Summary of DIVIMP Task Items:
This is an on-going list of work items - they are not listed in priority order.
- First Priority - should be done as soon as possible
- Second - short time frame - within a few days preferably
- Third - within a week to 10 days preferably
- Fourth - within a month
- Fifth - important but not yet scheduled
- APS poster - CMOD related.
Modeling the OUTER SOL for matching of horizontal and vertical reciprocating
probes. In addition, looking at matching other experimental data. The following are items
that need to be looked at:
- Outer SOL modeled plasma values compared to Horizontal and Vertical fast
The Dalpha signal is one large peak - corresponding to an intense
emission region in the private flux region - at or directly below
the X-point (with additional contributions from each target). Such a
radiating region can be seen in the 2D Dgamma plots. This feature
appears to dominate the target emission which from the
DIVIMP results at the moment is too low compared to
the C-top Dalpha by a factor of 2 or more and does not fill in or
peak in the valley as is seen in experiment.
- Dalpha/Dgamma ratio
I think this was extracted by Steve by integrating over the 2D Dgamma
data along the same LOS as the Dalpha. The current DIVIMP result
fails to replicate the sloping curve seen by experiment and instead
generates a very low Dalpha/Dgamma ratio in the private plasma
region that would normally be considered indicative of detachment.
We would need better plasma values in both the PP and the inner
target area to address this.
- 2D Dgamma
This 2D distribution reconstructed from a toroidal camera view
looks interesting. However, there is a large emission band near the
inner target which is missing from the current DIVIMP result - this
will require a better prescription/model of the inner target
to address. In addition, there is the radiating region near the
Xpoint that needs to be addressed for all of the spectroscopic
- Pressure gauge measurements
These are calculated by Eirene and according to Steve there are
two main issues at present - the CMOD plenum gauge pressure
calculated by Eirene appears to be too high - so high in fact that
particles eventually become almost perpetually trapped and thus
the code does not finish. Addressing this is an Eirene development
item that Steve may get to some time. In addition, because of the
location of the plenum entrance, and the CMOD geometry - both the
private plasma and inner target conditions will have a signicant
contribution to the pressure calculation. A consistent inner target
and private plasma are virtually required for reasonable modeling of
the pressure gauge results (at least this is what I gathered from
talking to Steve.)
- APS poster - DIIID related (Steve)
- C-Mod plenum neutral-neutral modeling:
Well, this is really a DIII-D issue, but since I am working on it it is listed here. Several problems with the n-n void modeling code have been identified and repaired. However, steady state modeling of the C-Mod plenum is still not possible due to a rapid pressure rise in the plenum. After 2 iterations, the pressure gets up to 2 Torr or so, causing EIRENE to spend most of its time following neutrals in the plenum (time spent modeling a single particle track can become longer than the total designated EIRENE run time). This rise in pressure does not seem reasonable at the moment, especially since the pressure in the plenum after the first iteration is only 50 mTorr or so, so that n-n collision MFP's are still longer than the dimensions of the plenum at that point.
- Review of n-n modeling procedure with Detlev
Once things are cleared up with respect to the C-Mod modeling (if possible) then I will go over the entire n-n modeling effort with Detlev. He has also suggested some DIII-D scenarios that he would like to see results for.
- DIII-D PFZ pressure gauge modeling
The cases have been setup and most have been run. I am waiting until the C-Mod n-n modeling issues have been resolved before proceeding in earnest. The only unexpected result so far is that a good deal of the noise in the pressure gauge calculation goes away when the plasma solution is kept constant, as opposed to iterating with a converged solution. The reason for this is not apparent at the moment, since variation in the plasma quantities is not obvious after convergence.
It is also now possible to puff into the PFZ vacuum grid, and the effect of this (as opposed to puffing in the PFZ in the absence of the grid) has not been explored yet. Perhaps a small puff will have more of an effect.
Also, the effect of varying the PFZ plasma has to be looked at in more detail.
As suggested by Peter, we can also apply Gary's (and now Larry's?) fluid solution to the inside, although Gary does not get a good match to Dalpha on the inside either. This may give a better estimate of the contribution from the inside to the PFZ pressure.
- DIII-D plenum gauge modeling
The plenum pressure is being calculated for four locations of the outer strike point, and this has already been done without a plenum vacuum grid. Presently, the vacumm grid has been developed for the plenum and these cases are to be re-run on the grid.
Noteworthy is that the agreement with experiment decreases the longer the time between the pressure measurement and the reference time (1650 ms). Although it is now accepted that the target conditions should be considered to be constant, I am wondering if the corrective target data shift that is applied at 1650 ms is also appropriate for the other times. It is straighforward to vary this and see how it affects the plenum pressure at the other strike point locations. A good check would be to compile DTS data for the other times and see what shift gives the best agreement between OSM and DTS. Not sure how much work it would be to get the data and do the mappings.
- Vacuum grid development (Steve)
- Vacuum grid 2D plots (almost done).
- Remove the restriction on the vacuum grid that grid lines have to intersect verticies on the standard (magnetic) grid. This will require writing a search routine for EIRENE so that the location in the vacumm grid can be found when a particle leaves the standard grid. (Not to be done before the APS.)
- Add vacuum grid capability to the main DIVIMP version.
- Provide on-going help for Taekyun's DIVIMP modeling of CMOD
- Find a second DIIID shot and time slice for modeling.
- collect and convert experimental data for the new shot
- Run introductory set of cases
- Collect heat flux and power source data in the output file
- Develop OUT code to model other diagnostics
- Midplane Probe
- Bolometers - Working from actual diagnostic geometry.
- 3D integration for direct comparison to toroidal camera views.
- UEDGE-OSM code-code comparisons.
- General Comparison Exercise
- Do a variant of the OSM-UEDGE comparison where we
take ALL of thevolumetric terms, for particles, momentum,
e- and i-energy, from UEDGE to use in OSM.
- Assign the 2D OSM P_cool_total (H+C) term based on the 2D expt bolo P_rad
- Develop simple plasma-wetted wall sputtering option
- Turn SOL23 into user-friendly, work-horse, main-line OSM.
- SOL24 and SOL24A
SOL24 is a method of specifying the plasma that has several free parameters that
define the ionization front and the size of the ionization zone. It then iteratively
solves the background one flux tube at a time obtaining a set of paramters on each
flux tube that are consistent with the source and sink assumptions in the prescription.
SOL22 is used at various stages as part of the solution process. SOL24 seems to mostly
involve constraining the sources and sinks that are passed to the standard solver.
For more details please see TN1519.
- Integrate into main code version. Currently it is only in Steve's thesis version of the code
- Revise option 24 to include option 24A as per TN1519 - this may involve a new SOL option as opposed to a revision of 24.
Use Thomson data over whatever spatial region it is available
and then blend this into a 2D model like SOL22 or an alternate
prescription at the edges so that a complete background plasma is
Multi-region SOL option that utilizes the DIIID divertor Thomson data to aid in
defining the background plasma. For more details see TN1514.
- Add hydrocarbon break-up option.
- Incorporate rate calculation code.
- Add a module for following Hydrocarbon transport including wall interactions.
- Add grid extension to the wall.
- Extract the 2D electron cooling term, P_cool from UEDGE
If all that's
available is P_rad - sum of hydrogenic plus impurity - then one has to
add the product of the ionization rate of hydrogen times the ionization
potential energy, 13.6 eV, plus same for all the charge states of carbon; I
hope that P_cool itself is available!. We then want to use that in 2 ways:
- plot it on Along Ring plots along with the UEDGE Te so we can try to
confirm that the sudden plunge in Te in front of the target is due to
P_cool concentrated there.
- As the volumetric loss term in the electron energy eqn in OSM.
- TN1518: calculating the line shape of CI line in DIVIMP
- EIRENE to UEDGE Neutral model comparison.
We need to run EIRENE on the UEDGE solution and compare the neutral
densities, ionization patterns with what UEDGE gets. I'm not sure whether
to call this critical-path stuff re PSI/EPS or not. What's certain is that
it's pretty critical for having us come across around here, and with Gary
and Livermore, that we are doing useful things. So, I think we need to give
this some priority.
- Code development items related to getting the current version of
DIVIMP/Eirene to work for Asdex Upgrade - Karl's visit from
Feb 21st to March 3rd 2000
- Changes to Steven's thesis code version to support Asdex Upgrade inner wall plasma contact
- EIRENE clean up and completion of code integration (Steve)
Turn SOL23 into user-friendly, work-horse, main-line OSM.
EIRENE output file printout (Steve)
- option to include recombination power terms in Qe and Qi,
- look at n-n mfp to see if void grid is required in
pressure gauge regions - look at Todd Evans grid generator
code if necessary
SOL24: Application to the inner target for C-Mod shot 990429019 (Steve)
Post to the Fusion Group Publications web page.
Several low priority tasks: (Steve)
- Print EIRENE output numbers for fluxes to entire torus, in [#/s] and [A].
- Reorganize the output table.