GPU‐accelerated simulation of polydisperse multiphase flows using dual‐quadrature‐based moment methods. (21st January 2020)

• Record Type: Journal Article
• Authors: Santos, Fabio P.; Lage, Paulo L. C.; Favero, Jovani L.; Senocak, Inanc
• Is Part Of: Canadian journal of chemical engineering. Volume 98:Number 5(2020)
• Page Start: 1211
• Page End: 1224
• Publication Date: 2020-01-21
• Subjects: moment methods -- multiphase flow -- particulate systems -- population balance; Chemical engineering -- Periodicals; Technology -- Periodicals; 660.05
• DOI: 10.1002/cjce.23697
• Languages: English
• ISSNs: 0008-4034

Abstract

Abstract: Polydisperse multiphase flows can be simulated by coupling a population balance model to the multi‐fluid model. For the first time, this simulation is carried out using the dual‐quadrature method of generalized moments (DuQMoGeM) and its direct version to solve the population balance model. The main disadvantage of these methods is the high computational cost of the embedded cubature. Herein, this challenge was addressed by parallelization on graphics processing units, which resulted in a significant acceleration of the simulations, with speedups that can be larger than 1000. Numerical simulations considering simultaneous particles breakage and aggregation were conducted with direct DuQMoGeM‐FC and DQMoM‐FC, whose results were different due to the existence of quadrature error in the latter.