A numerical study of solid-liquid phase change with Marangoni effects using a multiphase approach

Multiphysics modelling of Gas Tungsten Arc Welding on ultra-thin-walled titanium tubing

“…This can be understood intuitively with what happens to the meniscus of water in a tube 10 mm wide and a tube 10 cm wide, and this is in a case where g = 9.81 m s −2 if g = 4.4145 × 10 −4 m s −2 the difference would be overwhelming. Thus gtawFoam is completely unable to match the benchmark as stated in both [118] and [37]. With both references giving the same value it is unlikely both have made the same typo (or were read incorrectly) but intuitively at 10 mm width the simulation simply cannot work with a reasonable (the capillary action subsides at ≈ 1 × 10 −6 N m −1 .)…”

“…Another benchmark used in Saldi's thesis is for the validation of the solver for three phase Marangoni driven flow. The benchmark, performed by Tan et al [118], is an ANSYS FLUENT simulation of liquid bismuth, solid bismuth and argon gas in a differentially heated rectangular domain. Their simulation involved finding the steady state melt front for various linear temperature differences between the hot and cold walls in a microgravity environment.…”

“…Figure 5.21: 2D domain used for the Marangoni driven melting of bismuth benchmark with centimetres as opposed to millimetres as performed by Tan et al [118] and matched by Saldi [37]. As detailed in Section 5.4.2.1, with low gravity and a 15 mm width (as stated in both [118] and [37]) the surface tension force is overwhelmingly dominant. The solution used to resolve this is to scale the domain to 15 cm width.…”

“…In the benchmark case a microgravity environment where g = 4.4145 × 10 −4 m s −2 is used. An issue with this is that at the 15 mm domain width and thus a 10 mm liquid fraction width as stated in [37,118] the surface tension force of bismuth (0.0378 N m −1 ) [108] is overwhelmingly strong and the phase fraction drags up the walls through essentially capillary action. This results in very different results as the phase fractions behave like small water drops and nothing like the results in [37,118].…”

“…An issue with this is that at the 15 mm domain width and thus a 10 mm liquid fraction width as stated in [37,118] the surface tension force of bismuth (0.0378 N m −1 ) [108] is overwhelmingly strong and the phase fraction drags up the walls through essentially capillary action. This results in very different results as the phase fractions behave like small water drops and nothing like the results in [37,118]. This can be understood intuitively with what happens to the meniscus of water in a tube 10 mm wide and a tube 10 cm wide, and this is in a case where g = 9.81 m s −2 if g = 4.4145 × 10 −4 m s −2 the difference would be overwhelming.…”