Experimental validation of multiphase particle-in-cell simulations of fluidization in a bubbling fluidized bed combustor

As part of our long-term aim at developing a complex CFD simulation tool for bubbling fluidized bed boilers operating in oxyfuel regime, we have performed specialized fluidization experiments to validate our numerical algorithm for multiphase flow simulation. The algorithm combines and extends the function of several components of OpenFOAM, using the multiphase particle-in-cell method implemented into a custom solver derived from coalChemistryFoam. In the first part of this contribution, the mathematical model and the important improvements in the implementation of the numerical algorithm are discussed.The second part focuses on the design of the experiments and the methodology of the validation. The experimental device used is a model of the combustion chamber, the distributor, and the windbox. Its transparent walls are made of polymethyl methacrylate. In the experiments, three different materials (sand and two types of LECA suitable for use in fluidized bed combustion) were subjected to fluidization at different flow rates, using air at room temperature as the fluidization medium. The flow rates and the pressure drop at the fluidized bed were measured. Video recordings were taken and automatic image postprocessing algorithms allowed to identify the surface of the fluidized bed, its height and its shape as viewed from the front. Based on the sieve analysis, accurate particle size distributions were calculated and implemented in the model. Afterward, series of simulations were performed in order to find the correct settings of the particle stress model and the gas-solid drag model. In particular, to avoid excess complexity when treating non-spherical particles, the modified Ergun-Wen-Yu drag model was used with added amplification factors depending on the local regime of the multiphase flow (dense vs. dilute). The best obtained combination of parameter settings is material-dependent and allows to achieve satisfactory agreement with the experiments in terms of the time-averaged pressure drop and bed height. Qualitative assessment of fluidization in terms of the properties inherent to materials in different Geldart groups was also performed and found to be strongly influenced by the choice of the particle stress model. Finally, the results of the simulations were analyzed to evaluate the minimum fluidization velocity and the average bubble size. The obtained values were compared to semi-empirical correlations available in literature.