Case study of natural convection topology optimization based on finite volume method

Abstract

Topology optimization is crucial for improving natural convection cooling devices. This study applies natural convection topology optimization to various engineering cases, including 2D, 3D, inclined, and inverted heat sink with various boundary conditions. Notably, finite element method has long been used as the analysis module of the topology optimization since early studies, which is different from most computational fluid dynamics software that uses finite volume method (FVM). This gap restricts applications of the natural convection topology optimization in common CFD software. This work presents an FVM-based topology optimization framework for natural convection, which is rare in previous studies. The fluid is modeled as a steady-state incompressible laminar flow using FVM. The topology optimization model is established using Brinkman term and thermal conductivity interpolation. Meanwhile, a parallel computing method is also developed to accelerate computation. Combined with the FVM-based analysis, the topology optimization is conducted on 2D and 3D heat sink designs to demonstrate its effectiveness and accuracy. In 2D problems, heat sinks under different dominant heat transfer forms and working angles are discussed. In 3D problems, a case involving numerous design variables is conducted using common PC, which demonstrate the validity of this work in large-scale natural convection topology optimization.

Keywords

• Natural convection
• Topology optimization
• Finite volume method (FVM)
• Large scale
• Heat sink
• Multi-dimensional design