Extended DEM Simulations of Particle Motion and Heat Transfer in Solar Reactor Systems with Metal Oxide Particles

Abstract

Metal oxides are used in two-step thermochemical redox cycles for the solar hydrogen, syngas or fertilizer production. In several corresponding prototype systems, the metal oxide is present in granular or particle form. To model the performance of the system components, it is crucial to describe the particle motion correctly, because it often has a large effect on the overall heat transfer and hence the reactions. The Discrete Element Method (DEM) is an excellent method for this. Here we would like to present various extensions to incorporate heat transfer and chemical reactions into DEM simulations and how these were used to model different solar receiver and reactor systems. The calibration of DEM parameters is addressed as well. One example we are going to show is a particle mix reactor, in which heat carrier particles from a solar receiver should be mixed with SrFeO3 particles to reduce them under vacuum conditions. To model the system, we coupled the DEM code LIGGGHTS with a custom ray-tracing code for the radiative heat transfer and implemented a new heat conduction model which works with different sized particles and also accounts for the decrease in conductivity at low pressures. In a sensitivity study, we show how model parameters influence the results. The contact force model parameters were determined in a systematic bulk calibration procedure, which is also suitable for the parameters between two different types of particles. The results were validated by comparison to cold and high temperature experiments with temperatures of up to 1100°C. Mixture homogeneity was analyzed by a novel destruction-free segmentation technique, the mixing process was observed with a high-speed camera and transient data of thermocouples placed in the mixture of particles was recorded during heat exchange and reaction. Another example is ongoing work on a reactor system with ceria particles for the production of hydrogen and syngas. In the system a purge gas is used to reduce and quench the redox particles. Therefore, in addition to the models described before, the DEM has also been coupled to OpenFOAM for particle-fluid interaction. Preliminary and not fully validated simulations are shown for this process and an overview on ongoing work is given.