Alison Kubota's Technical and Data Analytics Site: 02/07/09

This is a repository for some of my work. As an image can convey a thousand words (more like a million), I would be happy to explain to you the any aspects of the work, including the operational aspects (when, where and how) such as the vertical consolidation of the projects and planning, as well as the technical details (what, why) that went behind these works that I am so proud to present to you. I developed all the algorithms and codes that went into the analysis and render pre-processing. By thinking carefully about the number of free dimensions you have to play with (generally alot more than 2), you can condense large and complicated data sets into meaningful representations that tell different stories.

These images were from different projects that I had developed and executed on supercomputers. The visualization aspect of the work adds to the projects themselves. My philosophy is to represent very extensive and complex data-sets in a both natural yet novel ways that tell complete stories of what is occurring in the data analysis. The brain is capable of handling very complex processes and pattern, and so the imagery from the visualization of high-dimensional data also allows us to detect these patterns and processes that may be hidden beneath the large expanse and noise of data.

There are many quantitative methods available to analyzing high-dimensional and time-series data. I have focused mainly on modeling and application of many of these methods to extensive data sets we produce from computers simulation. These methods are none the less applicable to a wide range of problems.

For anyone inquiring about this repository through my CV, you can contact me at the email or telephone listed on my hardcopy CV. Most of my files are over 200MB large, and are too large to be transmitted by email. These animations were designed to be displayed on a high-resolution high pixel count wall-sized "powerwall" system that should be the center-point of all corporate boardrooms. The animations here (and the ones I can fit on blank DVDs) are at much lower resolutions, since most monitor arrangements aren't large enough to display them, and most computers cannot handle the extremely high bandwidth.

Complex Inverse Problems

Hydrodynamic Instabilities Using Atomistic Simulations on Supercomputers: Richtmyer-Meshkov, Rayleigh-Taylor and Late-Time Kelvin-Helmholz Flavors

Here are some of the chump eye-candy movies on hydrodynamic instabilities.

Shockwave Deformation of Materials That Have Undergone Aging

This pdf contains a few frames that are part of a "power-wall" animation of a very large simulation consisting of over 130 million atoms. The calculations, post-processing, analyses and visualizations were done on very short notice and required 2 weeks straight of all-night work (periodic naps). There is a long story and presentation associated with the movie. Unfortunately the animation is up to 5 GB large, so I can only provide snapshots here. If you want the full animation, I can send it to you if you contact me through the email listed on my hardcopy resume.

111movieframes.pdf

Mapping of Atomistic Data to a Continuum Formalism.

One of the challenges associated with large-scale atomistic simulation is the extensive amount of data that is produced that require analysis. Simple filters to produce alternate or simplified representations can produce imagery that is often beautiful and provide useful insight.

Coarse-graining has been a technique used often in the atomistic fields in order to gain some idea of local properties of density, temperature, stresses, and so on. Coarse-graining is a nonparametric approach to estimating spatial and temporal distributions from these simulations, with the added constraints of required consistency with continuum mechanics. By performing this nonparametric density estimation on extensive data-sets, we reduce the dimensionality of the system (from 6N where N is in the billions) to something much more computationally tractable and useful for developing the underlying constitutive models that govern the behavior of large atomistic data sets.

This is an example of one such simulation results that employed nonparaametric density estimation of the longitudinal stress and strain rate. Strain has two contributions: elastic and plastic strain, though it is the rate of the strain contributions that are additive. The density estimation into a continuum form such as this are being used to provide insight as well as develop quantitative plasticity models for high strain-rate deformation (associated with shockwaves, detonations, and so on). This type of approach to develop complete single-crystal constitutive models in a bottom-up manner is quite unprecedented.

There are many different techniques that can be applied for nonparametric density estimation in the most general sense. In our cases, we use a simple non-continuous cubical window function as the kernel. However, we must remember the additional constraints that the estimations must obey continuum consistency, that that work on the system easily known atomistically must be the same as the continuum work based on the stresses and strain rates.

Dislocations Dynamics Under Dynamic Transitory Loads

This is an interesting representation of dislocations and stacking faults in lattices. Hidden are all the lattices that are not ear the dislocation. By using a common neighbor analysis treatment for the atoms in the perturbed lattice, you can pull out the atoms that are part of the dislocation. In this case, the green spheres. Then, the white surfaces are simple transparent triangles created by triangulation of atoms that are part of stacking faults - these atoms have a topology different than that of the face centered cubic lattice. After identifying these atoms, I wrote an algorithm that identified what the triangulations need to be in order to produce a surface that corresponds to the stacking fault itself. Using this treatment, you can get these very beautiful surfaces. The animations are even prettier.

Combined Stochastic and Deterministic Integration Methods to Model the Long-Time Evolution of Carbon Nanoparticles in Reactive Hydrocarbon Environments

To be added.

Radiation Damage in Plutonium and Other Materials.

This is an interesting treatment of algorithms to identify defects within the lattice, very similar to that used for the dislocations example above. The red and transparent blue spheres represent two types of defects. In addition, I have added a time-sequence of a high-energy tree - the branching structure representing the billiard ball like collisions produced by a single high-energy atomic collision. To do this, I filtered out all the atoms in all reported timesteps for which the atom had > 5eV of energy. Then I applied a sort algorithm to create a "worm" representing the path of each unique atom that I had filtered out. The coloration or hue corresponds to the energy of the atom. Then I sequentially rendered frames, letting the tree grow out with each image frame. What you get is the growing and branching tree produced by the collisions.

What one could do is have rendered, only the current and previous 10 frames, for example. This will create an effect of producing worm-like structures that travel around and multiply. This worm representation will allow you to see what is left at the core of the collisions. The full tree representation somewhat hides what was at the core of the cascade.

In the movie associated with the upper figure, I included a pan-and-zoom effect.

Repository of Invited Seminars and Lectures

Copies of invited seminars and lectures that I have given in the past. These are presentations that are targeted towards professionals. I have not included regular submitted presentations, such as conference submitted talks, group meeting presentations, and other informal non-invite type lectures.

To be included once I find a way to compress them to a small enough size.

Symposium Organizer, “Multiscale Modeling of Materials,” Materials Research Society Conference Fall Meeting, Boston, MA, 2006.
Invited Seminar, “Understanding the Atomic-Scale Aspects of Radiation Damage and Aging in Pu and Pu/Ga Alloys Through Computer Simulations,” Lawrence Berkeley National Laboratory, Berkeley, CA, 2006.
Guest Speaker, “Developments in Atomistic and Molecular Dynamics Simulations for Plasticity,” International Conference on Plasticity, Halifax, Nova Scotia, 2006.
Guest Speaker, “High Pressure Strength and Connecting Large Scale Atomistic Simulations of Shockwaves with the Continuum Scales,” International Conference on New Models and Hydrocodes for Shock Wave Processes in Condensed Matter, Dijon, France, 2006.
Invited Seminar, “High Pressure Strength via Internal Field Variable Analyses from Molecular Dynamics Simulations,” Sandia National Laboratory, Albuquerque, NM, 2006.
Workshop Organizer, “LLNL Summer Materials Research Institute,” Livermore, CA, 2005.
Invited Seminars, “Recent Advances in Connecting Classical Molecular Dynamics Simulations of Shockwaves with Reality,”
- Institute for Shock Physics, Washington State University, WA, 2005.
- Sandia National Laboratory, Albuquerque, NM, 2005.
- Department of Mechanical Engineering, Johns Hopkins University, Baltimore, MD, 2006.
- Department of Mechanical Engineering, University of Maryland, College Park, MD, 2006.
Invited Seminar, “Dynamic Strength by All-Atom Methods,” Department of Nuclear Engineering, Massachusetts Institute of Technology. June 11, 2007
Guest Speaker, “Single-Crystal Plasticity in Ramp- and Cyclically-Loaded Aluminum,” American Physical Society Conference on Shock Compression of Condensed Matter, Waikoloa, HI. June 28, 2007.
Directorate-Wide Seminar, “Wave propagation from the point of view of atomistic simulations,” Physics and Advanced Technologies Directorate, LLNL. July 24, 2007.
Gordon Conferences - Invited Seminar, Summer, 2008.

Repository of Peer-Reviewed Published Journal Articles

Alison Kubota, Wilhelm G. Wolfer, Steven M. Valone and Michael I. Baskes, "Modeling of aging in plutonium by molecular dynamics," J. Comp.-Aid. Matl. Design 14 (2007) 367.
Wilhelm G. Wolfer, Alison Kubota, Per Soderlind, Alex I. Landa, Benoit Oudot, Babak Sadigh, Jess B. Sturgeon and Michael P. Surh, “Density changes in Ga-stabilized delta Pu, and what they mean,” J. Alloy. Compd. (2007) to be published.
Alison Kubota, David B. Reisman and Wilhelm G. Wolfer, “Dynamic strength of metals in shock deformation,” Appl. Phys. Lett. 88 (2006) 241924.
Alison Kubota and W. G. Wolfer, “Transition pathways in the unfaulting of dislocation loops,” Mat. Sci. Eng. A, 400-401, (2005) 362-365.
F. Mota, M.-J. Caturla, J. M. Perlado, E. Dominguez and Alison Kubota, “Threshold energy of formation of an oxygen vacancy defect in SiO2 by atomic displacements using molecular dynamics,” Fusion Energy Design 75-9 (2005) 1027.
F. Mota, M.-J. Caturla, J. M. Perlado, E. Dominguez and Alison Kubota, “Atomistic simulations of threshold displacement energies in SiO2,” J. Nucl. Mater., 329-33 (2004) 1190-1193.
Michael J. Fluss, Brian D. Wirth, Mark Wall, Thomas E. Felter, Maria-Jose Caturla, Alison Kubota and Tomas Diaz de la Rubia, “Temperature-dependent defect properties from ion-irradiation in Pu(Ga),” J. Alloy. Compd., 368 (2004) 62.
Angela Violi, Alison Kubota, William J. Pitz, Charles K. Westbrook and Adel Sarofim, “A Fully-Integrated Molecular Dynamics - Kinetic Monte Carlo Approach for the Simulation of the Growth of Soot Precursors,” P. Comb. Inst. 29 (2003) 2343.
Alison Kubota, Maria-Jose Caturla, James Stolken, Babak Sadigh, Susana Reyes, Tomas Diaz de la Rubia and Jeffery F. Latkowski, “Threats to ICF reactor materials: computational simulations of radiation damage induced topological changes in fused silica,” Nucl. Instrum. Meth. Phys. Res. B 202 (2003) 88.
Jeffery F. Latkowski, Alison Kubota, Maria-Jose Caturla, Sham N. Dixit, Joel A. Speth and Stephen A. Payne, “Fused silica optics for inertial fusion energy: Radiation studies and system-level analysis,” Fusion Sci. Tech. 43 (2003) 540.
Thomas E. Felter, Lawrence Hrubesh, Michael Fluss, Alison Kubota, Lilian Davila, Maria- Jose Caturla and Tomas Diaz de la Rubia, “Laser damage threshold studies of fused silica modified by MeV ion implantation,” Nuclear Instr. Meth. Phys. Res. B, 207 (2003) 72.
Lilian P. Davila, Maria-Jose Caturla, Alison Kubota, Babak Sadigh, Tomas Diaz de la Rubia, J. F. Shakelford, S. H. Risbud and S. H. Garofalini, “Transformations in the mediumrange order of fused silica under high pressure,” Physical Review Letters 91 (2003) 205501.
Alison Kubota, Maria-Jose Caturla, Stephen A. Payne, Tomas Diaz de la Rubia and Jeffery F. Latkowski, “Modeling defect production in silica glass due to energetic recoils using molecular dynamics simulations,” Journal of Nuclear Materials, 307-311 (2002) 891.
Alison Kubota, Maria-Jose Caturla, James S. Stolken and Michael D. Feit, “Densification of fused silica due to shock waves and its implications to 3ω laser damage” Optics Express, 8 (2001) 611.
Alison Kubota and Demetre J. Economou, “Molecular dynamics simulations of ion-induced rearrangement of ultrathin oxide films on silicon,” IEEE Trans. Plasma Sci. 27 (1999) 106.
Alison Kubota and Demetre J. Economou, “A molecular dynamics simulation of ultrathin oxide
films on silicon: growth by thermal O atoms and sputtering by 100 eV Ar ions,” IEEE Trans. Plasma
Sci. 27 (1999) 1416.
Demetre J. Economou, Richard S. Wise and Alison Kubota, “Plasmoid formation and multiple steady states in a low pressure, inductively coupled electronegative plasma,” IEEE Trans. Plasma Sci. 27 (1999) 60.
Chang-Koo Kim, Alison Kubota and Demetre J. Economou, “Molecular dynamic simulation of silicon surface smoothing by low-energy argon cluster impact,” J. Appl. Phys. 86, 6758 (1999).
Alison Kubota, Demetre J. Economou and S. J. Plimpton, “Molecular dynamics simulations of low-energy (25-200 eV) argon ion interactions with silicon surfaces: sputter yields and product formation pathways,” J. Appl. Phys. 83 (1998) 4055.