FEM models, while maintaining a mean zone detection accuracy of 65<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\%\)</EquationSource> </InlineEquation> across the test specimens. By generating heat accumulation maps and identifying the primary affected zones, the tool enables fast and reliable thermal assessment during the earliest stages of the design process. This accelerates decision-making and supports rapid design iterations before any printing procedure begins.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A fast prediction model for detection heat accumulation zones in laser powder bed fusion printed parts.

  • 期刊论文

  • 发表日期：

    2026年1月21日

查看全文
  • Jesús García-Moreno Caraballo,
  • Sergio Corbera,
  • Rafael Barea,
  • Roberto Alvarez

摘要

Heat accumulation is a critical factor in additive manufacturing (AM) of metals, as it directly impacts quality metrics such as dimensional accuracy, microstructure, and surface roughness. Identifying potential heat accumulation zones before printing is essential to control and mitigate undesirable defects. In Laser Powder Bed Fusion (LPBF) technology, the heat applied by the laser to melt the powder layers dissipates primarily through the part to the printing base. Any interruption in this heat dissipation process can lead to the formation of heat accumulation zones. Geometric factors such as overhangs, printing area, and part height significantly influence heat accumulation. While current approaches, particularly those based on finite element methods (FEM), provide accurate assessments of heat accumulation, they are computationally intensive and often unsuitable for iterative design processes. This research introduces a graph neural network-based tool trained on FEM simulation data, designed to identify potential heat accumulation zones in LPBF parts using only geometric information. The tool achieves an average computational speedup of 4.44 \(\times \) compared to FEM models, while maintaining a mean zone detection accuracy of 65 \(\%\) across the test specimens. By generating heat accumulation maps and identifying the primary affected zones, the tool enables fast and reliable thermal assessment during the earliest stages of the design process. This accelerates decision-making and supports rapid design iterations before any printing procedure begins.