Meeting update

Last week was largely slow due to various issues with Grass. I worked with Rich a lot to get Grass up and going again. Of the few bugs we fixed, we are now running Visit 2.5 which is not stalling anymore. Not all of the kinks are worked out, but at least grass is functioning now.

I compiled a reading list for myself of a few textbooks and papers last week and began reading. I am pretty excited about these 3 papers I found on adsabs:

  • Foster, Prudence N.; Boss, Alan P., Triggering Star Formation with Stellar Ejecta: "We examine inducing the self-gravitational collapse of molecular cloud cores with stellar ejecta. We study the effect of winds of various strengths arriving at cloud cores modeled as marginally stable Bonnor-Ebert spheres, which are unstable both to collapse and to expansion. We find that some winds instigate collapse of the cloud core, while others result in expansion or destruction of the cloud. Collapse occurs when the incident momentum of the ejecta is greater than approximately 0.1 MMsun0 km s-1 for the standard γ = 1 wind and 1 Msun cloud scenario. The critical momentum, which divides those cases which induce collapse and those which do not, scales as the mass of the cloud times its sound speed, which is 0.2 MMsun0 km s-1 for the standard to K cloud. The critical momentum is exceeded for some supernova and many protostellar outflows, although if the wind has a velocity greater than approximately 100 km s-1, the effective adiabatic index will be γ = 5/3 and the cloud will be destroyed, through shredding into many pieces. The planetary nebulae of AGB stars appear to have momenta below the critical value. However, we found that a high wind temperature (T ˜600 K), possibly characteristic of AGB star winds, could instigate collapse even in low momentum winds." ​http://adsabs.harvard.edu/abs/1996ApJ...468..784F
  • "Radiation hydrodynamics of triggered star formation: the effect of the diffuse radiation field". Haworth, Thomas J.; Harries, Tim J., ​http://adsabs.harvard.edu/abs/2012MNRAS.420..562H, "We investigate the effect of including diffuse field radiation when modelling the radiatively driven implosion of a Bonnor-Ebert sphere (BES). Radiation-hydrodynamical calculations are performed by using operator splitting to combine Monte Carlo photoionization with grid-based Eulerian hydrodynamics that includes self-gravity. It is found that the diffuse field has a significant effect on the nature of radiatively driven collapse which is strongly coupled to the strength of the driving shock that is established before impacting the BES. This can result in either slower or more rapid star formation than expected using the on-the-spot approximation depending on the distance of the BES from the source object. As well as directly compressing the BES, stronger shocks increase the thickness and density in the shell of accumulated material, which leads to short, strong, photoevaporative ejections that reinforce the compression whenever it slows. This happens particularly effectively when the diffuse field is included as rocket motion is induced over a larger area of the shell surface. The formation and evolution of 'elephant trunks' via instability is also found to vary significantly when the diffuse field is included. Since the perturbations that seed instabilities are smeared out elephant trunks form less readily and, once formed, are exposed to enhanced thermal compression."

Last week got code up and running on grass and BH. I re-ran a stability check on the B&P sphere for 5 crossing times. I wasn't able to make a movie over my remote connection, for an error I don't yet understand:

-visit is waiting for a cli to launch on local host VisIt could not connect to the new client /usr/local/alt/visit2_5_0.linux-x86_64/bin/visit

But, the sphere did appear to be stable. It oscillated about its equilibrium, although it expanded much more than condensed. Not sure if this is significant or unexpected…

Next, wanted to tidy up the runs I worked on over the summer, by adding more sophisticated refinement criteria, and running out until a sink formed. I learned about the refinement objects and added the 2 criteria to my problem module: 1) refine on density gradients, and 2) refine on Jean's threshold. I didn't see any apparent conflict in the code that would prevent multiple flags to be called… is this the case?

When I tried to run the B&P matched case, however, the code froze on grass after initialing only the first 3/5 grids. It just stopped printing lines, although top showed it was still computing… I tried it in the debug queue on BH, but it seg-faulted. Not entirely sure why changing the ambient density from very light to heavy would change the computation, other than the Jean's refinement may be triggered and the entire domain is being refined. The way my data file was set up though, I thought would restrict Jeans based refinement to happen only for higher levels. . .

Lastly, I wanted to check before running large sims if restarts with self-gravity are working okay with the golden version. It seems though, that there is a bug with them (posted a ticket). This makes long jobs a little risky to run. .

Comments

No comments.