Scholarlink

<p>In the present paper, we propose a parametrized Level-Set (LS) approach based on a new Immersed Finite Element (FE)-LS strategy for Topology Optimization (ILSTO). Arbitrary design domains with generic boundaries are efficiently described by the proposed approach. Within this method, a generic FE discretization is mapped on a regular LS grid of knots, avoiding the TO problem from the necessity of postprocessing and filtering&#xa0;the obtained topology. In addition, using an approximating function for expressing the spatial derivatives and relying on a simple definition of reinitialization, the solution does not suffer from numerical instabilities, commonly associated with the calculation of the LS surface gradients. The proposed ILSTO method is benchmarked against conventional methods to demonstrate its improved geometrical resolution and numerical robustness, while maintaining computational efficiency.</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

Structural topology optimization based on an immersed FEM Level-Set method

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年8月12日

查看全文
  • Farzad Tatar,
  • Roberto Brighenti

摘要

In the present paper, we propose a parametrized Level-Set (LS) approach based on a new Immersed Finite Element (FE)-LS strategy for Topology Optimization (ILSTO). Arbitrary design domains with generic boundaries are efficiently described by the proposed approach. Within this method, a generic FE discretization is mapped on a regular LS grid of knots, avoiding the TO problem from the necessity of postprocessing and filtering the obtained topology. In addition, using an approximating function for expressing the spatial derivatives and relying on a simple definition of reinitialization, the solution does not suffer from numerical instabilities, commonly associated with the calculation of the LS surface gradients. The proposed ILSTO method is benchmarked against conventional methods to demonstrate its improved geometrical resolution and numerical robustness, while maintaining computational efficiency.