Multi-input convolutional network for ultrafast simulation of field evolvement.

There is a compelling need for the regression capability of mapping the initial field and applied conditions to the evolved field, e.g., given current flow field and fluid properties predicting next-step flow field. Such a capability can provide a maximum to full substitute of a physics-based model, enabling fast simulation of various field evolvements. We propose a conceptually simple, lightweight, but powerful multi-input convolutional network (ConvNet), yNet, that merges multi-input signals by manipulating high-level encodings of field/image input. yNet can significantly reduce the model size compared with its ConvNet counterpart (e.g., to only one-tenth for main architecture of 38-layer depth) and is as much as six orders of magnitude faster than a physics-based model. yNet is applied for data-driven modeling of fluid dynamics, porosity evolution in sintering, stress field development, and grain growth. It consistently shows great extrapolative prediction beyond training datasets in terms of temporal ranges, spatial domains, and geometrical shapes.

Development of a surface tension mediated technique for dry stabilization of mammalian cells.

Dry state preservation at ambient temperatures (lyopreservation) is a biomimetic alternative to low temperature stabilization (cryopreservation) of biological materials. Lyopreservation is hypothesized to rely upon the creation of a glassy environment, which is commonly observed in desiccation-tolerant organisms. Non-uniformities in dried samples have been indicated as one of the reasons for instability in storage outcome. The current study presents a simple, fast, and uniform surface tension based technique that can be implemented for lyopreservation of mammalian cells. The technique involves withdrawing cells attached to rigid substrates to be submerged in a solution of lyoprotectant and then withdrawing the samples at a specific rate to an inert environment. This creates a uniform thin film of desiccated lyoprotectant due to sudden change of surface tension. The residual moisture contents at different locations in the desiccated film was quantified using a spatially resolved Raman microspectroscopy technique. Post-desiccation cellular viability and growth are quantified using fluorescent microscopy and dye exclusion assays. Cellular injury following desiccation is evaluated by bioenergetic quantification of metabolic functions using extracellular flux analysis and by a Raman microspectroscopic analysis of change in membrane structure. The technique developed here addresses an important bottleneck of lyoprocessing which requires the fast and uniform desiccation of cellular samples.

Metal pad instabilities in liquid metal batteries.

A mechanical analogy is used to analyze the interaction between the magnetic field, electric current, and deformation of interfaces in liquid metal batteries. In the framework of a low-mode, nondissipative, linear stability model, it is found that, during charging or discharging, a sufficiently large battery is prone to instabilities of two types. One is similar to the metal pad instability known to exist in the aluminum reduction cells. Another type is new. It is related to the destabilizing effect of the Lorentz force formed by the azimuthal magnetic field induced by the base current, and the current perturbations caused by the local variations of the thickness of the electrolyte layer.

Evolution of a spherical hydrate-free inclusion in a porous matrix filled with methane hydrate.

The behavior of a small isolated hydrate-free inclusion (a gas bubble) within a porous matrix filled with methane hydrate and either water or methane gas is analyzed. Simplifying assumptions of spherical symmetry, an infinite uniform porous medium, and negligible effects of background temperature and pressure variations focus the investigation on the features of the dynamics of a single bubble determined by a phase transition. Two solutions are presented: an exact solution of the Stefan problem obtained when the effects of gas and water flow are neglected, and a numerical solution of the full problem. The solutions are in good agreement with each other and with known asymptotic dependencies, confirming that the effects of inertia and convection transport can be neglected in the case of small inclusions. It is found that, after an initial adjustment, the radius of any small bubble decreases with time following a self-similar solution of the Stefan problem. The lifetime of a bubble is evaluated as a function of initial radius and the system's physical parameters. Possible effects of such inclusions on the filtration of methane to the surface and other aspects of the dynamics of hydrate-bearing deposits are discussed.

Transition to two-dimensionality in magnetohydrodynamic turbulent Taylor-Couette flow.

Transition from a Taylor-Couette turbulent flow to a completely two-dimensional axisymmetric turbulent state is realized numerically by increasing gradually the strength of the azimuthal magnetic field produced by electric current flowing through the axial rod. With the increase of the Hartmann number, the Taylor-vortex-like structures shrink, move closer to the inner cylinder, and turn into unsteady but perfect tori at sufficiently high Hartmann numbers.

Patterned turbulence in liquid metal flow: computational reconstruction of the Hartmann experiment.

We present results of a numerical analysis of Hartmann's historical experiments on flows of mercury in pipes and ducts under the influence of magnetic fields. The computed critical parameters for the laminar-turbulent transition as well as the friction coefficients are in excellent agreement with Hartmann's data. The simulations provide a first detailed view of the flow structures that are experimentally inaccessible. Novel flow regimes with localized turbulent spots near the sidewalls parallel to the magnetic field and otherwise laminar flow are discovered. We finally suggest how these predictions can be tested in a transparent fluid using optical flow measurement.

Large-scale intermittency of liquid-metal channel flow in a magnetic field.

We predict a novel flow regime in liquid metals under the influence of a magnetic field. It is characterized by long periods of nearly steady, two-dimensional flow interrupted by violent three-dimensional bursts. Our prediction has been obtained from direct numerical simulations in a channel geometry at low magnetic Reynolds number and translates into physical parameters which are amenable to experimental verification under laboratory conditions. The new regime occurs in a wide range of parameters and may have implications for metallurgical applications.