Abstract

Experimental work measuring heat transfer due to internal convection on a smooth straight passage is recreated using unsteady Reynolds averaged Navier-Stokes conjugate heat transfer simulations. The experimental work utilizes 1-dimensional and 3-dimensional conduction models to determine internal heat transfer rates from external surface temperature measurements collected with an infrared camera. The numerical simulations recreated these experiments to verify the conduction model and investigate the differences between the k-ω shear stress transport turbulence model, Reynolds stress turbulence model, and the k-ε turbulence model. It is found that the conduction model can accurately predict the heat transfer in the passage within an average error of 6% but with reduced spatial accuracy. The lower spatial accuracy can be accounted for by utilizing both the conduction model to predict the magnitude of the heat transfer and the numerical simulations to capture the spatial distribution. No one turbulence model was found to provide consistently superior heat transfer predictions, but rather each model excelled in some scenarios and underperformed in others. Overall, the k-ε model was found to best match the experimental heat transfer calculations with an average error of 5.9% of the total heat transfer, and it takes a more conservative approach as it can over predict the external surface temperatures by approximately 0.4 K. The end goal of this study is to develop a way to derive heat-flux data from infrared measurements on a range of geometries. A simple and well-understood geometry is investigated here to provide a firm foundation for future work.

This content is only available via PDF.
You do not currently have access to this content.