Computer Simulation of Magnetic Resonance Imaging of the Flow of Fluidized Particles

Abstract

Rapid magnetic resonance imaging (MRI) and velocimetry (MRV) are non-invasive measurement techniques in 3D opaque systems with ∼10 ms temporal resolution, enabling new opportunities to challenge the accuracy of computational models of flow. Comparisons between rapid MRI/MRV and computer simulations are limited by (i) artifacts in rapid MRI techniques and (ii) non-trivial aspects of post-processing of computer simulation data to best match the measurement techniques. Here, we address these issues by starting with data from computer flow simulations of fluidized particles and feeding the data to a physics-based simulation of MRV measurements, which captures potential artifacts introduced by the measurement techniques. Flow simulation data is then post-processed in various ways, demonstrating that (1) velocity and particle position data must be taken from flow simulations at different points in time to match MRV measurements and (2) imaging must be faster than flow fluctuation for MRV to produce effectively instantaneous velocity fields.

Feasibility Analysis of a POD-Based Reduced Order Model with Application in Eulerian–Lagrangian Simulations

“…Multiphase flows involving solid particles are ubiquitous in nature and industry. The Eulerian–Lagrangian simulations by coupling computational fluid dynamics and the discrete element method (i.e., CFD-DEM) play an important role in simulating and understanding various gas–solid two-phase flows. − Recently, advanced numerical methods coupling the DEM have been further developed to simulate more complex gas–liquid–solid three-phase flows. − More applications of the DEM can be found in refs , . Even though the advanced DEM-based numerical methods are powerful, computational costs are usually huge.…”