Evaluation of Additively Manufactured Heat Exchanger Tubes Using Experimental and Conjugate Heat Transfer Analysis

Abstract Additive manufacturing offers many benefits for the design and production of gas turbine engine components. Direct metal laser sintering (DMLS) expands the available design space while reducing manufacturing cost and cycle time. Engineers continue to explore opportunities for additive components throughout the engine. Generally in regions of reduced stress, heat exchangers provide an opening for immediate implementation of DMLS components. The surfaces of components manufactured using the DMLS process are rough compared to those produced from traditional casting or subtractive manufacturing methods. Therefore, during the design phase, heat transfer designers must understand the heat transfer augmentation and pressure loss increase associated with the roughness from the DMLS process. In this study, DMLS tubes (as printed) were evaluated experimentally. The experimental results were used for validation of a conjugate heat transfer (CHT) model. With the CHT model accurately capturing the roughness effects of the tubes, the model can be confidently extended to other flow conditions and tube configurations. Within the study, a total of six printed tubes were experimentally tested and presented. Four tubes had a diameter of 4.0 mm, and the remaining tubes were 2.0 mm in diameter. The overall lengths of the tubes varied from 180 mm to 230 mm. The coolant flow rate through all tubes varied, yielding Reynolds numbers from 7,500 to 30,000. The overall pressure loss and heat transfer enhancement was measured in the printed tubes. These results were used to validate a CHT model using STAR-CCM+ for the same geometries and flow conditions. The measured Nusselt numbers were 1.25–2.4 times greater than those predicted by Dittus Boelter correlation for heat transfer in a smooth tube. The heat transfer enhancement came at the expense of increased pressure loss compared to smooth tubes. The frictional losses were 1.75–6.5 times greater than those predicted for smooth tubes. The thermofluidic performance of the additively manufactured tubes was generally well predicted by the Norris correlation for rough tubes. The CHT predictions also compare favorably with the Norris correlation. This experimental and numerical study shows the value of printed tubes for heat exchangers. It also highlights the challenges of measuring and predicting the thermal performance for various size tubes.