Magnetic resonance imaging studies of forced and free convective heat transfer in packed beds and fluid columns

Abstract

Prediction of fluid flow and associated energy transport is an essential component in many engineering applications where analytical solutions are not possible. In these systems experimentation and numerical simulations are a necessary part of the design process. This work focuses on the experimental study of mass and energy transport in packed beds and pure fluids under forced and natural convection using nuclear magnetic resonance (NMR) imaging (MRI) techniques. It further evaluates the efficacy of commercial computational fluid dynamics (CFD) software to simulate these processes. The study of heat transfer via NMR has proven difficult historically, despite sensitivity of NMR parameters to temperature. Here, a novel experimental setup is pioneered, which enables the study of heat transfer in packed beds. The method employs fluorinated pore-filling fluid and hydrogen-rich core-shell packing particles. Hydrogen and fluorine are NMR-active chemicals that can be imaged with the same experimental equipment by adjusting the resonance frequency; providing means to image the two domains separately. Pore- fluid velocities and particle-wax melting are observed in the same packed bed, at sub-millimeter resolutions, presenting a more complete picture of the conditions in these hard-to-measure systems. In the presented studies, this methodology is demonstrated under forced convection and proven capable in identifying and correlating spatial variations in heat transfer to pore-fluid velocity. The technique is then employed to assess the accuracy of a CFD model in the commercial software package, STAR CCM+, using the melt to quantify energy absorbed by the bed. In natural convection studies of a pure fluid and packed bed in the Rayleigh-Bénard configuration, the axial circulation pattern is found to change with axial position in the long narrow cylinder, a result that is rarely discussed in literature. A CFD model is shown to match well with these experimental findings. In porous media convection with sub-, near- and super- critical fluid, the rapidly changing thermal diffusivity was captured by the rate the particles absorb energy. Finally, a correlation is developed allowing particle-wax T 2 relaxation time to be converted into temperature.