(271f) Gas-Solid Flow Models Based on the Kinetic Theory of Granular Flows: What Have We Learned Since MFIX Version 1? | AIChE

(271f) Gas-Solid Flow Models Based on the Kinetic Theory of Granular Flows: What Have We Learned Since MFIX Version 1?


Fox, R. - Presenter, Iowa State University
Collaborating with his PhD advisor, Dimitri Gidaspow, and coworkers at NETL, Madhava Syamlal published the first opensource code for simulating gas-solid flows based on the kinetic theory of granular flows (KTGF) in 1993. The MFIX code (https://mfix.netl.doe.gov/) is now in release 22.1 and has been used by over 7000 researchers worldwide. Moreover, its success provided the foundation for many commercial and academic codes using the same fundamental modeling approaches developed for the Eulerian-Eulerian version of MFIX.

Although the KTGF only provides conservation equations for the solid phase, it is especially useful for modeling moderately dense gas-solid flows such as fluidized beds and risers. This is because the gas phase serves as a source of kinetic energy through drag to keep the solid particles in a state of “rapid granular flow”. As opposed to dilute flows, particles in moderately dense flows undergo sufficient collisions to allow for an Eulerian description that tracks only the mass, momentum and kinetic energy of the particle phase. Notwithstanding this relative simplicity, the numerical solution of the coupled gas-solid conservation equations is wrought with “hidden” difficulties that must be dealt with carefully to enable robust simulations. Indeed, as embodied in MFIX, Syam’s contributions to the numerical algorithms are one of the keys to its success.

In this presentation, I will give an overview of my group’s contributions to enhancing the capabilities of the KTGF models used in MFIX. Specifically, I will discuss the treatment of polydisperse particles (e.g., with a wide particle size distribution) and the extension of the particle-phase conservation equations to handle relatively dilute regions where collisions are infrequent relative to spatial transport. Using high-speed compressible gas-particle flows as an example, I will show results based on state-of-the-art hyperbolic flow solvers that illustrate the tremendous advancements achieved in this field since Syam’s seminal work on MFIX release 1.