Performance Improvement of Tree-Shaped Network With a Hybrid Lagrange Multipliers and Topology Optimization Method

Abstract

Abstract. Dendritic path architecture, like a tree-like branching channel network, has attracted considerable attention in thermofluids applications due to its ability to deliver uniform flow distribution, low energy consumption, and high heat dissipation efficiency. However, the existing design theories still have drawbacks, such as the effect of flow and heat transfer at the junctions of bifurcated channels was not considered. In recent years, topology optimization (TO) has emerged as a powerful tool to advance the design of thermal management systems, such as cold plates, enabling enhanced thermal performance and reduced energy usage. However, a key limitation of the TO approach is the significant computational time required in the design process. This work presents a hybrid design method for a bifurcated channel structure through the Lagrange multipliers (LM)-based optimization method, followed by TO. This hybrid approach can not only enhance thermal dissipation efficiency and minimize flow resistance but also achieve a better performance with reduced computational time. The LM-based design method takes into consideration the effects of developing flow, thermal conductivity, and fluid properties on the dimensions of bifurcated channels. Next, the bifurcated channel was further optimized using the TO method to achieve the improved outlet temperature and pressure drop. The hybrid approach provides a new path for designing tree-like branching channel networks for a wide range of thermal applications.