Project

General

Profile

GEOM 011012 » plan.txt

Blackmore, Victoria, 01 October 2012 13:45

 
1
Plan for magnetic field map "interface":
2

    
3
1. Code exists to calculate cylindrically symmetric field maps using the Biot-Savart law.
4
	-- Define a magnet as a series of coils
5
	-- Define a coil as a layers of current loops
6
	-- Define a current loop by its radius, and current flowing around it
7

    
8
	-- Generates field points in (r, z) = (B_r, B_z).
9

    
10
	-- To do: Currently re-writing this code to make it more user friendly. In other words, take it from 'proof of principle' to a workhorse for fitting field maps.
11

    
12
2*. Take field points in (r, theta, z) from CERN mapper tool. 
13
	-- Have an idea of the "range" of parameters that the magnet could have. So we know how many coils there are, but then have an envelope of radii, number of current loops, currents, etc. to fit too.
14
	-- Worst case scenario: Brute force iteration through all the options and a chi^2 fit for each option to the data. Best chi^2 = best coil model.  However, I'm thinking about ways that I could break the fits down into speedier portions (calculating the on-axis field is pretty quick and straightforward, which will at least give us a better guess at whether the coils are centred radially, etc. etc.)
15

    
16
3. Output the "best" coil model that suits the measured field map.  So this will be #coils, position of coils w.r.t. beam axis, any tilts on coils, number of layers per coil, number of coil loops per layer, current flowing around coil loops to generate that field.
17

    
18
4. Depending on the quality of the fit to the data, we may also have to look at the error field and its Fourier components. This is a bit more complicated and I haven't given it any more thought beyond my presentation at CM30: https://indico.cern.ch/getFile.py/access?contribId=36&sessionId=19&resId=0&materialId=3&confId=137610 (The section on the "Fourier-Bessel Model", p4 onwards)  Thoughts on this will have to wait until I've tried fitting a smeared field map or even a real magnet (i.e. MTA magnet, not a FC/SS).
19

    
20

    
21

    
22

    
23
* Before this point, test it out with a "perfect" field map.. so generate something with the biot-savart code, then try to fit to it, making sure that the fitted parameters come back as consistent with what is put in.  Secondly, test it out with a "smeared" field map. Finally, attempt to determine what the MTA magnet is -- Yagmur has a field map that I can use for this.
24

    
(1-1/2)