Session: VVUQ for Fluid Dynamics and Heat Transfer

Paper Number: 152278

152278 - Heat Transfer Correlations for Downward-Facing Curved Surfaces Under Upward Flow Conditions 

Abstract: 

This study presents heat transfer correlations for upward flow over a curved, downward-facing surface with uniform heat flux conditions. The scope of the study includes both forced and mixed convection regimes within the range of 1,000<Re<13,000, 2.56<Pr<4.36 and 0.001<Ri<10. In addition to experimental measurements, a comprehensive computational fluid dynamics study is conducted using six different RANS models along with Large Eddy Simulation (LES) to evaluate the capabilities of each of them in predicting heat transfer physics on curved geometries, which include separation and buoyancy effects. A comprehensive quantification of errors and uncertainties, employing VVUQ principles to ensure the reliability of the results. Two distinct correlations were developed: one for the forced convection regime (Ri < 0.1) and another for the mixed convection regime (0.1 < Ri < 10).

In the forced convection regime, the results indicate that curved surfaces have unique heat transfer characteristics compared to flat surfaces. This difference in behavior can be attributed to the changes in boundary layer development and flow separation influenced by the curvature of the surface, which alters the local heat transfer rates.

For the mixed convection regime, the heat transfer behavior was analyzed using a dimensionless Buoyancy number (Bo). The ratio of mixed-to-forced Nusselt numbers (Nu/Nu_f) was observed to follow a trend similar to that reported for other geometries. Specifically, Nu/Nu_f drops sharply as Bo approaches 1, attributed to the hindrance caused by buoyancy forces, and subsequently increases as Bo exceeds 1, where the buoyancy forces enhance convective heat transfer. To accurately represent this behavior, the mixed convection heat transfer was modeled using two separate correlations, one for Bo < 1 and another for Bo > 1, capturing the transition from buoyancy hindrance to buoyancy enhancement effectively.

The benchmarking of RANS models against LES was an essential part of this study to determine the most suitable approach for capturing the complex flow and heat transfer phenomena observed over curved surfaces. LES, known for its higher accuracy in resolving turbulent structures, served as a reference for evaluating the performance of the RANS models. The comparison revealed significant differences in the predictive capabilities of the models, which is explained based on each model's limitations.

These findings contribute to a more comprehensive understanding of heat transfer over curved surfaces, particularly in mixed convection regimes. By benchmarking multiple turbulence models, the study also provides valuable guidance for future numerical simulations involving complex geometries and mixed convection flow.

Presenting Author: Samah Albdour Khalifa University

Presenting Author Biography: Samah Albdour

Authors:

Sameer Osman Khalifa University
Imran Afgan Khalifa University
Yacine Addad Khalifa University