Scholarlink

<p>Manufacturing multi-physics simulations are widely used to predict the geometry and dimensions of the final product considering the manufacturing process parameters. The STereoLithography (STL) file format has become a widely adopted standard for representing three-dimensional (3D) geometries using a surface mesh of connected triangular facets and depending on the application, can serve as an input to these simulations. However, the simulation results cannot typically be aligned with STL format as a surface-based representation. This paper introduces a methodology for transforming the results of simulation defined at the nodes of hexahedral finite elements into a continuous field across the surface of all triangles in an STL mesh. The methodology generates a matrix-based framework to apply trilinear interpolation on the known nodal data and then feeds the outputs to a quadratic bivariate polynomial to construct a continuous field over the STL tessellation. The proposed approach maintains numerical stability by avoiding matrix singularities and is structured to support downstream integration with data-driven methods such as machine learning. As an application, the methodology is used to process the results of Laser Powder Bed Fusion (LPBF) simulation. The results are validated using various case studies showing a consistent and accurate STL representation of the predicted dimensions of the manufactured parts. The output can be directly used for manufacturing process control, error compensation, and or process optimization.</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

Mapping manufacturing finite element simulation into STL-based geometric deviation field

  • 期刊论文

  • 发表日期：

    2026年7月8日

查看全文
  • Yasaman Farahnak Majd,
  • Ahmad Barari

摘要

Manufacturing multi-physics simulations are widely used to predict the geometry and dimensions of the final product considering the manufacturing process parameters. The STereoLithography (STL) file format has become a widely adopted standard for representing three-dimensional (3D) geometries using a surface mesh of connected triangular facets and depending on the application, can serve as an input to these simulations. However, the simulation results cannot typically be aligned with STL format as a surface-based representation. This paper introduces a methodology for transforming the results of simulation defined at the nodes of hexahedral finite elements into a continuous field across the surface of all triangles in an STL mesh. The methodology generates a matrix-based framework to apply trilinear interpolation on the known nodal data and then feeds the outputs to a quadratic bivariate polynomial to construct a continuous field over the STL tessellation. The proposed approach maintains numerical stability by avoiding matrix singularities and is structured to support downstream integration with data-driven methods such as machine learning. As an application, the methodology is used to process the results of Laser Powder Bed Fusion (LPBF) simulation. The results are validated using various case studies showing a consistent and accurate STL representation of the predicted dimensions of the manufactured parts. The output can be directly used for manufacturing process control, error compensation, and or process optimization.