Page 1 of 1

Modifications to find_aperture

Posted: 01 Mar 2024, 03:27
by Skamarokha
Hello

Often dynamic aperture in X is larger than in Y. In n-line search mode of find_aperture function we now can choose the step for n-line mode search like step=xmax/n_x. This leeds to fine aperture result for large Y, but poor results for small Y. I propose adding ny for n-line search mode. Typically we choose the step for x,y and dp/p as 2*sigma_{x,y,z}. Described problem is illustrated on the following image.
DA example.jpg
Another proposal is to add X vs dP/P aperture search in find_aperture or make a separate similar function. We are interested in n-line search mode and area parameter for X vs dP/P for optimization. In our work we resently discovered that optimizing X vs dP/P area is a handy way to increase touschek lifetime, especially if your PC is powerful and you for some reason don't want to or can't use cluster for MOGA optimization.

Re: Modifications to find_aperture

Posted: 07 Mar 2024, 17:00
by michael_borland
Mikhail,

I think the only way you can get what I understand you are looking for is to use something other than line-search mode. By definition, line-search mode searches along lines starting from the origin. The radial spacing along the lines is xmax/nx. You can always reduce nx if you want finer steps.

The downside of the many-particle search in elegant is that it searches from the outside in, which can sometimes give misleading results. I'll look into providing an option to search from the inside out.

For (x, delta) DA, you can use the frequency_map command. You'll have to compute the area yourself, but it isn't hard. It's a good idea to make this more convenient, so I'll think about how to do that.

--Michael

Re: Modifications to find_aperture

Posted: 29 Mar 2024, 14:41
by michael_borland
Mikhail,

The ability to perform (delta, x) and (delta, y) aperture finding will appear in the next release.

Thanks for the suggestion.

--Michael