Can elegant perform dynamic Aperture optimization ?

[shancai]

Can elegant perform dynamic Aperture optimization or tune scan?
If can, can post some examples?
Thanks a lot!

[michael_borland]

Shancai,

At APS we use elegant and a genetic optimization script for combined dynamic/momentum aperture optimization. However, it requires a linux cluster, since the individual runs take about 1 hour each. You can also perform driving term minimization to improve dynamic/momentum aperture, although I don't use this myself because I find the genetic optimization works better. (You can find out more about driving term calculation from the twiss_output command manual entry).

As for a tune scan, I assume you are again referring to dynamic aperture. You'd have to write a script that runs elegant repeatedly with different tunes.

I'll try to post a few examples in the next week or so.

--Michael

[shancai]

I use the attached file to optimize dynamic aperture.
I want to correct chromaticities each time. But it Adjust chromaticities many times.
What's wrong with it?

Thanks a lot!

[michael_borland]

Shancai,

The chromaticity is corrected for each function evaluation. You want this, because you are changing quadrupoles. If the chromaticity isn't corrected each time, you'll almost certainly have a result you can't use (dynamic aperture optimized but wrong chromaticity).

If you want to correct the chromaticity just once, then you need to run elegant separately and save the results, then reload them for the DA optimization.

A few more suggestions for your input file:
* Consider simplex optimization instead of random-walk
* Set n_steps=1 instead of n_steps=10. Otherwise, you'll run the same optimization 10 times.
* You don't need radiation_integrals=1 in twiss_output, since you haven't included an SREFFECTS element and aren't optimizing any radiation-related quantities.
* Set concat_order=2 in twiss_output to get slightly better performance

--Michael

[michael_borland]

Shancai,

Attached are some files to demonstrate performing a tune scan to optimize dynamic aperture. The scripts work on linux, but may require modification to work on Windows.

Also attached is a plot showing what the result looks like.

A word of caution: this analyzes only the dynamic aperture. The momentum aperture will also vary with tune. A proper lattice design has to compromise between the two of these.

--Michael

[shancai]

I test the example on my windows system.
The only changing is change elegant to elegant.exe.
It can run no problem.

[hsayed]

Hi Micheal,

I'm interested in the same kind of problem. For some reason the attached files are no longer available.
Thanks
Hisham

[michael_borland]

Hisham,

Please try again. It should work now.

--Michael

[csun]

hello Michael,

Thank for the explanation. We are also interested in the same kind of problem. Could you post an example using resonance driving terms to improve the dynamics aperture? You mentioned the genetic optimization works better than driving terms optimization. May I ask for your genetic optimization script. Thanks!

CSUN

[wguo]

Attached is an example file that optimizes dynamic aperture through resonance term minimization.

Instructions can be found in the file.

Good luck,

Weiming Guo

DA-optimize-example.ele

Dynamic aperture optimization through resonance term minimization

(4.8 KiB) Downloaded 1090 times