LT Calcoli poster at symposium on fusion technology SOFT 2010. https://www.youtube.com/watch?v=UP0JSVwUWZg Check out the SOFT 2018 participation by following us at https://twitter.com/LTCalcoli https://www.linkedin.com/company/l.t.-calcoli-s.r.l./

1. F. Lucca1
, E. Briani1
, C.Gianini1
, C. Jong2
, J. Knaster2
, A. Marin1
(1) L.T. Calcoli S.a.S., Piazza Prinetti 26/B, 23807, Merate (LC), Italy
(2) ITER Organization, Route de Vinon, CS 90 046, 13067 St. Paul lez Durance Cedex, France
26th
Symposium on Fusion Technology
Porto – Portugal –
Sept 27 – Oct 1 2010
L.T. CalcoliL.T. CalcoliITER TF Coil Double Pancake assembly:
laser welding numerical simulation
In the frame of the ITER coils production, one crucial point is the Cover Plates (CP) welding to complete a coil layer, the so
called Double Pancake, once the reacted and insulated conductor is fit inside the Radial Plate (RP) grooves. The scope of
the activity here described is the assessment of the deformation induced by this laser welding through a FE numerical
simulation, in order to find the welding sequence that minimize the global distortion.
In the frame of the Finite Elements method, the welding can be simulated through a
“classical method” or a local-to-global approach.
The classical method consists in a transient heat transfer + subsequent structural
analysis. The global distortion induced by the welding can be simulated with a good
approximation, but “long” time analyses are required: the CPU times can be
acceptable (tenth of hours) on small samples (tenth of cm welding paths), but they
become unrealistic on long welds (~ some meters in length).
The Local-Global approach has been developed to overcome this problem
The 2D detailed model reproduces the complete Double Pancake, with a very
detailed mesh in the P-side region.
A detailed thermal + structural analysis of the P side passes has been performed and
the out of plane final distortions have been compared with the experimental ones,
showing a good agreement.
AVG εx
AVG εy
The thermal contractions have been applied on a 2D global model, able to reproduce
the complete structure.
Thermal model:
temperature field
Structural model:
displacement (m) in X direction [chamfer closure]
The welding process is simulated applying the
computed contractions at weld seams (the
yellow elements in figure) of P-plane and N-
plane following the experimental welding
sequence.
The experimental welding sequence
has been simulated on the 3D model.
Radial dimension change after P plane weld and
release evaluated along 6 reference radial paths
The comparison between
the numerical and the
available experimental
data [1] has shown a good
agreement in describing
the “deformed shape”.
The global result of the
correlation can be
considered positive.
A numerical procedure aimed at evaluating the distortion induced by non-deep (up to few mm) laser welding has been
developed and applied to the Middle and Side DP of the ITER coils. The results are very promising, and a complete asses
of the procedure could be performed in the frame of an experimental-numerical research campaign.
In this activity, the Local-Global approach
technique has been adopted using Abaqus
Code, in order to be able to compare several
long welding paths in acceptable time frames
in spite of partial accuracy loss in favor of
speed and ease of model definition.
The numerical methodology has been tuned
and assessed on experimental data from a
welding sample (1m long mock-up
reproducing the Middle DP curved section) in
which a 2.5 mm deep welding has been
performed
The deformation induced by 1 and 2 mm
deep welds (the ones that will be used in the
complete DP) has been extrapolated and
these local strains have been applied on the
global models reproducing the Middle and
Side DP, and their deformation has been
computed.
Why Local-Global? Procedure asses
on Sample Test
22 + 22 welding (each 1m
long) were experimentally
[1] and numerically
performed on the upper (P)
and lower (N) mock up
faces .
Reduced 3D models have been developed for the thermal contraction evaluation.
TU
N
I
N
G
EX
PE
R
I
M
E
NT
AL
Vs
N
U
M
E
R
I
C
AL
Procedure application
on Full DP
The complete welding sequence (22 + 22 welds and final constraint removal) has been
simulated, with different boundary constraint and welding depths.
• The maximum out of plane deformation occurs after P side welding and constraints
release; it is of few mm.
•The maximum out of plane displacement after N side welding and constraint release is
between some hundreds of mm and 1 mm in the various analysed cases.
• The maximum X
contraction occurs after N
side welding and
constraint release; it is of
few mm and varies
linearly with the
contractions.
On the Side DP 3D model, two
different boundary conditions
and two welding depths (2 and
2.5 mm) have been applied.
In the Middle Double Pancake
three regions (A, B and C) can
be found, slightly different in
geometrical section and
characterised by different
welding depths: 1 or 2 mm.
To find the welding sequence that
minimizes the deformation three
welding sequences have been simulated on the 2D models: from centre to
outer (seq1), from outer to centre (seq2), from outer left to outer right
(seq4).
The final differences among the three sequences are lower than 1%.
Out of plane displacement of a zoomed region in
the inboard branch after the complete welding
(both N and P side) and constraint release
Due to the section non-symmetry,
higher final deformations (from 10
to 100% in the several analysed
cases) are present on the Side DP
respect to the Middle DP.
Local detailed 2D and 3D models
have been developed, to compute
the new contractions.
FU
L
L
D
O
U
BL
E
P
A
N
C
A
K
E
W
EL
D
I
N
G
References:
[1] N. Koizumi et al, Fusion Engineering and Design –
Volume 84, 2009, pages 210-214.
In the Side Double Pancake
the inboard and outboard
section are quite different
from each other and very
different from the middle DP.
Inboard Side RP - 2.5 mm deep welding - Vertical displacement
-7.0E-04
-6.0E-04
-5.0E-04
-4.0E-04
-3.0E-04
-2.0E-04
-1.0E-04
0.0E+00
1.0E-04
0 5 10 15 20 25 30
welding passes
verticaldisplacement(m)
03_seq1
05_seq1
25_seq1
03_seq2
05_seq2
25_seq2
03_seq4
05_seq4
25_seq4
CONCLUSIONS