Session: 06-02: Heat Transfer in CSP Applications 1

Paper Number: 142263

142263 - A New Mpi-Accelerated Dem Heat-Transfer Model to Compute Radiation & Evaluate Nu-Correlations in Dense Moving Beds of Particles in Heat Exchanger 

Abstract: 

Next generation Concentrating Solar Power system employs ceramic particles as heat transfer medium under high temperatures. The intrinsic properties of ceramic particles —high thermal conductivity and capacity— make them ideal candidates for heat transfer and storage applications in the solar power industry. To better understand the complexities of the flow and heat transfer behaviors in the particle-based heat exchanger, Discrete element method (DEM) has gained great popularity because of depicting interparticle interactions in a more realistic way. A growing number of heat transfer models have been offered to calculate the radiative heat exchange. However, the state-of-the-art models still have several limitations: (1) Stefan-Boltzmann equation is widely used while ignoring the re-emission from reflected radiative energy, and (2) the particle emissivity is usually assumed to be constant while in the heated environment of heat exchanger, the emissivity value can fluctuate with the varying particle temperature. For the first restriction, our previous research has implemented our radiative heat transfer model using the radiosity equation into a DEM-based open source software, LIGGGHTS. However, the computational cost is largely increased because of solving a system of equations. In addition, the original message passing interface (MPI) acceleration technique used in LIGGGHTS is not compatible with the radiation model we implemented. Therefore, our research is confined to systems with <10,000 particles. To cope with these restrictions, we modify the radiosity equation and use an iterative method to approach the result numerically. This technique is found to cut down nearly 90% of the computational cost for computing radiative heat fluxes while sustaining satisfactory accuracy. We also manage to parallelize the radiative heat transfer model through MPI and therefore to enhance our capacity to simulate larger systems. Particularly when there is a variable particle emissivity distribution, the new technique shows promise as it yields precise outcomes without imposing the extensive computational burden that is characteristic of the radiosity equation. In this presentation, we will compare the results and computational cost between this newly developed methodology and the conventional strategies. To illustrate the effectiveness of our model, a gravity-driven, moving-bed heat exchanger model filled with alumina-based particles will be used as a showcase. This example will demonstrate the influence on a CSP heat exchanger from various geometric and physical properties, such as particle size, thermal conductivity and flow velocity. The Nusselt number will be evaluated with the presence of radiative effects along with the temperature dependence of thermophysical properties. With the development of efficient modeling techniques that accurately represent the complex physics governing these systems, we are closer to further develop the potential of CSP technology for sustainable energy production.

Presenting Author: Bingjia Li University of Michigan

Presenting Author Biography: Ph.D., Mechanical Engineering, University of Michigan, Ann Arbor, 2020-
M.S., Mechanical Engineering, University of Michigan, Ann Arbor, 2018-2020
B.S., Building Environment & Energy Application Engineering, China University of Mining and Technology, 2014-2018

Authors:

Bingjia Li University of Michigan
Zijie Chen University of Michigan
Rohini Bala Chandran University of Michigan