Effect of heat flux and mass flux on the heat transfer characteristics of supercritical carbon dioxide for a vertically downward flow using computational fluid dynamics and artificial neural networks.

Drastic variation in the thermodynamic properties of supercritical fluids near the pseudo critical point hinders the use of commercial computational fluid dynamics (CFD) software. However, with the increase in computational abilities, along with the use of Artificial Neural Networks (ANN), turbulence heat transfer characteristics of supercritical fluids can be very accurately predicted. In the present work, heat transfer characteristics for a vertically downward flow of carbon dioxide in a pipe are studied for a wide range of heat flux and mass flux values. Firstly, six different turbulent models available in the commercial CFD software - Ansys Fluent are validated against the experimental results. The k-? Standard model with enhanced wall treatment is found to be the best-suited turbulence model. When experimental results were validated in CFD, an average error of 1% in the bulk fluid temperature and 2% in the wall temperature were recorded. Further, K-? Standard Turbulence Model is used in CFD for parametric analysis to generate the data for ANN studies. Mass flux range of 238 to 1038 kg/m2s, and heat flux range of 26 kW/m2 to 250 kW/m2 are used to generate 81,432 data samples. These samples were fed into the ANN program to develop an equation that can predict the heat transfer coefficient. It was found that ANN can predict the heat transfer coefficient for the considered range of values within the absolute average relative deviation of 2.183 %. [ABSTRACT FROM AUTHOR]