As stated in the header, I am a computational physicist in the Atlanta area. I started out as a “music major” at the local community college, but I realized that would just land me the job I already had of selling music instruments. One night, while hanging out with a friend, the following conversation was had
Me: You know, I was always good at physics in high school.
Friend: Hey! You should do that. It would be lucrative!
Me: You know what? I think I will!
Or something to that effect. The end results was that I woke up, thought about it a bit more, discussed it with my girlfriend (now wife), and got down to business. Now, I am in the final throws of wrapping up my PhD in physics at the Georgia Institute of Technology. My current research is on the effects of surface shape on enhanced near-field radiation. I am currently on the prowl for what will be my next step once I am done with school.
Activities and interests
As I mentioned above, I am a recovering professional sales man. There are times where I miss the excitement of the sales floor, but then I remember grind that was associated with it and go back to my research. I was a music major, and that gave me a chance to learn to play quite a few different instruments. I still enjoy pulling out the euphonium and the trombone, but I don’t get around to it that often. Most of my free time is spent in the back yard with the wife and pups. The little one really likes fetch!
As for the stuff you really care about, my current fascinations revolve around computing. I am trying to teach myself the intricacies of proper parallel programming; however, that is slow going because I do not own a super-computer with which to practice. I am also intrigued by heat transfer. There is an open problem out there on how to couple the radiative exchange in the thermal bands to the conductive modes of energy exchange numerically. I got time as an undergrad to play with a finite difference approach to the problem. The program was kludge of FORTRAN 77 and Fortran 90 (mostly 77) and had hard coded limits appropriate for the computers of the 1980s and 1990s. These limits are not trivial to change thanks to the use of common blocks. The radiative exchange is coupled to the conduction modes by the view factors known to computer graphics. The calculation of the view factors takes a lot of time and, for large scenes, does not account for long range interactions of hot objects. An object that is hot enough may well emit enough energy that it will reflect around to interact with a surface that is not immediately visible. The other issue with this approach deals with the thermal shadow. The large nodes necessary for the thermal network solver do not resolve the thermal shadow well. We need to figure out a method to determine the power delivered to a surface and the shadowing when only a fraction of the surface participates in the exchange.
Coupled with my thesis work, I want to investigate methods of solving linear systems where the matrix is a block matrix. There has to be a systematic method to solve this type of system. Specifically, I have a system where the matrix is comprised of 2 by 2 block elements that are symmetric up to a minus sign. This symmetry in how the elements are computed that is not being exploited. This means that swapping the indexes of the block simply gets you the symmetric block of the matrix. Currently, I am using a brute force method to solve the system. I would love for the chance to search for a more efficient method to solve the system.
My Work at GTRI
Currently, I am investigating the effects of shape on near-field radiative transfer. We all know from undergraduate physics that all objects are surrounded by a thermally generated electromagnetic field. At large separations, this is simply the black-body exchange predicted by statistical mechanics. When the separation between the objects is small enough, the surface reach a point where they are within the decay range of the evanescent waves trapped on the surfaces. We want to establish a long range structure that can direct the energy through a system of particles. This is useful for thermal control of low temperature experiments, micro and nano scaled structures, and it even applies to evolution of interstellar gasses if the literature can be believed. To do this, we need to understand what the effect of particle shape is.
My other current project is to develop a framework in Python to process hyper-spectral imagery. My task was to establish the development plan, layout the structure, and help guide the development of additional utilities. We decided to go with Python so that we could get out of the choke-hold of MATLAB. This is partly due to financial reasons and partly to memory limitations. We cannot process the highest resolution data cubes with in MATLAB because it cannot request enough memory to do the necessary work. Long term goals include real time processing of hyper-spectral data for in situ checking during field work and processing spectral data to predict broad band sensor response. The end goal is to use any means necessary to enhance performance including acceleration with graphic processing units.
A few other topics that I have studied are:
- Electromagnetic wave scattering from non-spherical particles. This is commonly known as either Mie scattering or the T-matrix approach. This topic came up because the formalism to establish the parameters of the matrix is very similar to my thesis work. I was looking for tricks to help me out.
- Studied the performance parameters of counter flow cooling towers to understand the energy emission under normal load. I no longer remember our original goals with this topic. I do remember that I hypothesized that you could predict the performance load by measuring the enthalpy change from the surrounding environment to the center of the plume. That failed.
- Vehicle dynamics. My very first task joining GTRI was to develop a first principles based vehicle dynamics model. The first version was a highly simple model written in MATLAB. Once I transitioned to a graduate student, I went back and expanded the model into a five degree of freedom, four wheel, untracked land vehicle model under normal conditions and hooked it into a MySQL database. The goal was to help with simulating visible and infrared imagery in complex urban environments.
That just about wraps it up for me. I hope you enjoy your stay!