Specifics of forced-convective heat transfer to supercritical CO2 flowing upward in vertical bare tubes

Abstract

Heat transfer in the forced convection regime of fluids at supercritical conditions has been studied extensively for the past 60 years. The dominant approach to summarize the experimental results was by proposing empirical correlations for the data within the investigated range of parameters. It was soon realized by researchers worldwide that heat transfer coefficients become non-linear functions of wall and bulk-fluid temperatures at certain combinations of experimental parameters within the region of the peak of specific heat at supercritical pressures. Thus, it has become a standard approach to remove nonlinear experimental heat transfer coefficient values treating them as a sign of a deteriorated (as opposed to normal) heat transfer regime. There were recent attempts to address this shortcoming and extend the applicability of conventional empirical correlations to the deteriorated heat transfer regime. However, these attempts were not satisfactory. In this thesis, a new methodology has been developed that allows the use conventional empirical correlations without distinguishing entrance effects or deteriorated heat transfer regime. The methodology is based on binning experimental data according to the parameter X = (h_b - h_pc) / (q/G) and then combining correlations based on wall and bulk-fluid temperature on each bin to minimize RMS and maximal overprediction of heat transfer coefficients within each of the bins. Using this methodology, 95% of normal heat transfer data were predicted with a spread of ±19%, which is 1.74 times narrower compared to the prediction by the empirical correlations developed based on the conventional methodology and on the same data; while all the data (2786 points, including entrance effects and deteriorated heat transfer) were predicted with a spread of ±20% (based on 2σ-level). The data correlated based on the new methodology where obtained within the following range of experimental parameters: P = 7.58 – 8.91 MPa, Tb = 20 – 142 ˚C, Tw = 32 – 231 ˚C, G = 885 – 3048 kg/m2s, q = 26 – 616 kW/m2K, D = 8.1 mm. The experimental data were obtained based on a series of tests on supercritical CO2 flowing upwards in a bare tube at the MR-1 loop (located in Chalk River) of the former Atomic Energy of Canada Limited (AECL). Normal, deteriorated, and improved heat transfer regimes were covered in the experiments.