Finish initial arfc draft of Moltres introduction manuscript

[lindsayad]

Tasks

• [x] Decide whether we want to submit a JOSS paper first
• [x] Adding some comparison with expected results will be necessary. Ideally, MSRE or 1-D analytical solutions. Are the fluxes the appropriate magnitude? How correct is the peak/average flux ratio? Is the temperature profile as was seen in MSRE?
• [x] Performance analysis. It runs fast, but how fast? Is it appropriate to do a scaling study?
• [x] Add description of how the group constants were generated for this particular simulation.

More to be added through comments.

Tag @katyhuff

[katyhuff]

Adding some comparison with expected results will be necessary. Ideally, MSRE or 1-D analytical solutions. Are the fluxes the appropriate magnitude? How correct is the peak/average flux ratio? Is the temperature profile as was seen in MSRE?

[katyhuff]

Maybe also some more performance analysis detail.

[lindsayad]

What are you thinking about JOSS?

[katyhuff]

We should definitely publish it in JOSS. Needs some attributes to be qualified. This is how we evaluate them: http://joss.theoj.org/about#reviewer_guidelines .

So, tests, docs, install instructions, etc.

On Thu, May 18, 2017, 2:30 PM Alex Lindsay notifications@github.com wrote:

What are you thinking about JOSS?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/arfc/publications/issues/6#issuecomment-302547272, or mute the thread https://github.com/notifications/unsubscribe-auth/AAYCq_RFWMWWaiJqyLfahITSYe8lQtC6ks5r7LhzgaJpZM4NYZcB .

[lindsayad]

Nestor, MSRE preliminary physics report (calculations):

• 8% fuel salt volume fraction (final design 22.5 %)
• 1 mole% UF4
• 93% U235 enrichment
• average thermal flux = 2.9x10^13 nts/cm^2
• peak thermal flux = 7.3x10^13 nts/cm^2
• ratio = 2.5
• average core power density = 4 W/cm^3
• peak core power density = 10 W/cm^3

Ours:

• 22.5% fuel salt volume fraction
• .9 mole% UF4
• 33% U235 enrichment
• average thermal flux = 1.0x10^13 nts/cm^2
• peak thermal flux = 2.3x10^13 nts/cm^2
• ratio = 2.3
• peak core power density = 17 W/cm^3

Really finding very little experimental data, but it must exist

[lindsayad]

@katyhuff What would a scaling study look like?

[katyhuff]

Typically, I see the following:

A very simple problem (specific materials, geometry, number of timesteps) is run with with a certain number of degrees of freedom (e.g. mesh resolution). Then, you investigate how quickly the problem runs when you spread that problem across an increasing number of processors. You want to see a strong decrease in runtime as the number of processors increases.

Considering the things that might impact scaling, the modeler (you) will often choose a couple of different canonical small simple problems which demonstrate the bounds of scaling behavior. For example (just an example), if you know that the scaling is 'bad' for problems with asymmetric geometries, you might run your scaling study with two different problems -- one sphere and one complex/asymmetric structure. Then, you would plot the scaling results together, to communicate that you expect most problems will fall between these scaling behavioral trends.

Here are two good examples: http://www.sciencedirect.com/science/article/pii/S0927025611004204 http://www.sciencedirect.com/science/article/pii/S0022311512000165

[katyhuff]

you've finished this draft! thanks @lindsayad !!!