<p>In the aerospace and defense industries, flow forming is a critical process for the integral fabrication of large-scale tubular components. Simulating this process using the finite element method (FEM) requires an excessively fine mesh to accurately capture the contact interactions, resulting in an infeasibly high computational cost. To address this challenge, a modified multimesh method is proposed to accelerate the simulation by reducing the overall mesh scale while preserving local refinement. This method employs a customized three-mesh system that enables the rapid and automatic construction of locally refined meshes using hierarchical 4- and 9-refined templates. Additionally, radial basis function interpolation (RBFI) combined with the restricted additive Schwarz method (RASM) to efficiently transfer state variables among meshes via localized subregion sampling. The complete multi-mesh framework was implemented in Python and integrated into ABAQUS/Explicit. Validation across multiple flow-forming cases demonstrated that both the final geometry and state variable distributions aligned closely with those obtained from traditional fine-mesh simulations. In the studied forward flow-forming examples, the proposed method reduced the computational time by up to 59.09%.</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 modified multi-mesh method for accelerating numerical simulation of flow forming

  • Yun-Da Dong,
  • Mei Zhan,
  • Xiao-Guang Fan,
  • Zhuo-Lei Zhai,
  • Yi-Yang Yang,
  • Heng Cao,
  • Fei Ma

摘要

In the aerospace and defense industries, flow forming is a critical process for the integral fabrication of large-scale tubular components. Simulating this process using the finite element method (FEM) requires an excessively fine mesh to accurately capture the contact interactions, resulting in an infeasibly high computational cost. To address this challenge, a modified multimesh method is proposed to accelerate the simulation by reducing the overall mesh scale while preserving local refinement. This method employs a customized three-mesh system that enables the rapid and automatic construction of locally refined meshes using hierarchical 4- and 9-refined templates. Additionally, radial basis function interpolation (RBFI) combined with the restricted additive Schwarz method (RASM) to efficiently transfer state variables among meshes via localized subregion sampling. The complete multi-mesh framework was implemented in Python and integrated into ABAQUS/Explicit. Validation across multiple flow-forming cases demonstrated that both the final geometry and state variable distributions aligned closely with those obtained from traditional fine-mesh simulations. In the studied forward flow-forming examples, the proposed method reduced the computational time by up to 59.09%.