Hi Andreas,
Solving Poisson’s equation in the co-moving frame works great in most cases, but for relatively long bunches in the alpha
magnet there is an inaccuracy because due to the curved trajectories the particles will have a significant velocity in the rest frame. For long bunches with a small curvature radius this can lead to underestimating the magnetic field component of the space
charge. Typically this effect is ignored, but it might not be negligible in the specific situation that I try to simulate. I am looking for a way to evaluate the effect of electrostatic assumption.
Actually, I did simulations with CST PS which is full electrodynamic, so I can compare the results from CST PS with another
code using the electrostatic assumption. Which is pretty much what I did by comparing CST PS with TraceWin, and I found non negligible difference. The other difference between these two codes is that CST PS works in the laboratory frame, whereas TraceWin works
in the rest frame. To determine the effect of the electrostatic assumption only, I would need to do simulations in the same frame with and without this assumption. This is possible in CST PS, which has a full electrodynamic PIC solver, and an electrostatic
PIC solver. I am currently running my first simulations with the electrostatic PIC solver of CST PS. I can let you know about the results soon.
Anyway, I would like to benchmark CST PS with another full electrodynamic code.
I understand that developing the full electromagnetic solver in OPAL will take time. I wish you will find some additional
man power. I would be interested to participate in this effort, but this wouldn’t be compatible with the schedule of my thesis work.
Cheers,
Abel
De : Adelmann Andreas <andreas.adelmann AT psi.ch>
Envoyé : mardi 23 mai 2023 16:35
À : PIRES Abel 611230 <Abel.PIRES AT cea.fr>
Cc : opal AT lists.psi.ch
Objet : Re: [Opal] Questions about space charge in an alpha magnet with OPAL
Hi Abel,
Thank you for your answer, and sorry for the delay of my answer. I have been reading and discussing a bit about the simulations of space charge.
I attended IPAC 23 in Venise, and had the opportunity to discuss with one of your colleague during a poster session, he was presenting the poster
“OPAL and Future Directions Towards the Exascale Area”. Sadly I cannot remember his name. The fact that OPAL is massively parallel is exciting, since I have access to super calculator at TERATEC, such as Topaze.
For now, simulating CSR is not the priority for me. But I need a full electromagnetic solver, the electrostatic assumptions might be inappropriate
for an alpha magnet.
it would indeed be very helpful to know why the ES assumptions is breaking down.
I am glad that you consider this as a very interesting project. Do you have plans to adapt the full electromagnetic solver to other elements, including
imported magnetic field maps ? I can only guess it would be difficult and need an important amount of work.
Without more additional man power what will take in minimum 2 years.
Hi Abel at the moment coherent synchrotron radiation is done by using a analytic formulas
colleagues simulated the magnet this could be a solution but of course only approximative.
Adapting to the alpha magnet will be a major but very interesting project.
Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Head a.i. Labor for Scientific Computing and Modelling
Paul Scherrer Institut OHSA/D09 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09
-------------------------------------------------------
Friday: ETH HPK G 28 +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge,
the more profound will be its consequences.
Leon Lederman
============================================
Dear Andreas and Christof,
Thank you for your answers. I am happy to learn that it is possible, and that it has already be done !
Are you in contact with Simon Friederich, or should I contact him by myself ? I would be very happy to know how he worked it out, and if he used
a cyclotron element or ring element, or something else.
I can provide some more detail. Here is a link for a publication concerning the installation of the alpha magnets :
P. Guimbal et al., “Status of the ELSA-2 Project,” in Proc. EPAC’02, Paris, France, Jun. 2002, paper TUPRI066, pp.1768-1770.
I am also joining some other figures :
fig_1 is the layout of the first part of the accelerator, until the first alpha magnet. It is a simulation run with TraceWIN/PARTRAN. The input
main beam parameters are included in the screenshot. From the input to GAP8, the beam goes from 2 MeV to 17 MeV.
fig_2 is the layout again, with the emittances. One can notice a large emittance growth for the Y axis due to the transport through the alpha
magnet, while the X axis emittance is conserved.
fig_3 is the magnetic field produced by the alpha magnets
fig_4 is the electric field produced by the bunch, in the middle of the alpha magnets. Simulation done with CST PS.
I can provide the numerical values computed by TraceWin partran for the 6x6 matrix of the alpha magnet
At the moment, we did simulation using CST PS and TraceWIN/PARTRAN. I found non-negligible differences for the results of these two codes. Basically,
we found more important emittance growth for the Y axis using CST PS. We submitted a paper for peer review for IPAC 2023 with these results, I should be able to communicate it soon.
As you said, we want to investigate the effects of the small bending radius. And we would like to investigate synchrotron radiation in a second phase.
Do you think it would be possible with Opal ? Which version of Opal would be the best fitted for that ?
Simon Friederich from University of Mainz has done simulations of an alpha magnet using Opal-T. He could possibly provide you sample files for this.
Dear Abel sorry for the delay in answering.
Indeed this is an interesting problem, maybe even a challenge!
OPAL does not have a predefined alpha magnet. However if a field map could read
I would guess that there is a need to model synchrotron radiation (maybe in a second phase) and what certainly needs
to be addressed is the small bending radius. Our field solver uses a Cartesian grid, we need to understand the implication
Overall I am certain that OPAL can be used/updated to solve that problem, however some modification are maybe required order.
Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Head a.i. Labor for Scientific Computing and Modelling
Paul Scherrer Institut OHSA/D09 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09
-------------------------------------------------------
Friday: ETH HPK G 28 +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge,
the more profound will be its consequences.
Leon Lederman
============================================
I am contacting you because I am interested by your code. I am working at CEA in France, close to Paris. I work on a 30 MeV electron LINAC. After first acceleration
stage, the beam reach 17 MeV and then travels through two alpha magnets. With appropriate chirp, we can compress beam from 30ps rms to 7ps rms. At high charge (>1nC) there is large emittance growth within the alpha magnets.
We want to simulate properly the beam transport through this magnetic element, which bend the beam trajectory into the shape of a alpha letter. I am joining an
image to illustrate.
A good simulation of the transport through alpha magnet is necessary, we could use field maps in the laboratory frame, or the alpha transfer matrice. Then a good
simulation of space charge effects is necessary, with the least assumptions.
I have three questions, if you could answer them that would help me decide if I should use your code for these simulations.
Do you know if your code would enable such simulations ?
What are the assumptions done in your code that could be inappropriate for this situation ?
What are the strength of your code regarding this simulation ?
If you have any questions or need more details don’t hesitate to contact me.
<fig_1.JPG><fig_2.png><fig_3.png><fig_4.png><fig_5.png>
|
