Restart with New Geometry!

After a lot of work, we are starting with a new geometry. In my previous lab note, I pointed out some geometry-related issues that I was having with the newest geometry. All of those issues have been overcome, and I fixed a few other minor problems along the way.

During testing, I found that some of the transport quantities (diffusion, thermal conductivity, viscosity) were not using the most accurate estimates. Everything was within an order-of-magnitude, so the results obtained thus far are not useless; but they should be considered more quantitative. The new model uses well-established expressions to calculate these quantities based on first-principle theories.

I've also updated the code to include scripts that automatically generate visualization files and stop AWS servers after the computation is complete. This should save time by creating visualizations from the data without the need for me to download data files. It also will save a bit of money because the AWS server will shutdown immediately after the computation is complete.

I've completed a few test runs of the new geometry and code, including on the AWS server. Here is a visualization of the result.

We currently have 4 servers running, and I expect to add more in the next few weeks with slight differences to the geometries. The new geometries are large enough that we need to use a slightly larger server, so we've upgraded to the AWS r5.large, a memory optimized sever with 16 Gb of RAM. In testing, I assured that this memory was almost completely utilized (on average about to 75%).

 

These new simulations should add valuable information regarding the dynamics of hyperpolarized xenon through these setups.