<p> CHF and Post-CHF tests were performed in water at 9.7 MPa using two vertical test assemblies having identical, internally heated annular flow channels, one heated directly, the other indirectly. Experiments were conducted to determine the effect of these methods of heating on CHF and Post-CHF heat transfer. </p> <p> For the range of the test conditions investigated, the results show that the direct and indirect heaters have similar CHF performance. At heat fluxes above CHF and mass fluxes of 2.0 and 3.5 Mg.s^-1 .m^-2 , the indicated maximum wall temperatures of the heaters were similar, but at the highest mass flux for the tests, 5.0 Mg.s^-1 .m^-2 , the indirect heater had lower indicated maximum wall temperatures than the direct heater for a given heat flux above CHF. </p> <p> A multi-fluid model, of the type used previously in the prediction of CHF, was derived and tested against the experimental data. The model, which considers droplet entrainment, deposition and evaporation in the annular flow regime, assumes dryout to occur when the liquid film flow on the inner rod approaches zero. The CHF predictions were in fairly good agreement with the experimental results. In general, the model under-predicted CHF at low inlet subcoolings and over-predicted CHF at high inlet subcooling. The error trend is consistent with that of the CHF prediction models of other researchers. </p> <p> In addition to the CHF prediction model, a Post-CHF model to predict the vapour temperatures, and hence, the heated wall temperature is also presented in the report. The theory is based on a physical model of heat transfer in the liquid deficient regime. In the model, heat in the dry region is assumed to transfer from the heated wall to superheat the steam and some of this heat, in turn, is used to evaporate the droplets which are entrained in the vapour core. Droplet entrainment and deposition at the shroud (outer tube) film-vapour interface are modelled. Heat transfer enhancement due to increased turbulence downstream of the rod centering spacers is incorporated through an empirical correlation. The predicted results were compared to the direct heater experiments. In general, the predicted wall temperatures were in agreement with those in the experiments. </p> / Thesis / Master of Engineering (MEngr)

The problem of low Mach number (non-Boussin´esq) conjugate laminar natural convection combined with surface radiation in a vertical annulus with a centrally located vertical heat generating rod is studied numerically, taking into account the variable transport properties of the fluid. Such problems arise often in practical applications like spent nuclear fuel casks, cooling of electrical and electronic equipment, convection in ovens, cooling of enclosed vertical bus bars and underground transmission cables. The physical model consists of a vertical heat generating rod, a concentric outer isothermal boundary and adiabatic top and bottom surfaces. The heat generation in the rod drives the natural convection in the annulus. Surface radiation is coupled to natural convection through the solid-fluid interface condition and the adiabatic condition of the top and bottom surfaces. A mathematical formulation is written using the governing equations expressing the conservation of mass, momentum and energy for the fluid as well as the energy balance for the solid heat generating rod. The governing equations are discretized on a staggered mesh and are solved using a pressure-correction algorithm. Steady-state solutions are obtained by time-marching of the time dependent equations. The discretized equations for the dependent variables are solved using the Modified Strongly Implicit Procedure. A global iteration is introduced on the variables at each time step for better coupling. The parameters of the problem are the heat generation and gap width based Grashof number, aspect ratio, radius ratio and the solid-to-fluid thermal conductivity ratio. The coupling of radiation introduces the wall emissivity and the radiation number as the additional parameters and also necessitates the calculation of radiation configuration factors between the elemental surfaces formed by the computational mesh. The radiant heat exchange is calculated using the radiosity matrix method. A parametric study is performed by varying Grashof number from 106 to 1010 , aspect ratio from 1 to 15, radius ratio from 2 to 8, the solid-to-fluid thermal conductivity ratio from 1 to 100, with the Prandtl number 0.7 corresponding to air as the working medium. The characteristic dimension and the outer boundary temperature are fixed. For Radiative calculations, and the emissivity is varied between 0.25 and 0.75. Converged solutions with laminar model could be obtained for high Grashof numbers also as the heat generation based Grashof number is generally two orders of magnitude higher than the temperature difference based Grashof number. Results are presented for the flow and temperature distributions in the form of streamline and isotherm maps. Results are also presented for the variation of various quantities of interest such as the local Nusselt numbers on the inner and outer boundaries, the axial variation of the centerline and interface temperatures, maximum solid, average solid and average interface temperature variations with Grashof number and the average Nusselt number variation for the inner and outer boundaries with Grashof number. The results show that simplification of conjugate problems involving heat generation by the prescription of an isoflux boundary condition on the rod surface is inadequate because a truly isoflux condition cannot be realised on the one hand and because the solid temperature distribution remains unknown with such an approach. The average Nusselt numbers on the inner and outer boundaries show an increasing trend with the Grashof number. For pure natural convection, the Boussin´esq model predicts higher temperatures in the solid and lower average Nusselt numbers on the inner and outer boundaries, compared to the non-Boussin´esq model and the Boussin´esq approximation appears to be adequate roughly upto a Grashof number of 109, beyond which the non-Boussin´esq model is to be invoked. The average pressure in the annulus is found to increase with an increase in the Grashof number. Radiation is found to cause convective drop and homogenize the temperature distribution in the fluid.