Magnetic Field Simulation Using Radia and Tosca For Kolkata Superconducting Cyclotron

Cyclotrons and Their Applications 2007, Eighteenth International Conference
MAGNETIC FIELD SIMULATION USING RADIA AND TOSCA FOR

KOLKATA SUPERCONDUCTING CYCLOTRON
J Pradhan*, A Dutta, U Bhunia, J Debnath, S Paul, Z A Naser, M K Dey,
C Mallik and R K Bhandari
Variable Energy Cyclotron Centre, 1/AF, Bidhan Nagar, Kolkata-700 064, INDIA
commercially available finite element code whereas
RADIA [3] is freely available and is interfaced with
MATHEMATICA [4], uses boundary integral method to
solve. The RADIA requires geometry creation using
simple MATHEMATICA code requires less CPU time of
the solver however have some weak points as well. The
finite element code requires significant skills and time for
the input of geometry, materials and proper meshing [5].
The median plane view of the main magnet constructed in
RADIA is shown in figure 1. The pole tip hill is machined
into two pieces to approximate its spiral profile. The hill
is chamfered on both azimuthally edges at 45. There are
several small assemblies and sub-assemblies like hillshoe, pole tips, central plug with its hill and valley part
and so on which produce assembly errors besides several
holes for trim coil inserts and others. Apart from this, the
geometrical complexity of the magnet poses dimensional
uncertainties. It is likely that the splitting of the poletips, different holes etc may have definite influence on the
local saturation both in strength and direction, because of
the topography of the coil generated field [6]. During
field measurement several shims of different shape and
sizes have been added to remove field imperfection.
However we modelled the full magnet and compared the
field data thus obtained with measured ones. Because of
some limitations (computer memory) and other reasons as
mentioned we could not arrive to the accuracy as desired
for the internal field. So, in order to calculate accurately
the field arising from such complex structure it is assumes
that the pole tip surfaces to be fully saturated. So
Abstract
Detail magnetic field measurement of the Kolkata
Superconducting Cyclotron, in the acceleration and the
extraction region is carried out for several excitations of
the cyclotron. However it was not possible to measure the
fields at all excitations spanning over the entire operating
region of the cyclotron. Moreover, measurement in the
extraction trajectory is limited by the in-accessibility of
the region to the measurement jig system. So the
magnetic field simulation is performed using 3-D
magneto static codes (RADIA and TOSCA) to compute
the field data required for running the beam dynamics
codes. This paper presents a scheme to generate field data
by excluding the part of the model which is nearer to the
median plane and saturated, thus avoiding detail/minute
modelling of that region (like shims, trim-coil inserts
etc.). The comparison between all the measured data and
the corresponding simulated ones shows the deviations
are within a few gauss, which justifies the reliability and
acceptability of the scheme to extrapolate the field for
higher excitation.
INTRODUCTION
The main magnet of Kolkata Superconducting
Cyclotron (SCC) consists of two coils (-coil & -coil),
which has been excited to different measurement grid
points and field mapping has been carried out [1].
However it is not possible to measure the field at all
excitations and the measurement in the extraction region
is limited by the in-accessibility of the region to the
measurement jig system. The axial field is measured and
compared along the central plug of the SCC. This paper
presents the 3-D simulation of magnetic field to compute
the field data for all the excitations spanning over the
entire operating region of the cyclotron. The simulation
data provides the iron-produced field, both for the
azimuthal modulation and the average field required for
beam dynamic codes. The measured data is compared
with calculation to establish the reliability of the
simulation. The edge field is also simulated and
compared, which is required for the studies of extraction
system.
Magnet coil
Channel inside
the yoke hole
SIMULATION SCHEMES
The field is calculated by 3-D modelling of Kolkata SCC
main magnet. Both OPERA-3D (TOSCA) and RADIA
codes are used for simulation. OPERA 3-D [2] is a
______________________________________________
*[email protected]
441
Figure 1. Lower half portion of the magnet.
part of the model (pole tip surface) which is nearer to the

median plane is excluded to avoid detail/minute
modelling. The field from these saturated surfaces is
obtained from one of the measured field data. In addition
this also incorporates the field due to all the shims added
in the pole tips. The small assembly errors, if any, around
the pole tip region, which has major contribution, will be
absorbed. This constant field, B sat (r, ) is added with the
field obtained from the model calculation to get the
internal field. The coil field, B coils (r, , I, I) is
calculated analytically, which is used to get the iron
generated field, B iron(r, , I, I). The contraction of coil at
liquid helium temperature that is confined to 0.3% [6] is
not taken into account in calculation. The uncertainties of
the coil dimension and its placement with respect to
surrounding yoke produces non- negligible errors
especially in the higher radii. We can write the total field,
B(r, , I, I) as the sum of B sat (r, ) + B coils (r, , I, I) +
B iron(r, , I, I). The iron field, Biron(r, , I, I) is consist
of the field generated by the variable saturation of the
yoke and primarily depends on the magnetisation of iron.
various iron shims. This is an iterative process. The final

measured data is smoothened, analysed and compared
with simulated ones. The comparison of Iron average for
different currents combination of -coil & -coil is shown
in figure 3. The top five curves (5 to 8) are extrapolated
data for which measurement data are not available. The
observed discrepancy in iron average field is less than
0.5% for all sets of measured data as shown in figure 3.
The variable saturation of the yoke depending on the coil
currents affects both the absolute level and average slope
of iron field. The comparison for the 3rd, 6th and 9th
harmonics is shown in figure 4 for one set of coil currents
and it is found that the maximum deviation is less than
0.2%.
Percentage deviation
0.5
RESULTS AND DISCUSSION

Coil Field:
Biron (kG)
17
375
403
481
459
-0.3
200
10.0
I=459A I=471A
13
12
0
200
400
Radius (mm)
C3
B(kG)
6.0
4.0
C6
2.0
275
340
356
471
C9
0.0
0
200
400
600
Radius (mm)
Figure 4. Measured and calculated amplitude of the 3rd,

6th and 9th harmonics, as function of the radius at the
listed coil currents [ calculated, oo measured] .
8
5
4
3
2
1
14
600
8.0
16
15
400
Figure 3. Percentage difference between measured and

calculated iron average field for the given set of currents.
B0 (kG) I (A) I(A)
16.47
16.85
17.16
17.49
-0.1
Radius (mm)
The internal field is measured upto radius of 673 mm in

the median plane, and coincides roughly with the
extraction radius of cyclotron. Beyond this radius internal
field mapping is not feasible because of the presence of
coil tank inner wall. Datas are taken at 2.54 mm radial
and 1 azimuthal interval respectively. Different imperfection present in the field is reduced by applying
18
0.1
Iron Internal Field:
#
1
2
3
4
-0.5
The field due to two independent -coil & -coil is

calculated analytically. This is air-core field. It is likely to
have some discrepancies form this calculated field on
account of thermal contraction, dimensional in-accuracy,
assembly errors and so on. This field is used to get iron
field from the total field as mentioned above.
19
0.3
Axial Field:
The magnetic field along the axis has sharp gradient
and appreciably large fringing field extending up to
several meters. The field data is required to study the
beam injection system into the cyclotron. The
measurement has been performed using hall probes from
the median plane up to the height of 1.7mts with interval
of 5mm.The measured and simulated axial field
600
Figure 2. Average Iron field profile for the listed coil

currents. Top 4 curves are obtained from simulation.
[ calculated, oo measured]
442
distribution is shown in figure 5, where 3 and 4 sets are

extrapolated data for which measurements are not
available.
Axial Field (kG)
50
40
30
20
I(A) I(A)
1
2
3
4
200
300
500
700
excitations plays a substantial role in shaping the ironproduced field.
Channel Field inside the Yoke hole:

One active magnetic channel of the extraction system is
placed in the yoke hole of the cyclotron as shown in
figure 1. This channel is a combined function magnet
used primarily to align the beam with external beam line.
The magnetic field inside the channel and outside stray
field has been measured for different channel currents as
well as main magnet excitations using hall probes. The
field along the length of the channel with the channel
centre at S=0 is shown in figure 7 for different channel
currents and main magnet excitations.
200
300
500
700
3
2
10
0
0
500
1000
1500
2000
Axial distance (mm)
Figure 5. Axial field distribution [ calculated,

measured]
B(kG)
-2
oo
-4
Outside the
magnet yoke
-6
-8
Edge Field:
-10
-250 -150
It is necessary to calculate the path of the extracted beam,

which requires edge field data. Edge field measurement is
carried out by inserting the hall probes through a different
radial holes existing in the median plane of the cryostat.
Datas are taken at 12.7mm radial intervals at affixed
azimuth. The possible errors in the edge field
measurement like radial and azimuthal shift is
incorporated by comparing with internal data of the same
excitation. The comparison of the data for two different
25
Figure
B (kG)
15
600
350
Field inside the active

[ calculated, oo measured]
450
channel.
REFERENCES
= 203
[1] C Mallik et al., Magnetic Field Mapping of Kolkata

Superconducting Cyclotron presenting in this
conference.
[2] OPERA 3D Software for electro-magnetic design by
Vector Field
[3] RADIA Technical Reference Manual ESRF,
Grenoble, France
[4] MATHEMATICA is a registered trademark of
Wolfram Research, Inc.
[5] P.Elleaume et al., Computing 3D Magnetic Field
from Insertion Devices IEEE (1998)
[6] G. Bellomo et al,. Magnetic Field Mapping of the
K-500 Cyclotron at MSU, Nuclear Instrument and
Methods 180(1981), 285-304
5
0
500
250
The computational scheme used to simulate the

magnetic field is quite satisfactory to compute field data,
where the measurement data is not available. The detail
comparison between measured and simulated data is
performed to validate the simulation.
20
10
150
CONCLUSIONS
I=459A; I=471A
= 147
7.
-50
50
S(mm)
700
800
900
1000
Radius (mm)
Figure 6. Comparison of Edge iron field at two different

angles [ calculated, oo measured]
azimuths for a given excitation is shown in figure 6.
Possible sources of error in the edge field region due to
coil field contribution are already mentioned. At the
magnet edge especially in the proximity of iron structures
(coil tank walls) the influence of the different coil
443

Magnetic Field Simulation Using Radia and Tosca For Kolkata Superconducting Cyclotron

Uploaded by

Copyright:

Available Formats

Magnetic Field Simulation Using Radia and Tosca For Kolkata Superconducting Cyclotron

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Magnetic Field Simulation Using Radia and Tosca For Kolkata Superconducting Cyclotron

Uploaded by

Copyright:

Available Formats

Cyclotrons and Their Applications 2007, Eighteenth International Conference

MAGNETIC FIELD SIMULATION USING RADIA AND TOSCA FOR

Figure 1. Lower half portion of the magnet.

Cyclotrons and Their Applications 2007, Eighteenth International Conference

part of the model (pole tip surface) which is nearer to the

various iron shims. This is an iterative process. The final

RESULTS AND DISCUSSION

Figure 4. Measured and calculated amplitude of the 3rd,

Figure 3. Percentage difference between measured and

B0 (kG) I (A) I(A)

The internal field is measured upto radius of 673 mm in

Iron Internal Field:

The field due to two independent -coil & -coil is

Figure 2. Average Iron field profile for the listed coil

Cyclotrons and Their Applications 2007, Eighteenth International Conference

distribution is shown in figure 5, where 3 and 4 sets are

Axial Field (kG)

excitations plays a substantial role in shaping the ironproduced field.

Channel Field inside the Yoke hole:

Axial distance (mm)

Figure 5. Axial field distribution [ calculated,

It is necessary to calculate the path of the extracted beam,

Field inside the active

[1] C Mallik et al., Magnetic Field Mapping of Kolkata

The computational scheme used to simulate the

Figure 6. Comparison of Edge iron field at two different

You might also like