Optimizing Additively Manufactured Heat Exchangers for Micro-Channel CFD

Thesis Abstract

The study addresses the design, manufacturing, and performance evaluation challenges of heat exchangers produced by additive manufacturing (AM) with micro-channel geometries, focusing on achieving enhanced thermal efficiency while ensuring structural integrity and manufacturability under practical operating conditions. The aim is to optimize the geometry, material selection, and process parameters of AM heat exchangers to maximize heat transfer coefficients and minimize pressure drops, with robust validation across multi-physics simulations and experimental testing. Specific objectives include (1) to identify micro-channel geometries and lattice supports that balance heat transfer, pressure loss, and mechanical strength; (2) to quantify the effects of AM build orientation, post-processing, and surface roughness on thermal performance; (3) to develop a calibrated reduced-order model (ROM) and surrogate model for rapid prediction of thermal-hydraulic performance; (4) to validate the models against experimental data from a test rig comprising 10 printed samples with varying channel dimensions; (5) to propose design guidelines and process-structure-property correlations grounded in mechanistic and statistical analyses. A mixed-methods research design is employed, integrating computational fluid dynamics (CFD), finite element analysis (FEA), and empirical testing. The population comprises AM-produced copper and aluminum micro-channel heat exchangers, with a purposive sample of 10 specimens representing three channel geometries (circular, trapezoidal, and wavy-walled) and two alloy systems (CuCrZr and AlSi10Mg). Data collection instruments include high-resolution digital microscopes for surface roughness (Ra), X-ray computed tomography (XCT) for internal geometry verification, a calibrated thermal-fluid test bench for steady-state performance measurements (inlet temperature range 60–120°C, mass flow rates 0.2–1.0 kg/s), and digital pressure sensors for pressure drop data. Instrument validation incorporates traceability checks, calibration routines, and repeat measurements (n=3 per specimen) to ensure reliability. Methodologically, the study integrates design of experiments (DOE) to explore micro-channel parameters, regression analysis and analysis of variance (ANOVA) to identify significant factors influencing Nusselt number, pressure drop, and overall heat transfer effectiveness, and multi-objective optimization (pomdp or NSGA-II) to balance thermal and hydraulic performance. A local Reynolds-averaged Navier-Stokes (RANS) CFD model with k-? SST turbulence closure is used to simulate laminar and transitional regimes within micro-channels, incorporating surface roughness effects and contact resistance at interfaces. Thermo-mechanical coupling via FEA assesses structural integrity under thermal cycling and pressure loading, generating safety factors and predicting deformation. A reduced-order model (ROM) based on proper orthogonal decomposition (POD) and Kriging surrogate modeling provides rapid performance predictions for design iterations. The theoretical framework combines convective heat transfer theory, porous media concepts for lattice segments, and the design of experiments theory, with anchoring hypotheses drawn from the theory of fluid-saturated porous media and surface roughness-dependent heat transfer. Expected findings indicate that optimized micro-channel geometries with controlled surface roughness (Ra in the range 1.5–4.0 ?m) and orientation-optimized builds yield a 12–28% increase in effective heat transfer coefficient compared to baseline smooth channels, while maintaining pressure drops within 15–25% of conventional counterparts. The ROM and surrogate models are anticipated to predict performance within ±6% of full CFD results across the design space, enabling rapid design exploration. The study is expected to reveal critical process-structure-property relationships, including the influence of lattice topology on flow obstruction and thermal resistance, and will establish correlations between XCT-derived geometric deviations and observed thermal performance. Contributions to knowledge include (i) an integrated methodology for optimizing AM micro-channel heat exchangers combining CFD, FEA, and experimental validation; (ii) empirically grounded design guidelines linking AM parameters, micro-channel geometry, and thermal-hydraulic performance; (iii) validated ROM and surrogate models enabling rapid multi-objective optimization in industrial contexts. The study recommends adopting standardized post-processing protocols to minimize detrimental roughness effects, implementing orientation-aware build strategies for complex micro-channels, and leveraging the developed design framework for next-generation AM heat exchangers in automotive and aerospace cooling systems.