1 March 2006 in Journal of Computational Physics
1 June 2015 in International Journal of Heat and Mass Transfer
9 October 2015 in Soft Matter
In this paper, the self-propelled motion of Leidenfrost droplets on ratchet surfaces is numerically investigated using a thermal multiphase lattice Boltzmann model with liquid–vapor phase change. The capability of the model for simulating evaporation is validated via the D2 law. Using the model, we first study the performances of Leidenfrost droplets on horizontal ratchet surfaces. It is numerically shown that the motion of self-propelled Leidenfrost droplets on ratchet surfaces is owing to the asymmetry of the ratchets and the vapor flows beneath the droplets. It is found that the Leidenfrost droplets move in the direction toward the slowly inclined side from the ratchet peaks, which agrees with the direction of droplet motion in experiments [Linke et al., Phys. Rev. Lett., 2006, 96, 154502]. Moreover, the influences of the ratchet aspect ratio are investigated. For the considered ratchet surfaces, a critical value of the ratchet aspect ratio is approximately found, which corresponds to the maximum droplet moving velocity. Furthermore, the processes that the Leidenfrost droplets climb uphill on inclined ratchet surfaces are also studied. Numerical results show that the maximum inclination angle at which a Leidenfrost droplet can still climb uphill successfully is affected by the initial radius of the droplet.
1 October 2014 in Journal of Computational Physics
We present simulations of the implosion of a dense shell in two-dimensional (2D) spherical and cylindrical geometry performed with four different compressible, Eulerian codes: RAGE, FLASH, CASTRO, and PPM. We follow the growth of instabilities on the inner face of the dense shell. Three codes employed Cartesian grid geometry, and one (FLASH) employed polar grid geometry. While the codes are similar, they employ different advection algorithms, limiters, adaptive mesh refinement (AMR) schemes, and interface-preservation techniques. We find that the growth rate of the instability is largely insensitive to the choice of grid geometry or other implementation details specific to an individual code, provided the grid resolution is sufficiently fine. Overall, all simulations from different codes compare very well on the fine grids for which we tested them, though they show slight differences in small-scale mixing. Simulations produced by codes that explicitly limit numerical diffusion show a smaller amount of small-scale mixing than codes that do not. This difference is most prominent for low-mode perturbations where little instability finger interaction takes place, and less prominent for high- or multi-mode simulations where a great deal of interaction takes place, though it is still present. We present RAGE and FLASH simulations to quantify the initial perturbation amplitude to wavelength ratio at which metrics of mixing agree across codes, and find that bubble/spike amplitudes are converged for low-mode and high-mode simulations in which the perturbation amplitude is more than 1% and 5% of the wavelength of the perturbation, respectively. Other metrics of small-scale mixing depend on details of multi-fluid advection and do not converge between codes for the resolutions that were accessible
1 February 2009 in Journal of Computational Physics
24 June 2004 in International Journal for Numerical Methods in Fluids