| 研究課題/領域番号 |
21J13418
|
|---|
| 研究機関 | 京都大学 |
|---|
研究代表者 |
李 昊 京都大学, 工学研究科, 特別研究員(DC2)
|
|---|
| 研究期間 (年度) |
2021-04-28 – 2023-03-31
|
| キーワード | topology optimizaion / parallel computing / mesh adaptation / thermal fluid problem / distributed elements / body-fitted mesh / level set method |
|---|
| 研究実績の概要 |
We propose a parallel distributed and open-source framework for full-scale 3D structural topology optimization (TO). This can be achieved by properly combining parallel computing and mesh adaption techniques by adopting a reaction-diffusion equation (RDE) based level-set method. Our proposed method can be applied to solve 2D and large-scale 3D multiphysics optimum design problem, i.e., mean compliance problem, minimal power dissipation problem, fluid-structure interaction problem, strongly coupled natural convection problem. Various the-state-of-the-art works have been published to demonstrate the validity of our methodology.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
1: 当初の計画以上に進展している
理由
Our research is ongoing as what we planned in our research proposal. First, the cornerstone of this work is to build up an open-source topology optimization framework which is highly scalable and easy-to-use. This has been achieved in our first paper published on Finite Element in Analysis and Design (Elsevier). Next, we extended from the elasicity to fluid-based problem, which requires much more numerical efforts on the solver. This part of the work has been carried out and has been published on Applied Mathemaical Modelling (Elsevier). After that, we further challenge a more complicated strongly coupled thermo-fluidic problem and this part of the work has been published on International Journal for Numerical Methods in Engineering (Wiley).
|
| 今後の研究の推進方策 |
we intend to break through the existing bottleneck of the sequentially performed body-fitted mesh evolution method. A fully parallel framework will be constructed. For accessing our newly updated framework, a classical aerodynamics problem lift-drag problem will be investigated. We incorporate three different remeshing techniques (isotropic, anisotropic, or body-fitted adaptive mesh refinement) into the reaction-diffusion equation-based (RDE) fluid topology optimization framework. This is owing to the flexibility of the RDE method for handling both “separate” and “hybrid” flow modeling strategies.
The lattice infill structure is of great interest from the biomimic standpoint. We will introduce a maximum length scale constraint in this workflow.
|
