An efficient and reliable predictive method for fluidized bed simulation

Abstract

In past decades, the continuum approach was the only practical technique to simulate large-scale fluidized bed reactors because discrete approaches suffer from the cost of tracking huge numbers of particles and their collisions. This study significantly improved the computation speed of discrete particle methods in two steps: First, the time-driven hardsphere (TDHS) algorithm with a larger time-step is proposed allowing a speedup of 20-60 times; second, the number of tracked particles is reduced by adopting the coarse-graining technique gaining an additional 2-3 orders of magnitude speedup of the simulations. A new velocity correction term was introduced and validated in TDHS to solve the overpacking issue in dense granular flow. The TDHS was then coupled with the coarse-graining technique to simulate a pilot-scale riser. The simulation results compared well with experiment data and proved that this new approach can be used for efficient and reliable simulations of large-scale fluidized bed systems. V C 2017 American Institute of Chemical Engineers AIChE J, 63: 5320-5334, 2017 Figure 20. Computation speed (CPU time to simulate 10 thousand steps) using MFIX-DEM program with increased number of particles.Figure 21. Computation speed of case W125 and W1000.

Toward Constitutive Models for Momentum, Species, and Energy Transport in Gas–Particle Flows

“…When one tracks the motion of only a representative particle or subset of particles through this latter approach, it is equivalent to the multiphase particle-in-cell (MP-PIC) approach, where the representative particles are commonly referred to as parcels (24,25). A hybrid option, where collision between the parcels is tracked to quantify direct parcel-parcel interaction (26)(27)(28), obviating the need to postulate a stress model, is also very popular. The PSD and dynamic evolution of particle properties are handled more efficiently by the particle-based methods.…”

“…As noted in the previous section, coarse-grained EL simulation approaches that are suitable for gas-particle flows in large process vessels have also been developed in the literature. In these approaches, Newton's equations of motion are solved for a manageable number of representative particles (i.e., parcels); the MP-PIC method employs a user-specified model for particlephase stress and a relaxation step to capture parcel-parcel interaction (25), whereas the CFD-DPM approach tracks and accounts for parcel-parcel contact interaction (26,27). More specifically, DPM treats each parcel as a particle having the same density as the primary particle it represents and the same total volume of all the primary particles in that parcel.…”

“…Bubble-like voids in dense fluidized beds and clusters in more dilute regions have indeed been captured by MP-PIC simulations (69). Recently, Lu et al (28) assessed the predictability of a coarse-grained EL approach with a hard-sphere contact model (27) by performing simulations of gas-particle flow in a small-scale fluidized bed and showed that the coarse-grained EL approach also captured bubbles and clusters.…”

Coarse-Grain DEM Modelling in Fluidized Bed Simulation: A Review

“…A very significant improvement in speed for computations is obtained by combining coarse graining and hard-sphere event-driven models for collisions, without compromising the accuracy of the predictions in fluidized beds, according to Lu et al [87,88] (Figure 12). Together with coarse-graining, a corrected hard-sphere model specifically derived by the authors, formed the CGHS (coarse grain hard-sphere) model, capable of providing an overall impressive speed-up of 24 using f = 2.…”

Impact of Contact Scaling and Drag Calculation on the Accuracy of Coarse‐Grained Discrete Element Method

“…Other authors consider the voidage within the CG-parcel e CGP and argue with the invariance of kinetic energy and energy dissipation during the collision process [69,70]. Instead of an explicit scaling of the calculated contact forces, they scale the restitution coefficient by…”