Simulating bubble dynamics in a buoyant system. (31st October 2019)

• Record Type: Journal Article
• Authors: Arai, Erin; Villafranco, Dorien; Grace, Sheryl; Ryan, Emily
• Is Part Of: International journal for numerical methods in fluids. Volume 92:Number 3(2020)
• Page Start: 169
• Page End: 188
• Publication Date: 2019-10-31
• Subjects: bubble dynamics -- buoyant systems -- multiphase flows -- rising bubble dynamics -- smoothed particle hydrodynamics (SPH) -- volume of fluid (VOF); Fluid dynamics -- Mathematics -- Periodicals; 532
• DOI: 10.1002/fld.4778
• Languages: English
• ISSNs: 0271-2091

Abstract

Summary: Multiphase flows are critical components of many physical systems; however, numerical models of multiphase flows with large parameter gradients can be challenging. Here, two different numerical methods, volume of fluid (VOF) and smoothed particle hydrodynamics (SPH), are used to model the buoyant rise of isolated gas bubbles through quiescent fluids for a range of Bond and Reynolds numbers. The VOF is an Eulerian grid–based method, whereas the SPH is Lagrangian and mesh free. Each method has unique strengths and weaknesses, and a comparison of the two approaches as applied to multiphase phenomena has not previously been performed. The VOF and SPH simulations are compared, verified, and validated. Results using two‐dimensional VOF and SPH simulations are similar to each other and are able to reproduce numerical benchmarks and experimental results for sufficiently large Morton and Reynolds numbers. It is also shown that at low Reynolds numbers, the two methods, SPH and VOF, diverge in the transient regime of the bubble rise. Regimes that require simulations capable of representing three‐dimensional drag are identified as well as regimes in which results from VOF and SPH diverge. Abstract : This work compares the volume of fluid (VOF) and smoothed particle hydrodynamics (SPH) methods for multiphase flows in the context of bubbles rising through viscous liquids. Model results are compared with each other and both experimental and numerical data from the literature. Observations of agreement and disagreement between the two approaches are made and recommendations for method usage based on parameter regimes are issued.