Fully-resolved simulations of heat transfer in particle-laden flows
Johns Hopkins University
Solid particles suspended in a fluid flow are encountered in many industrial applications, environmental processes and natural systems, such as fluidized beds, cloud formation, dust and pollutants dispersion, industrial mixers, oceanic plankton and many others. In the present dissertation we carry out fully resolved numerical simulations of several problems in this general area both with and without particles-fluid heat transfer. An important aspect of the work is that the finite size of the particles is properly accounted for and that the fluid dynamic forces acting on them are based on an accurate solution of the fluid equations rather than parameterized. The general approach used in this study is based on the PHYSALIS method. This method uses local analytic solutions as “bridges" between the particle surfaces and a fixed underlying Cartesian grid. For the isothermal case, we study the rotational dynamics of a particle free to rotate around a fixed center in a turbulent flow. Fixing the particle center and carrying out parallel simulations of the flow without the particle enables us to fully characterize the flow incident on the particle. We determine the scales of eddies interacting most with the particle and explore the effect of vortex shedding on the rotational dynamics. The Magnus mechanism is not found to play a significant role. To account for particles-fluid heat transfer phenomena, we have extended PHYSALIS to deal with the energy equation. This new direct numerical simulation method for non-isothermal systems is described in detail and extensively validated against experimental studies and analytical solutions. The method is implemented numerically on a GPU-centric code, which is compatible with BLUEBOTTLE – a highly efficient GPU-centric computational fluid dynamics framework. An example of particles transported by a Rayleigh-Bénard convective flow is shown to demonstrate the potential applications of our method. A further application to the thermal wake of particles in turbulent flow is also given.
Fully-resolved simulations of extended particles, Heat transfer in particulate flows, Physalis method