Remember to press refresh on your browser if you have visited this page previously to ensure you seeing the latest version. Otherwise you may be viewing an older cached version.
08/04/2013 Over the last couple of weeks I have been running models 61 to 66. These are identical to models 51 to 56 except that they have corrected boundary conditions and the boundary is a little further out. As of today, only model 62 has not yet solved.
22/03/2013 Committed comi code to Launchpad Repository via Bazaar. Revision 21. (Hall Model 61)
19/03/2013 The plot autogenerator has been updated to create errB/B plots aswell as Bmod and Vector Plots. This has been tested on model 51 over the weekend and the results are now online. I will shortly try the plot autogenerator on the quad sub-models as it should now run on these with minimal tweaking. This should give some useful information for comparison with the mesh convergence runs.
I have also started solving Hall Model 61, this is pretty much identical to Hall Model 51 except that the boundary conditions have been corrected and they are now slightly further out. Boundary location is outer_air x 1.2 instead of x 1.02.
15/03/2013 Quad Model 05 - This model has been run to do a boundary comparison check with Quad Model 02. Quad Model 05 will be run with a Normal boundary condition, all other model details are identical to Quad Model 02.
14/03/2013 Please see the notes on the magnetic meeting from 12/03/2012 for details of first comparisons between the quad model 01, 02, 03 and model 51.
08/03/2013 The autogen code has been updated to generate vector plots aswell as Bmod plots. It is clear that the code will need some work to make it work with the new quad sub-models.
The Hall model has been updated so that the boundary conditions are now controlled in the comi code. This will allow us to re-run models 51-56 when the opportunity arises to see if this has fixed the problems we were observing at low field.
Committed comi code to Launchpad Repository via Bazaar. Revision 20. (Quad Sub Model 04)
07/03/2013 The first quad model compared favourably with the hall model in the vicinity of Q9, there was disagreement further up the beamline but this was understandable in the context that there was a lot of iron in the hall model to the east of Q7 that is missing in the Quad model. I was generally happy with the agreement along the beam axis.
The vector plots are also much improved with a true dipole shape being visible in the vector map even at low field level. The jury is out as to whether this is due to the change in the boundary conditions or due to the simpler model. I'm presuming the former but it will take a rerun of the hall model with the new boundary conditions to prove this.
A new variable has been added to the code that controls the meshing resolution so that the original mesh size is divided by the cube of this variable. In principle this gives a model that should have a solve time that is linearly proportional to the value of this variable. To try this out I'm now running 3 more quad models, with this variable set to 0(default mesh settings), 2 and 4. These are Quad Models 2, 3 and 4. The point of adding this variable is to make it easier to test a model for mesh convergence. A second default quad sub-model has been run as the quad sub-model air has been extended slightly.
04/03/2013 I have built the first sub-model which is intended as a model for looking more closely at the fields around the quads Q9-Q7. It is a pared down model of the MICE Hall model. Quad Sub-model 01 contains the NSW, SSW, Virostek Plates, TOF Plate, EMR, Q9 to Q7 with baseplates. This has been run at the same meshing resolution as model 51 so as to provide a direct comparison. The results in the vicinity of the quads should be comparable if the outer structures are irrelevant.
In building this model there are two issues worth pointing out:
1) A bug was found in the EMR code. A section of air was missing in the central region. Default behaviour of OPERA would have
been to fill this with air so this is of minor concern but it has been fixed.
2)The Quad sub model is in control of its boundary air volume. This means that the meshing resolution, potential type, etc, have now all
been fixed in the code.
22/02/2013 I have realised this morning that models 57/58 have not solved correctly. The reason they have not solved is that without the Virostek plates in the model, there is no reduced potential region around the coils. I now need to rerun models 57 and 58 with the Virostek (& TOF) Plates.
21/02/2013 Committed comi code to Launchpad Repository via Bazaar. Revision 19. (Model 58)
21/02/2013 It has become apparent that the model is not giving reliable results at low field levels.
It is clear that any field reported below ~1 gauss is likely to be FEA noise, although the exact point at which this happens is unclear.
To try and get a better handle on this noise limit and to try and understand whether the model is elsewhere limited
by meshing resolution I am going to run a series of model tests.
To do this I have added a few variables to the code:
1) #TightenConvergence. If set to 1 this increases the convergence criteria from 1E-08 to 1E-10.
This may improve results but will increase the solve time.
2) #TightMeshResolution. If set to 1 this will halve the meshing resolution.
I'm not sure at the moment whether this will give meshing issues but if it doesn't then it is expected that this
variable will signficantly increase the solve time.
3) #x(y)(z)_scale_factor. This defines the size of the outer model body in comparison to the outer air.
Previously it was fixed at 1.02 but I'm wondering if this value was too small as there is evidence of a non-tangential field at the model border.
These variables allow this ration to be easily altered. I'm going by increasing this value to 1.2.
For simplicity it is my intention is to run a step IV solenoid 240 Mev/c model with just the NSW and SSW
(edit 22/02/2013: also need Virostek Plates) in situ to see if these variables make any noticeable difference to the results that we are seeing.
04/02/2013 Model 56 has shown a solve time of 21 minutes which is clearly in error. The error occurred because the .comi could not find the 240MeV/c Step 6 flip mode conductor file. This file was incorrectly named and so the model ran with no conductors. The model is now being re-run with a correctly named conductor file.
Some of the models ran in earkt April are listed in the March section to keep the model numbers in order. (These were the models ran over the easter break as a batch.)
Model 66 - 08/04/2013
test model 66 op3 - solution file.
test model 66 - associated build files.
Solve Time = 52 hours.
Magnet Configuration: Step VI Flip Mode - 240MeV/c - No return yoke.
As model 56 but see notes for model 61.
Model 65 - 08/04/2013
test model 65 op3 - solution file.
test model 65 - associated build files.
Solve Time = 48 hours.
Magnet Configuration: Step IV Flip Mode - 240MeV/c - With return yoke.
As model 55 but see notes for model 61.
Model 64 - 08/04/2013
test model 64 op3 - solution file.
test model 64 - associated build files.
Solve Time = 59 hours.
Magnet Configuration: Step IV Flip Mode - 240MeV/c - No return yoke.
As model 54 but see notes for model 61.
Model 63 - 22/03/2013
test model 63 op3 - solution file.
test model 63 - associated build files.
Solve Time = 39 hours.
Magnet Configuration: Step VI Solenoid Mode - 240MeV/c - No return yoke.
As model 53 but see notes for model 61.
Model 62 - 08/04/2013 - Model Still Running Solution Not Yet Available
test model 62 op3 - solution file.
test model 62 - associated build files.
Solve Time = xx hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - With Holger's return yoke.
As model 52 but see notes for model 61.
Model 61 - 19/03/2013
test model 61 op3 - solution file.
test model 61 - associated build files.
Solve Time = 60 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
As model 51 but the boundary conditions have now been set as total potential. The boundary is now slightly further out than before. The boundary use to be a distance of outer_air x 1.02, it is now outer_air x 1.2.
Quad Sub Model 05 - 15/03/2013
Quad Sub Model 05 op3 - solution file.
Quad Sub Model 05 - associated build files.
Solve Time = 23 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
As Quad model 02 but with Normal boundary conditions.
Quad Sub Model 04 - 07/03/2013
Quad Sub Model 04 op3 - solution file.
Quad Sub Model 04 - associated build files.
Solve Time = 116 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
Quad model 02 with increased mesh resolution (mesh/4^(1/3)). The model is being run for mesh convergence testing.
08/03/2013 Committed comi code to Launchpad Repository via Bazaar. Revision 20. (Quad Sub Model 04)
Quad Sub Model 03 - 07/03/2013
Quad Sub Model 03 op3 - solution file.
Quad Sub Model 03 - associated build files.
Quad Sub Model 03 - comi snapshot.
Solve Time = 45 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
Quad model 02 with increased mesh resolution (mesh/2^(1/3)). The model is being run for mesh convergence testing.
Quad Sub Model 02 - 07/03/2013
Quad Sub Model 02 op3 - solution file.
Quad Sub Model 02 - associated build files.
Solve Time = 22 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
New Baseline model with slightly extended air boundary - shouldn't make much difference.
Quad Sub Model 01 - 04/03/2013
Quad Sub Model 01 op3 - solution file.
Quad Sub Model 01 - associated build files.
Solve Time = 14 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
Quad Sub-model 01 contains the NSW, SSW, Virostek Plates, TOF Plate, EMR, Q9 to Q7 with baseplates. This has been run at the same meshing resolution as model 51 so as to provide a direct comparison. The results in the vicinity of the quads should be comparable if the outer structures are irrelevant.
Model 60 - 21/02/2013
** MODEL TERMINATED DUE TO LONG SOLVE TIME - CP MODEL 59 **
** MAY RERUN THIS MODEL AT A LATER DATE **
test model 60 op3 - solution file.
test model 60 - associated build files.
Solve Time = xx hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
As model 58 but the meshing resolution has been doubled.
Model 59 - 21/02/2013
test model 59 op3 - solution file.
test model 59 - associated build files.
Solve Time = 166 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
As model 57 but the meshing resolution has been doubled.
21/02/2013 Committed comi code to Launchpad Repository via Bazaar. Revision 19. (Model 58)
Model 58 - 21/02/2013
test model 58 op3 - solution file.
test model 58 - associated build files.
Solve Time = 29 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
As model 57 but the convergence variable has been tightened from 1E-08 to 1E-10.
Model 57 - 21/02/2013
test model 57 op3 - solution file.
test model 57 - associated build files.
Solve Time = 24 hours.
Magnet Configuration: Step IV Solenoid Mode - 240MeV/c - No return yoke.
This is the baseline model upon which I will base a comparison for the next few models. The model incorporates the solenoids, the NSW and the SSW only. It is being run at the standard meshing resolution and the standard solve convergence.