Heat Transfer Enhancement of Single-Phase Internal Flows using Shape Optimization and Additively Manufactured Flow Structures

Abstract

• We optimize the shape of internal fins using generative design. • The optimized fins are additively manufactured from alsi10 mg. • Rigorous measurements and simulations characterize heat transfer performance. • The optimized fins have 65–77% lower thermal resistance than conventional designs. This paper reports heat transfer enhancement of forced internal single-phase flow enabled by additively manufactured internally finned channels. We consider internal liquid coolant flow through a tube with constant heat flux applied to the exterior surface. A genetic algorithm was developed and used to optimize the fin geometry in order to minimize total thermal resistance for laminar flow or convection thermal resistance for turbulent flow. The genetic algorithm sampled more than 500 fin geometries, evaluated thermal resistance, and selected the optimum fin geometry. We selected three designs for experimental testing: a complex fin structure suggested by the genetic algorithm; a straight fin design; and a smooth circular channel as a reference. The devices were made from aluminum silica (AlSi10Mg) using direct metal laser sintering additive manufacturing. The devices were tested using a water ethylene glycol mixture (80:20) over the hydraulic diameter based Reynolds number range of 400 – 18,000, covering both laminar and turbulent flow regimes. Finite element simulations of heat transfer and pressure drop helped to understand and interpret the measurements. Compared to the smooth pipe reference design, the total thermal resistance of the genetically optimized fin design was 77% lower for laminar flow at Re h = 2,000 and 65% lower for turbulent flow at Re h = 10,000. Compared to the straight fin design that is commonly used for enhancing convection heat transfer in internal flows, the total thermal resistance of the genetically optimized fin design was 55% lower for laminar flow at Re h = 2,000 and 29% lower for turbulent flow at Re h = 10,000. We explore the potential design space for different types of heat exchangers by investigating how total heat transfer varies with exterior heat transfer coefficient and channel wall thermal conductivity. This research shows how shape optimization and additive manufacturing can result in devices with improved internal convection heat transfer.