eddy current septum magnet optimization

19
Eddy Current Septum Magnet Optimization Powering Options of SMH42 and the Influence of the Septum Thickness on the Fringe Field Zsolt SZOKE (TE/ABT/ SE)

Upload: reese-morse

Post on 02-Jan-2016

41 views

Category:

Documents


1 download

DESCRIPTION

Eddy Current Septum Magnet Optimization. Powering Options of SMH42 and the Influence of the Septum T hickness on the Fringe Field. Zsolt SZOKE (TE/ABT/SE). Outline. Eddy Current Septa Magnets Our Goal Baseline Design Performance Analysis in Time Domain - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Eddy Current Septum Magnet  Optimization

Eddy Current Septum Magnet Optimization

Powering Options of SMH42 and the Influence of the Septum Thickness on the Fringe Field

Zsolt SZOKE (TE/ABT/SE)

Page 2: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 2

Outline

• Eddy Current Septa Magnets• Our Goal• Baseline Design Performance• Analysis in Time Domain• Comparing Full Sine and Half Sine Excitation• Comparing 5mm and 3mm Septum Blades• Comparing 2ms and 7ms Wavelength

19/08/2014

Page 3: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 3

Eddy Current Septa Magnets

• Different types of septa:– direct drive (DC, pulsed)– eddy current (only pulsed)

• Eddy current type - advantages:– coil dimensions are not critical– the pulsed coil has such a

magnetic field which induces eddy currents in the septum counteracting the fringe field

– septum can be very thin19/08/2014

Page 4: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 4

Our Goal

• Optimize different eddy current septum magnet parameters.

• 3 comparisons made with the baseline design.• Baseline: 2ms, full sine, 5mm septum.

• Examination of the fringe field: By and ∫Bydl.

19/08/2014

7ms half sine 3mm septum

Page 5: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 5

Baseline Design Performance

• ʃBy,gapdlmax = 502.67Tmm

• By,gap,max = 542mT

• ʃBy,fringedlmax = -1.37Tmm (after the current pulse)

• By,fringe,max = -1.4mT (after the current pulse)

• Idriving,max = 30251A• Gap fringe

field shape:

19/08/2014

Page 6: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 6

Analysis in Time Domain (1/2)• Opera finite element simulation, spanning 3×

the excitation time• Discrete moments interpolation in MATLAB• 2 types of interpolation:

– PCHIP: for plotting– SPLINE: for peak

determination

19/08/2014

Page 7: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 7

Analysis in Time Domain (2/2)

• 6 values for each simulated moment:– t [ms]– I [kA]– By (gap)

– ∫Bydl (gap)

– By (fringe)

– ∫Bydl (fringe)

19/08/2014

“gap”: the middle of the aperture

“fringe”: 5mm from the septum

Page 8: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 8

Comparing By of Full Sine and Half Sine

• Fringe field extents after excitation.

• Full sine: By = -1.4mT

• Half sine:By = 15mT

19/08/2014

0 1 2 3 4 5 6-40

-30

-20

-10

0

10

20

30

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-2

0

2

4

6

8

10

12

14x 10

-3

Mag

netic

flu

x de

nsity

[T

]

0 1 2 3 4 5 60

20

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

0

0.01

0.02

Mag

netic

flu

x de

nsity

[T

]

Page 9: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 9

Comparing ʃBydl of Full Sine and Half Sine

• Integrated fringe field extents after excitation.

• Full sine: ∫Bydl = -1.37Tmm

• Half sine:∫Bydl = 14.28Tmm

19/08/2014

0 1 2 3 4 5 6-50

0

50

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-20

0

20

Fie

ld in

tegr

al [

Tm

m]

0 1 2 3 4 5 60

20

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

0

10

20

Fie

ld in

tegr

al [

Tm

m]

Page 10: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 10

Comparing Full Sine and Half Sine

• Huge time constants in both cases:– 1 or 2ms excitation time (half or full sine)– time constant: >4ms

• Fringe field peak values are 10.4-10.7 times lower using full sine wave instead of half sine.

• Using ‘direct damping’ of the fringe field (full sine excitation) proves to be very effective.

19/08/2014

Page 11: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 11

Comparing By of 5mm and 3mm Septa

• Fringe field extents after excitation.

• 5mm septum blade: By = -1.4mT

• 3mm septum blade:By = -3.6mT

19/08/2014

0 1 2 3 4 5 6-40

-30

-20

-10

0

10

20

30

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-2

0

2

4

6

8

10

12

14x 10

-3

Mag

netic

flu

x de

nsity

[T

]

0 1 2 3 4 5 6-40

-20

0

20

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-0.01

0

0.01

0.02

0.03

Mag

netic

flu

x de

nsity

[T

]

Page 12: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 12

Comparing ʃBydl of 5mm and 3mm Septa

• Integrated fringe field extents after excitation.

• 5mm septum blade : ∫Bydl = -1.37Tmm

• 3mm septum blade :∫Bydl = -3.51Tmm

19/08/2014

0 1 2 3 4 5 6-50

0

50

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-20

0

20

Fie

ld in

tegr

al [

Tm

m]

0 1 2 3 4 5 6-40

-20

0

20

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-10

0

10

20

30

Fie

ld in

tegr

al [

Tm

m]

Page 13: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 13

Comparing 5mm and 3mm Septa

• A thinner septum blade is advantageous for the beam: lower continuous losses.

• 3 mm septum blade has higher current density.• Fringe field peak values are 1.7-2.6 times lower

using 5mm septum instead of 3mm.

19/08/2014

Page 14: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 14

Comparing By of 2ms and 7ms Wavelength

• Fringe field extents after excitation.

• 2ms wavelength: By = -1.4mT

• 7ms wavelength:By = -15.6mT

19/08/2014

0 1 2 3 4 5 6-40

-30

-20

-10

0

10

20

30

40

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-2

0

2

4

6

8

10

12

14x 10

-3

Mag

netic

flu

x de

nsity

[T

]

0 5 10 15 20 25-40

-20

0

20

40

Cur

rent

[kA

]

Time [ms]0 5 10 15 20 25

-0.02

0

0.02

0.04

0.06

Mag

netic

flu

x de

nsity

[T

]

Page 15: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 15

Comparing ʃBydl of 2ms and 7ms Wavelength

• Integrated fringe field extents after excitation.

• 2ms wavelength : ∫Bydl = -1.37Tmm

• 7ms wavelength :∫Bydl = -15.14Tmm

19/08/2014

0 1 2 3 4 5 6-50

0

50

Cur

rent

[kA

]

Time [ms]0 1 2 3 4 5 6

-20

0

20

Fie

ld in

tegr

al [

Tm

m]

0 5 10 15 20 25-40

-20

0

20

40

Cur

rent

[kA

]

Time [ms]0 5 10 15 20 25

-20

0

20

40

60

Fie

ld in

tegr

al [

Tm

m]

Page 16: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 16

Comparing 2ms Wavelength and 7ms

• 3.5 ms pulse half sine shape would be a pulse length similar to present SMH42.

• Shorter pulse length (w.r.t. baseline design) wasn’t investigated, since it will be very difficult to build a compatible magnet.

• Fringe field peak values are 3.2-11.1 times lower using 2ms full sine wave instead of 7ms.

19/08/2014

Page 17: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 17

Final Conclusion

Full sine vs. Half sineDirect fringe field cancellation very effective

5mm septum vs. 3mm septumThe thicker the septum, the lower the fringe field

2ms wavelength vs. 7ms wavelengthThe shorter the pulse, the lower the fringe field

The base line design appears a good compromise.

Next: the BMP42 septum bumper analysis.19/08/2014

Page 18: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 18

References• Full documentation: Z. Szoke: Eddy Current Septa Magnet

Optimization• M. J. Barnes, J. Borburgh, B. Goddard, M. Hourican, in Proceedings

of the CAS-CERN Accelerator School: Magnets, Bruges, Belgium, 16-25 June 2009, edited by D. Brandt, CERN-2010-004, pp. 167-184

• Finite element simulations: Cobham Opera 16• Data processing: MATLAB R2013b

19/08/2014

Page 19: Eddy Current Septum Magnet  Optimization

LIU-PS Meeting 19

Thank You for Your Attention!

Q&A

19/08/2014