5. Engineering Design of Tokamak Device
The HT-7U superconducting tokamak device features a superconducting magnet system, which
consists of sixteen TF coils and fourteen PF coils located about the plasma mid-plane
symmetrically. The central solenoidal (CS) assembly consists six PF coils. A vacuum vessel with
sixteen rectangular horizontal ports and thirty-two bathtub-shaped vertical ports is located in the
bore of the TF coil. A cryostat with two thermal shields of 80K encloses all of the superconducting
coils, the vacuum vessel and the support structures. The superconducting magnet system, the
vacuum vessel and the thermal shields are supported on the cryostat independently. The HT-7U
device has a height of 10 m (with the main support), a diameter of 7.6 m and a total weight of 360
tonnes. Fig. 1 shows the vertical cross-section view of the HT-7U device with supports. The main
parameters of the HT-7U device are list in Table I.
Fig. 1 The vertical cross-section view of the HT-7U device with supports
Table I.
Main Parameters of the HT-7U Device
Toroidal Field, Bo
3.5 T
Plasma Current, IP
1 MA
Major Radius, Ro
1.7 m
Minor Radius, a
0.4 m
Aspect Ratio, R/a
4.25
Elongation, Kx
1.6 - 2
x
0.6 - 0.8
Heating and Driving:
ICRF
3 MW
LHCD
3.5 MW
ECRH
0.5 MW
Pulse length
10- 1000 s
Configuration
Double-null divertor
Pump limiter
Single null divertor
5.1 TF System
The TF magnet system of HT-7U consists of a toroidal array of sixteen coils, seeing Fig. 2.
The TF magnets are designed to support the magnetic centripetal forces and overturning torques by
wedges at their inner legs and between their outer arc segments. The TF assembly with all attached
PF magnets is supported on eight low rigid supports.
Fig. 2 The TF magnet system of HT-7U
Table II
TF magnet system characteristics
Major radius (R0)
1.7 m
Minor radius (a)
0.4 m
Magnetic field at the plasma center (Bt)
3.5T (4.0 T)
Maximum field at the coil (Bmax)
5.8 T (6.5 T)
Number of TF coils (N)
16
Total number of Apere-turns
30MAT (34MAT)
Operating current (Io)
14.3 kA (16.4 kA)
Total stored energy
300 MJ (390 MJ)
Total length of CICC of the TF system
19.2Km
Total weight of the TF system
160 ton
The conventional lowest in-plane stress TF coil shape is well-known “bending moment free
D”. It is approaching to a pure tension coil for reducing bending moment in the coil. The contour of
the D-shaped coil consists of five arcs and a straight leg. The straight leg of the coil with wedge is
squeezed each other to form an inner arch. The adjacent intercoil structure is of wedge-shaped
across-section. External wedge chocks of the torus are also squeezed each other with coil to form
an outer arch. Two arches resist strong centering force and overturning torque. Total sixteen
insulating breaks apart in the torus of TF system. A returning turn in the TF system is used to
compensate the toroidal component of the current of crossover lines between coil and coil. There
are sixteen horizontal ports at the midplane and thirty-two vertical ports at the top and the bottom
wedge chocks between coil and coil for passing vacuum vessel ports. A cryogenic-cooling channel
system on each TF coil is consisted of fourteen cooling lines connected in parallel. Local
instrumentation and control for the TF magnet system is comprised of diagnostics, protection
systems, sensor interfaces and acquisition systems.
The TF coil is wound in pancakes, and the TF coil electrical insulation is composed of
multiple layers of polyimide film-glass impregnated with epoxy resin. Each turn of conductor is
wrapped with half lapped of glass cloth tape built up to a thickness of 0.5 mm. Glass epoxy felt
(1mm thick) is inserted between pancakes. A glass cloth roving is inserted into the corners formed
by the juncture of the glass epoxy sheets and the rounded corners of the conduit. A wrap of
half-lapped glass cloth tape built up to a thickness of 8mm on the coil surface forms the ground
insulation. The allowable ground voltage in normal operation is about 5KV.
The TF coil cases are 316LN stainless steel plate welded-up structures with full penetration
welds. It’s structure is shown on Fig.3. Shape of cross-sections of the TF coil-case is a “keystone”
for the straight leg and a rectangle for the rest parts. The four walls thickness of the coil cases is not
equal. The steel plate thick of inboard leg is 60mm. The 16 TF coil case inboard legs are wedged
together
and wedging loads are reacted by hoop compression yielded from centering force.
The steel plate thick of outboard leg is 40mm. The steel plate thick of two sides is 25mm for the
wedged parts and 30mm for the rests parts. The TF coil case is made of two halves with welding
structures. The inter-coil structure is also a welded-up structure made by 316LN stainless steel plate,
thickness of which is 20mm.
Fig. 3 The structure of the TF case
The structural support concept uses mutual support between the magnet components to ensure
that in normal operation, all electromagnetic forces are reacted within the TF magnets, and there
are no resultant force transmitted through the gravity support system.
The gravity supports of TF system are composed of pedestals with flexible plates, so that they
can deflect in the radial direction to allow thermal expansion of the TF magnet system. But they are
rigid versus out-of-plane bending caused by TF coil torsion or seismic motion. The thermal load of
the gravity supports from the room temperature to 4K is intercepted by a thermal anchor consisting
of cooling channels for 80K helium gas.
ITER (International Thermal-nuclear Experiment Reactor) TF structural design criteria will be
adopted as the global level of TF structure design in HT-7U. According to ITER TF structural
design criteria, stresses in TF structure are divided into three main groups: primary stress,
secondary stress and peak stress. Primary stress is a stress developed by the imposed loading which
is necessary to satisfy the laws of equilibrium between external and internal forces and moments.
The basic characteristic of a primary stress is that it is not self-limiting. In our case the stresses due
to the electromagnetic loading fall into this group. There are:
·Primary membrane stress: e.g. average stress in the TF coil case due to in-plane loading;
··Primary bending stress: e.g. bending stress in the TF coil case due to out-of-plane loading.
Secondary stress is a stress developed by the constraint of adjacent material or by self-constraint of
structure. It must satisfy an imposed strain pattern rather than being in equilibrium with an external
load. The secondary stress is self-limiting. The generally thermal stress often represents this kind of
stress. Peak stress is that increment of stress which is additive to the primary plus secondary
stresses duo to local discontinuity or local thermal stress, including effects of stress concentration.
A peak stress doesn’t cause any noticeable distortion but is a possible source of a fatigue crack. The
stress limits are based on Von·Mises stress, defined as the root of a half of quadratic sum of the
algebraic difference between any two of the three principal stresses. The maximum allowable
Von·Mises stress in the material (base metal) is defined as 2/3 times the 0.2% yield stress of the
material. In normal operation, the allowable value is increased by a factor of 1.3 for combined
membrane and bending stresses and 1.5 (i.e. up to yield) for combined primary and secondary. For
welds, in sections characteristic of the coil cases, there two values are decreased by a factor of 0.8.
The following loads are considered in various combinations in HT-7U TF structural design:
A) Dead loads (D): primarily the weight of the component and attached elements.
B) Design pressure (P): the pressure imposed on a component due to hydraulic or gas
pressures in a worst-case fault scenario.
C) Normal operating thermal effect (To): loads imposed by temperature changes before,
during, and after a pulse.
D) Electromagnetic loads (EM): loads on magnets due to the currents flowing in the TF
conductors interacting with the magnetic fields crossing them. These loads also include
electromagnetic effects of discharge cleaning.
E) Electromagnetic loads during faults (EM-F): electromagnetic loads induced during
abnormal events such as: control failures, quench, etc.
F) Interaction loads (IR): loads superimposed by interactions between components; for
example, the load due to a PF coil supported on the TF coil.
G) Electromagnetic loads due to plasma disruption (EM-D): the electromagnetic loads
induced during a plasma disruption.
H) Seismic loads: loads induced by the safe shutdown earthquake (SSE) or most intense
earthquake (MIE).
The TF magnet components shall be designed for both normal operating conditions and
off-normal events. The load combinations are defined below (P= the estimated probability of
occurrence per year):
Normal operating events (Normal) (P≤1): D+P (where applicable) + To +EM (or EM-D) + IR
Anticipated events (Upset) (1≥P>10-2):
D+P (where applicable) + To + (EM-F) + IR
Unlikely events (Severe) (1≥P>10-2):
D+P (where applicable) + To + FSSE + IR
Extremely unlikely (Faulted) (10-4≥P>10-6): D+P (where applicable) + To + EM + IR+ FSSE
D+P (where applicable) + To + (EM-D) + IR+ FSSE
Incredible events (P<10-6): loads related to these events do not require consideration.
Before the structure analysis ,the essential electromagnet analyses and calculation for the
magnetic field in TF system will be fulfilled.
The strengthened austenitic steel AISI 316LN is used in TF coils case (including PF coils
supports, TF gravity support stacked plates, joint bolts and keys between TF coils and LHe stub
elements), and TF conductor Jacket Material is AISI 316LN, too. AISI 316LN is a general purpose
austenitic steel which has found extensive application at cryogenic temperatures. The chemical
composition range is quite large and may need restricting to achieve the required mechanical
properties so that C + N wt%≥0.22. This materials shall be homogenous to the maximum extent
possible and be essentially fully austenitic, and magnetic permeability , for magnetizing forces not
greater than 30MAT (total number of Ampere-turns) shall not exceed 1.05 at 295K. AvestaPolarit
(Sweden) Asia Pacific Ltd supplied solution annealed hot rolled stainless steel plates (316LN) to
fabricate TF coil case and The φ22mm seamless 316LN stainless steel pipe is produced by N0.5
steel manufactory of Shanghai (China) for TF CICC conductor jacket.
The essential electromagnet analyses and calculation for the magnetic field in TF system
are fulfilled. The EFFI code, which was developed by Lawrence Livermore National Laboratory
(USA) in the early 1980’s, is used as the main calculating tool to evaluate the magnetic field in TF
of HT-7U. In up-to-date design the torodial magnetic field is produced by a set of 16 D-shaped
super-conducting toroidal coils, and each coil consists of 130 turns CICC. The requirement of the
TF system design is to produce Bt=3.5 Tesla at the position of R=1700mm. For minimization
expenses and optional construction decision in the future, the torodial magnetic field of 4.0 Tesla in
the center of plasma (R=1700mm) will be considered during the TF system design. According to
the geometric dimension of the TF coil winding, its analysis model is divided into 164 current
elements. By computing, for the toroidal magnetic field of Bt=3.5 and 4.0Tesla at the Plasma center
(R=1700mm and Z=0mm) the operating current of TF coil should be 14.3kAand 16.4kA separately
for different operating current separately. Magnetic field map of the TF coil based on the operation
current of 14.3077 KA is shown in Fig.4, it is shown the maximum field point 5.85 and 6.72 at
R=1120mm and Z=850mm. The ripples on the boundary of plasma in the mid-plane of the device
are 2.03% at R=1300mm and 4.55% at R=2500mm, and the ripple is 0.05% at the center of plasma .
The total inductance of TF system in HT-7U is 2.91525 Henry (disregarding to some current leads
and connection wires), so the total stored energy (We) of TF system is 298.39MJ and 389.71MJ.
The distribution of toroidal field and the value of ripple in midplane (Z=0mm) are shown at Fig. 5.
Fig.4. Magnetic field map of the TF coil
Fig.5. Bt and ripple along device equator
Each TF coil experiences a bursting force as well as a resultant centering force towards the
center of the machine. The resultant centering forces are reacted by the cylindrical vault formed by
the inboard straight legs of the TF coils. The bursting force on each coil is reacted by a combination
of the coil case and winding pack. The magnetic forces occur on the winding pack, and the winding
pack tends to be pushed outwards against the outer case ring. The overall coil centerline shape is
more vertically elongated than that of the idealized “bending moment force” shape for an 16 coils
TF system and it is actually chosen to satisfy the plasma ripple requirement and the vertical
divertor space.
The PF coils and the plasma create time varying out-of-plane (i.e. toroidal ) forces on each of
the TF coils. These forces vary around the poloidal direction. The resultant force on each TF coil is
zero but there is a resultant moment about the radius (the so-called overturning moment). The
overturning moments on the TF coils are supported by some shear keys in the support structure
between the TF coils.
If Bt=3.5, and 4.0 Tesla, the centering force acted on per coil of TF system is 989.4 and
1293.2 ton and the vertical tensile force per half of coil is 701.4 and 916.87ton, separately. Total
overturning torque distribution of TF coil at some severe time (from discharge start t=0.0s to end of
the shot t=17.64s) has been calculated , seeing Fig.6 and Fig.7. The maximum value is 288.3 and
332 ton·m at time=13.6S with different Bt of TF system. Fig.8 and Fig.9 shows the distribution of
out-plane loading normal to the center-line of the TF coil at time=13.64s and the different times.
As well known, Quench of TF magnet is a serious problem, because an enormous magnetic
field energy is stored in the TF magnet system. When a quench occurs, termination of
superconductivity is accompanied by conversion of the stored energy into a thermal energy which
is spent on heating of both the normal state zone in the TF coil. Thus the quench analysis for a
certain length CICC with 14.3KA operating and 5.85 Tesla background magnetic field becomes
very important. To analyze the quench of TF coil it is required to have more detail magnetic field
distribution along the conductor. Furthermore, to obtain magnetic field distribution along the
conductor and inside one single pancake of TF winding, the magnetic
filed magnitude at many
conductor of one TF winding single Pancake (Total length=100m) intermediate points at the
conductor have been calculated. The distributions are described by linear approximation. The
results are shown in fig.10 and fig.11. The influence of the strand transposition inside the cable is
not taken into account. Two peak values can be seen around 1m and 7m of each turn conductor
separately. There are two lower peaks
around 2.5m and 6m along the conductor.
Fig. 6 The distribution of in- plane and out-of-force in TF coil for Bt=3.5T
Fig. 7 Total overturning torque distribution of
Fig.8 The out-plane loads of TF coil at
TF coil at different time Fig.6
Fig.9
time=13.64s
Out-of-plane forces at discharge start (t=0.0s) and end of the flat top (t=13.64s)
in TF coil for Bt=3.5T
Fig.10
Magnetic field distribution along the TF
coil winding single Pancake (Total length=100.003m)
Fig.11 Magnetic field at turns in the TF
winding for Bt=3.5T
Since some necessary stress analysis and other analysis for TF system will go further and need
dates about electromagnet, some major results calculated for HT-7U TF coils in this paper are
summarized in Table.III
Table. III Main parameters of TF coils
Magnetic field at the plasma center (Bt) (unit: Tesla )
3.5
4.0
Maximum field at the coil (Bmax) (unit:
5.85
6.72
Tesla)(R=1.12m,z=0.85m)
Operating current (Iop) (unit: KA)
Total inductance of TF (unit: Henry)
14.3077
16.3511
2.91525
Total stored energy of TF (unit: MJ)
298.39
389.7084
The centering force of coil (unit: ton)
989.366
1293.196
The vertical force per half of coil (unit: ton)
701.375
916.767
The max overturning torque of coil (Time =13.64s) (unit:
288.366
332.0498
ton·m)
Ripple (R=1.3m,Z=0),%
2.0266
Ripple (R=1.7m,Z=0),%
0.0499
Ripple (R=2.5m,Z=0),%
4.5458
The TF magnet system must be designed, built and operation so that credible magnet system
failures, which could occur under normal and abnormal condition, cannot cause damage. The design
and analysis assessment of TF magnet structure with the specified safety factors, form the main
defense against the development of the faults during the operating life of the HT-7U device. However,
the possible fault conditions of TF magnet are still occurred. Except for manufacturing fault, design
error, electrical fault and so on, the most possible fault conditions is one or more toroidal coils failed
due to the quench of CIC conductor and the short-circuit of joints simultaneity occurrence. it based on
the hypothesis that the current of all the other coils is kept constant when one or more coils failed, and
non-symmetric loads in the TF coils have existed, and these forces distribution have been taken into
account.
The calculation of the TF magnetic field inside the NbTi cable-in conduit conductor (CICC) has
been carried out in its normal operation. When the operating current of TF coil is equal 14.3, and
16.4kA, the toroidal magnetic field at the Plasma center is 3.5, 4.0 and 4.5 Tesla. If one coil fails, the
TF magnetic field distribution will be changed, seeing fig.12. No doubt, in normal conditions (16
coils energized), disregard for the act of the poloidal fields current, the force on the toroidal coils is
almost only pure tensile force and the tangential force is zero. But in the fault conditions these two
kinds force obviously exist. When the current of one coil is zero and the others are full current, large
tangential forces generate will push one against the other. These forces reach the maximum value in
the two coils symmetrically located adjacent to the failed coil and are zero for the opposite coil. This
is shown in fig.13 and fig.14, in which tangential forces acted on the remaining coils are given.
Fig.12 The map of TF magnetic field when one coil fails, (Iop=14.3KA)
Fig.13 When one coil quenches, the electro-magnetic loads distribution of the TF coils
system (Iop=14.3077KA)
Fig. 14 Total tangential forces acting on the toroidal coils when one of them
quenches. N= 2(or 16) indicates the nearest coil.
Fig.15 The tangential forces distribution in per unit length on the
nearest coil (Iop=14.3077KA, Times=13.64s)
If more than one coil fails, the maximum value point of the toroidal magnetic field will decrease
with the number of failed coil increasing. These tangential forces will increase to the maximum value
when one half of the toroidal field system fails, then will gradually decrease to zero.
The in-plane loads (include the vertical component and the radial component) have also been
calculated for all t previously mentioned fault conditions. It can be seen that these forces reach the
maximum values in the equilibrium situation (all the coils energized), but when fifteen coils failed
they are equal the minimum values. In fact, whether in normal condition or in fault condition, the
total vertical component of in-plane loads in one energized coil is zero because of the symmetry of
TF coils system, and the total radial component is sustained by the central supporting cylinder, which
is formed by squeezed TF cases along the straight leg.
As previously mentioned, these calculations have been carried out in the hypothesis that, when
one or more coils fail, the currents of all the other coils is kept constant. This is the most reasonable
assumption, although it is very hard to achieve in practice operation of HT-7U. In fact, when any fault
condition occur, due to mutual inductance coupling between TF coils, the current distribution in the
TF coil system is not symmetric with respect to the fault coil. Thus, the analysis in detail needs to go
further. Meanwhile, this analysis must be confirmed with the final TF power supply and protection
configuration of HT-7U since this may prevent some of these fault conditions.
The superconducting TF magnet system is comprised of two variant parts: the case of the TF
coil and the winding that includes CICC conduits and insulating layers. The TF coil-case’s metallic
material selected is AISI 316LN as forged plates, its young’s modulus and Poisson’s ratios as a
function of temperature are presented in following Table.IV. Yield and Ultimate strength of 316LN at
4K arrive at 800MPa and 1100 MPa, and its fracture toughness (KIC) reaches 200 MPam1/2.
Table.IV Young’s modulus and Poisson’s ratio of 316LN
Elastic Property
Temperature (K)
295
Young’s modulus (GPa)
Poisson’s ratio
80
4
196
209
207
0.294
0.283
0.282
Because the winding of TF is multi-component, it is far from being isotropic. The effective
elastic characteristics of the winding are calculated by means of the theory of composite material. As
a result of the stress analysis is carried out for 4K operating condition, the material properties that
introduced into the calculations are given in:
316LN stainless steel case
the winding
E=207MPa
E=84.1MPa
υ=0.282
υ=0.3
In fact, the winding of TF may be regarded from the structural point of view as unidirectional
composites, and they are made up of a large number of superconductors in the axial direction, bound
together by insulating material. But the internal structure of the winding is very difficult to calculate
directly by finite element method because of its complex structures. Considering all these facts, the
model for simulating the effective elasticity modulus of the periodic superconducting coil cable is
based on the Homogenization method. Due to a typical repetitive cell of the winding, finite element
analysis of effective mechanical and thermal characteristics of micro heterogeneous superconducting
toroidal field coils are carried out by building the FE model of such a repetitive cell with a
three-dimension boundary condition of homogenization method. By the means of this method., the
effective orthotropic materials properties of the winding are seeing Table V.
Table.V the Orthotropic mechanical properties of the Winding at 4 K
The effective modulus
the winding of
TF
E1
(GPa)
22.4
E2
(GPa)
56.3
E3
(GPa)
21.7
υ12
0.109
υ23
0.273
υ31
0.336
Alf1
( %*)
0.327
Alf2
( %*)
0.311
Alf3
( %*)
0.340
* —
integral coefficient from RT to 4 K , where the symbols 1,2 and 3 mean:
1 —
normal to the winding direction (in the coil plane);
2 —
in the winding direction;
3
— in the toroidal direction (normal to the coil plane).
The TF coils are cooled down from room temperature to 4.5K, while maintaining a maximum
temperature difference of 40K between any two points in the TF magnet system. No structural
analysis has been performed to verify the level of thermal stresses during Cool down but the value of
40K is widely used in magnet operation (including the CS and PF coils) and is expected to be
acceptable.
The coil cooling down to 4 K was modeled. On cool down, the inward redial movements (xdirection) of TF coil are –0.23cm in the inner leg of TF coil and –1.015cm in the outer leg of TF
coil respectively, and the vertical movements (y-direction) of TF coil are –1.071cm in the most
upper section of TF coil without regard of the vertical contraction of TF gravity support. In a word,
by the composite result of redial (x-direction) and vertical (y-direction) movements of TF coil, the
coil maximum contraction is 1.25cm in the upper camber of the TF coil.
During operation, under the action of the in-plane electromagnetic forces yielded by the
interaction of the toroidal magnetic field and the current in TF coil itself, the D-shape TF coil will
be changed into the “circularization” due to the dominant breathing deformation of the inboard and
outboard legs in the coil. This data has been calculated within the finite element model of the TF
coils. When the current of TF coil is 14.3KA, the coil deformation (excluding the thermal
contraction deformation) under the in-plane electromagnetic loads is shown in Fig.16, and the inner
leg moves inward up to 0.02cm and the maximum outer leg outward displacement amounts to
0.073cm.
Fig.16 The TF coil deformed shape due to the “in-plane” loading only
The radial and vertical displacement (cm) of the equatorial plane inner and outer leg of TF coil is
shown in table.VI
Table.VI
Equatorial Plane Inner Leg
EQUATORIAL PLANE OUTER LEG
Cool Down
TF-only
End of Flat Top(t=13.64s)
Urad
-0.23
-0.0223
-0.0223
Uvert
-0.5355
0.09894
0.09894
Urad
-1.015
0.07262
0.07262
Uvert
-0.5355
0.09894
0.09894
Because of the complex TF magnet structure system bear the three dimension
electromagnetic load in normal operation, the more detail three dimension model is built using the
FE code. In order to improve the accuracy, the method, which mixing 8-node isoparametric
hexahedral solid element with 4-node tetrahedral solid element, is used to construct the 3D FE
model, where the winding and the case of the coil are represented by missive element with
isotropic material characteristic. The elements of both components between winding pack and the
TF coil case are fixed at the same nodes, which means that friction effects between the winding
pack and casing are not taken into account. This FE model is provided with 2520 hexahedral and
tetrahedral elements and 4044 nodes, and the behavior of the whole TF coil is achieved by suitable
boundary condition, namely the centripetal in-plane loads are resisted by wedging the coils along
their side walls at the inboard straight leg, and the out-of-plane loads are resisted friction faces
along the inboard legs. The complete FE model used for post-processing is represented with
associated mesh for the TF coil. The cases of load have static load (dead loads), in-plane loads due
to the self field of the TF coils and the out-of-plane loads due to the interaction of the TF coil
current with the polodial magnetic field. By computing, the maximum out-of-plane force will occur
in end of Flat Top (t=13.64s). The most critical area is found to be the casing upper inboard curved
region where the combination of in-plane loads and out-of-plane loads is particularly severe, and
the maximum Von·Mises stress and deformation is shown in Table.VII, fig.17.
Table.VII
The case of TF coil
The winding of TF coil
3.5 → 4.0（Tesla）
Loads
(Magnetic field at the plasma center)
The
maximum
·
Von
Mises
stress
（MPa）
The maximum
deformation
（cm）
The
maximum
Von·Mises stress
（MPa）
load 349.5 → 420.8
Without
In-plane
regard to
+out-of-plane
Dead load
load
Add Dead
In-plane
load
+out-of-plane
load 352.3 → 424.5
1.027 → 1.196
220.6 → 264.8
1.029 → 1.198
222.2 → 266.5
load
Fig.17 The distribution of Von·Mises stress and deformation of TF coil
under charged (Bt=3.5T)
The 3D FE model as described in this section does not provide information about local
stresses in the radial plates, conductors and insulation. To obtain information about the in-plane
behavior and the stresses in the radial plates, a detailed local 2D finite element model of the TF coil
inboard leg cross section has been developed and analyzed, and this model is a fully-bonded FE
model.
The radial plate design is of wedge-shaped across-section. It consists of the four stainless
plates whose thickness is not equal. 130 turns CICC conductor embedded into the radial plate, Each
conductor consists of a square 1.5mm thick 316LN jacket wrapped with half lapped of glass cloth
type built up to a thickness of 0.5mm, between pancakes 1mm thickness glass epoxy sheets are
inserted. There will be a ground wrap of half-lapped glass cloth type built up to a thickness of 6mm.
All of gaps, such as: the 0.5mm gap between conductors and the 6mm gap between winding pack
and coil case and so on, are filled with epoxy resin by the vacuum-pressure impregnation. The
current in the conductor is 14.3 kA, The thickness of the nose is 63.3 mm.
As a result of symmetry, the out-of-plane loads is zero in the TF coil inboard leg cross section at
equatorial plane, and the in-plane Lorentz forces have been calculation. The distribution of in-plane
forces along radial direction in this section is shown as Fig.18. Fig.19 shows the 2D finite element
mesh of the TF coil cross section at equatorial plane, the material properties used in the analysis are
tabulated in Table. VIII.
Fig. 18 The distribution of in-plane forces along radial direction in this section(Bt=3.5T)
Fig.19 The 2D finite element mesh of the TF coil cross section at equatorial plane
(the right: fractionated gain)
Table.VIII Material properties used in 2D FE model of the TF coil cross section at 4K
Material
E（GPa）

Alf (%)
Cu
138
0.338
-0.295
NbTi
82
0.333
-0.188
207
0.282
-0.306
62
0.210
-0.120
The Case of TF coil
and the jacket of CICC
Insulation
The local stresses in the nose of the casing, the TF conductor jacket, radial plates and
insulation system of the TF-coil cross section at equatorial plane have been calculated in more
detail by using a 2D finite element model. The most critical stressed jacket and insulation region is
identified as that as the nose of the inboard leg. Where the TF coil centring force is transmitted
through the plates and ground insulation to the casing. The radial centring force is reacted by
wedging resulting in a toroidal compressive force. The jacket and insulation stresses are broadly
compressive.
The local detail analysis of the TF coil cross section showed that the maximum stress intensity
and maximum toroidal compressive stress occur at the corner of the casing at the inner radius. The
maximum Von·Mises stresses at this region is 279.09MPa, and the maximum deformation which
occur at the middle of the nose is 0.38573cm. These results are shown in Fig.20. At the same time,
the maximum shear stress in the insulation of TF coil is 93.5MPa and is very localized, and this
shear stress is higher than the static allowable shear stress of the insulation. In order to reduce the
shear stress in the insulation, some modifications that the winding pack can separate from the
casing must be done, and up-to-date computing result has shown that the maximum shear stress in
the insulation is 31.5MPa, this result is within acceptable levels.
Fig.20 The distribution of Von·Mises stress and deformation of the 2D FE model (Bt=3.5T)
To obtain input for detailed local model, such as: the poloidal shear keys, the toroidal support
parts and so on, A detail 3D linear finite element model of the TF coil magnet system has been
development to model the intercoil structures, the TF casing and the winding pack. The winding
pack is modeled with 3D solid element with orthotropic isotropic material properties. Fig.21 shows
the 3D finite element model. The total 3D FE model is linear, and the size of the model: No. of
nodes: 11413 and No. of elements: 5946. A suitable boundary condition is adopted to achieve the
behavior of the single TF coil under charging. As a conservative estimation, it is assumed that the
out-of-plane loads(t=13.64s). are taken up by the straight part of the inboard leg and TF toroidal
support parts, and friction effect in inboard region and TF toroidal support regions is not include in
the model. In this load condition, the maximum Von·Mises stress is 148.22MPa, and occur at the
red region of Fig.22. In the design of TF coil, most of part of the out-of-plane forces will been
taken up by eight poloidal shear keys which are distribute uniformly in upper and lower toroidal
support regions. The dimension of poloidal shear key’s cross-section is 60mm×60mm. The
poloidal keys have been modeled with the help of 3D solid element. The result of the stress
analysis shows the first poloidal key bear the maximum Von·Mises stress, and this value is
144.76MPa, seeing fig.23.
Fig.21 3D global linear finite element model of TF coil magnet system
Fig.22 The distribution of Von·Mises stress of the 3D FE model
Fig.23 The distribution of Von·Mises stress of the TF coil (the right: fractionated gain)
5.2 PF System
The poloidal field(PF) system was consisted of fourteen superconducting coils, including 6
pieces central selenoid coils, 4 pieces divertor coils and 4 pieces outer-big-rings. Its are used with
two types superconducting conductor which size and the configuration of cabling are slight different.
One is 20.4×20.4 mm in size, 4 stages cabling and first stage with 2 superconducting strands add 1
copper strand. It was used for selenoid coils and divertor coils. Another is 18.6×18.6 mm in size, 4
stages cabling and first stage with 1 superconducting strand add 2 copper strands. This type CICC was
used for 4 pieces outer rings only.
Figure 1 is shown the cross section of the PF coils for HT-7U tokamak. Figure 2 is shown the
CICC’s configuration, cabling and size. The NbTi cable-in-conduit conduct (CICC) cooled by
supercritical helium at 4.5 K is chosen as superconductor for all of the PF magnets. The PF system
was supported by the case of TF coils.
2890
454
1809
994
820
711
550
PF13
1703
1488
PF3
PF1
PF11
1710
1487
50
PF5
3324
1339
PF9
PF7
2085
126
126
101
10
1199
952
3235
3019
Plasma
Fig. 1 The cross section of the PF coils for HT-7U
The maximum capacity of the volt seconds for PF is about 10 Web and the stray field in
plasma Initiation region is less than 35 gauss. The peak magnetic field in the body is about 4.5 tesla.
Four type modes for shape of plasma can be chose during the operating. The plasma current is
about 1MA and the duration of the plasma is about 10 seconds for ohmic heating discharge
according to the design of the PF system.
The PF coils are pancake wound only. We have developed two special machines for the
superconducting magnet pancake winding. It can be easily to pre-winding the square conductor
CICC that is quite strong and difficult to winding in normal way. The CICC was wound start
from the inner circle to outer circle and continue from outer to inner. For example, each solenoid
coil with seven turns in radial and 20 turns in height, so it should be continue wound only use one
piece CICC from inner to outer and outer to inner ten times.
The boring machine is another tools for liquid helium inlet on the CICC. To welding the
liquid helium tube is also should be carefully during winding. Some inspecting and leakage
detection should be done after these two processes to find any quality problem that have to be
avoid fully.
It should be satisfy on operating in each case of the machine for the structure ability of the
coil insulation. The highest voltage on PF coils probably during the plasma initiation and
disruption is about 2400 V on the diverter coils.
The turns electrical insulation consists of kapton with 2 layers and multiple layers of
fiberglass. During winding, the conductor was wrapped with 0.1 mm in thickness of fiberglass at
first. Then 1 layer fiberglass and 1 layer kapton be half-overlap wrapping. 2 layers fiberglass with
detection wires wrapping for the turn insulation at last. Between the pancakes, insert a 1.0 mm
fiberglass carpet to improve the pancakes insulation. In the end of winding was wrapped a 5 mm
in thickness of the fiberglass as ground insulation. The current connections and liquid helium
tubes were set inside the center of the solenoid but except the two pairs of outboard ring coils for
which were through outer of the case of toroidal coils. Even liquid helium inlets were through the
center of solenoid where the field is quite high but the structure will be sample.All the PF coils
will be done vacuum pressure impregnate (VPI) to fill with epoxy resin and solid the windings to
improve the capability of the insulation and structure strength. The conductor turn insulation is
subject to some shear stress and tensile stress normal to the insulation layer.
For the HT-7U PF coils system there are three requirements which are the capability of
ohmic heating volt-second, the stray field in region of plasma initial and the equilibrium field for
plasma. The concept of the PF system design is to reduce the total ampere-turns and the magnetic
energy stored in the system to be used for both buildup plasma current and equilibrium for it. The
first step in defining each PF coil location is to calculate the maximum radial position and size of
the solenoid. The overall layout of the PF system for HT-7U is shown in table 1. In this table, Ioh is
the current in each coil at discharge moment of ohmic heating system, R1 and Z1 are the first turn
center position in R and Z directions, Nr and Nz are the turns number for each coil in R and Z
direction, Bw and Bh are the distance between two neighbour turn in R and Z direction.
Table 1. Coil current and PF coils parameters
No. Coil
Ioh(kA)
R1(cm)
Z1(cm)
Nr
Nz
Bw(cm)
Bh(cm)
PF1
11.4
56.3
3.085
7
20
2.17
2.27
PF3
12.7
56.3
52.385
7
20
2.17
2.27
PF5
11.6
56.3
101.685
7
20
2.17
2.27
PF7
5.73
97.2
172.02
11
4
2.17
2.27
PF9
5.73
97.2
182.1
17
12
2.17
2.27
PF11
3.15
290.54
170.5
6
10
2.17
2.27
PF13
.694
325.
83.35
4
8
2.17
2.27
In nominal cause of operating, the plasma current of HT-7U is 1 MA. The machine is capable
of achieving this plasma current for at least 10 seconds by utilize inductive drive alone. This is
accomplished with 10 web capable of PF coil system during operating. The major radius of the
plasma is 1.78 meter and the minor radius is 0.4 meter. A slight change was made to accommodate
better conditions for the mechanical design of divertor and vacuum vessel.
The ohmic heating system for HT-7U tokamak device was designed to satisfy the capability
of flux swing and stray field, even the plasma initial can be helped by lower hybrid wave or
electron cyclotron wave. The ohmic heating system should be provided with enough flux swing and
very low stray field in the region of the plasma breakdown. An about 10 volt-seconds and quite
lower stray field was designed in PF coils system. The ohmic heating stray field is shown in table 2
and figure 3. The PF coils system design has been made as compact as possible. To minimize
ampere-turn of the PF coils and power requirements and the overall device radius, the parameters of
the ohmic heating system have been optimized to achieve maximum volt-second capability and to
satisfy the constraints for structural and superconductor performance. The peak field on PF coils is
about 4.5 Tesla in the final design.
Table 2. Ohmic heating stray field
R(cm)
Z(cm)
Br(T)
Bz(T)
138.
.0
.0
-0.00118
170.
.0
.0
.0004
178.
.0
.0
.00053
194.2
.0
.0
.00045
250.5
0.
0.
-0.00291
Z (cm)
R (cm)
Fig. 3 Ohmic heating stray field and flux contour
HT-7U will be operated in a circle section of the plasma, but a non-circle cross
section with a high elongation is a prefered in the future of the operating. The PF system
should support the inductive operation under the plasma burn pulses lasting about 10
seconds without EC wave or other conditions. Extension of the burn pulse will be
towards upto 1000 seconds as the steady-state. HT-7U operation baseline will keep the
plasma current of 1MA with the elongation kx is about 1.8 to 2.0 and the triangularility δx
is about 0.6. The HT-7U plasma will be shaped and controlled by PF coils system only in
a pulse length of 10 seconds. There are two candidate modes of operation of the HT-7U
plasma. In the basic mode, the plasma minor radius is 0.4 meter and the major radius is
1.78 meter. In other mode, the plasma minor radius can be 0.45 meter and the major
radius can be 1.91 meter, where the required currents for plasma control are moderate and
the corresponding out-of-plane loads on the TF coils are structurally tractable. The value
of the poloidal beta is about 1.6 for two modes. The calculation results are shown in table
3 and the plasma shapes are shown in figure 4. In table 3, Rx is the distance from X-point
to the center point of the tokamak device in horizontal and and Zx is the distance from
X-point to the equator plane of the device.
The plasma geometry and the position are given from EQT(EQuilibrium of Tokamak)
code and PFFIFFPS code. The PF coils system should be satisfied with the requirement of
physics design and engineering constraintions, in which the PF coils system should be
satisfied with (1) the capacity of the volt-second is about 10 web, (2) the stray field of the
ohmic heating system in the region of plasma initial is lower than 5×10-3 T, (3) maximum
field is limited lower than 5 T at the superconducting magnets and (4) the current of the
CICC conductor can’t be more the 15 kA on the field of magnets of 4.5 T.
Table 3. Parameters of three type plasma shapes
R(m)
a(m)
kx
δx
Rx(m)
Zx(m)
Big-K
1.798
0.418
1.92
0.516
Circle
1.95
0.545
1.0
0.
Big-V
1.91
0.448
1.79
0.62
1.583
0.804
1.632
0.801
DN
cm
PF3
Ip=0
BetaR=
a=
Ipf1=
Ipf2=
Ipf3=
PF1
Ipf4=
PF2
Ipf5=
Ipf6=
PF4
Ipf7=7
DN
PF9
PF7
PF5
a,
a,
PF11
PF12
PF6
Big-K double null
Ipf9=6.3
Big-V double null
circle
Fig. 4
Ipf10=3.6
Single null
plasma shapes of HT-7U
Ipf8=1
PF10
PF8
Ipf11=-9
Ipf12=-9
Figure 5 is the inductive discharge waveform. After a fixed OH current and plasma
cm
shape, then the current waveforms for PF coils are fixed. The field on coils is quite
different in different time points from the current change during operating. The peak field
on the PF coil occurs near the ends of the solenoid on the inboard sides and the allowable
maximum field at the superconducting PF coil is 4.5T. Figure 6 shows the field on each
PF coil in the inductive discharge case. The field maximum varying rate for PF coil is
about 7T/s and time duration is about 60ms. A well-controlled startup was obtained with a
loop voltage as low as 8 V, which among tokamaks is quite a challenging technology in
experiment.
B (T)
T-No.
Fig. 5 Typical inductive discharge scenario
Fig. 6
PF Coils Field in
(R=1.79m, a=0.41m, kx=1.84, δx=0.58)
Operation
All the PF coils will be attached to the case of toroidal coils that will be support the vertical
gravity loads and magnetic forces. The solenoid coils will be bolted together with preload is about
680 T force in total and then tight connect by two ends with the front of the toroidal coils case
through the fiberglass plates. The center solenoid is self-supporting against the coil radial forces
and most of the vertical forces, with the support to the toroidal coils reacting only the weight.
The other coils were sample support by the case through the supports connect parts. During
assembly, these connect can be adjustment easily and inspective quickly. The current in/out and
helium tube were wrapped insulation material individual down to the place where the insulator set
at the bottom of the machine.
The CICC joints will be used for all PF coils if need. It was about 350 mm in length and
50mm70 mm in cross section. It was set at the bottom of the machine for each coil connects with
the current leads. There are other joints for the coil need the CICC longer than 600 m that will be
set nearby the magnet, but the type are same with above. We have made serial joints prototype and
will testing the later. Figure 7 is show the inlets for current and liquid helium on solenoid coil.
Fig. 7
The inlets for current and helium tube on solenoid coil
The main load on coils and structural are the pre-loads force and the magnetic hoop force,
which creates tension in the structural material. The result from analysis was shown all PF coils
were safety enough during operation at high field that means the stresses and movements were
susceptible. For solenoid coils, a peak stress is about 300 MPa and the movement on the structure is
about 6 mm. The detailed analysis for the solenoid has been carried out and the result is shown in
Fig. 8 and Fig. 9.
Fig. 8 Analysis for solenoid on stresses and movement
Fig. 9 Analysis for the stresses on outboard coil and the insulation on solenoid
The engineering design of the PF coils system of HT-7U was finished that is satisfied with
the requirement of physics design and global constructions. It was being fabricated in industry
factory now. The solenoid coils and diverter coils will be testing one by one in CASIPP.
5.3 Vacuum Vessel
The vacuum vessel is torus-shaped with “D” shaped cross-section, double wall, upper
vertical ports, lower vertical ports, horizontal ports and flexible supports. It shows as Fig1. The
vessel is symmetrical by equator. Overall exterior dimensions of the vacuum vessel are 2.63m
height with 1.95m inner radius and 2.75m outer radius. The torus consists of 16 segments and each
segment consists of inner shell, outer shell, ribs and ports. Two ribs separate outer shell and inner
shell, and give the required mechanical strength. The ribs are skipping welded to inner shell and
outer shell. Other two ribs (end ribs) are tight welded both to inner shell and outer shell at segment
end. Every two segments are welded together by end ribs. Eight low rigidity supports are connected
to lower vertical ports alternatively.
Fig.1 Configuration of Vacuum Vessel
The space that made up by shells and ribs will be filled with boride water both for vessel
cooling and reduce neutron radiation. One octant is a cycle unit. Water inlet is on inner shell at
vessel bottom, boride water will flow along the channels that made up by ribs and shells, then
through holes on ribs go to outlet. During bake-out hot nitrogen gas go through the same way.
The principle of choice of material for vacuum vessel has a significant influence on
performance, fabrication characteristics, mechanical strength at operating temperature, chemistry
properties, and low cost relative to other candidates. Compared with Ti6Al4V, Inconel, SS-316LN
and SS-304L, 316L stainless steel has lower cost, good mechanical properties, good chemistry
properties and good fabrication characteristics in different extent. It is selected as main material of
the vessel. Table 1 shows the main parameter of vacuum vessel and material.
Table 1 Vacuum Vessel Parameters
Size:
-Toroidal Extent of Sector
16°
- Shell Thickness
8 mm
- Rib Thickness
15 mm
Material
SS-316L
Surface Area/Volume:
- Interior Surface Area(Include
162.4 m2
Ports)
36.1 m3
-Interior Volume
5.2m3
-Interlayer Volume
Mass:
- Main Vessel
13.2 Tons
-Ports, Flanges and Supports
20.1 Tons
- Shielding Water
5.2 Tons
- Total
38.5 Tons
Resistance:
-Toroidal
~85.4 μΩ
- Poloidal
~20.7μΩ
Considering symmetrical of vacuum vessel. A model of 1/16 vacuum vessel shows as Fig2 is
used for structure analyses. Bellows on ports and low stiffness supports are considered as spring
components. At each end of ports in X, Y, Z directions proper rigidity was applied on. The value is
showed as table2. Rotate freedom degrees of each port end ware not confined. From the working
conditions as described in next paragraph it can be seen under load condition 3 and 5 that the stress
is quite high. In these cases thermal deformation is the main reason that cause peak stress at each
port end is higher than that under other load conditions. But detail analyses indicate peak stress was
caused by stress concentration.
Fig2 Structure Analysis Model
Table 2 Stiffness of each end of port (N/mm)
Vertical
Horizontal
ports
ports
X
1492
2630
1603
Y
32
2577
58479
Z
1680
160
146627
Directions
Supports
The vacuum vessel must withstand many individual and combined loading conditions during
vacuum test, normal and off-normal operation. Because the vessel is double wall, several different
working conditions must be considered. During vacuum test space between double wall will be
pumped into vacuum, during plasma operation space between double wall will be filled with
shielding water, and during bake out hot nitrogen will flow through the interlayer. Being
superconducting tokamak when the device in operation both inside and outside of the vessel will be
pumped into vacuum. While plasma breakdown and disruption electromagnetic forces is much
strong. Table 3 shows different load conditions, load values and stress values.
Table 3 Different load conditions
Load
condition
1
2
3
4
5
vacuum
0.1MPa
vacuum
0.1MPa
vacuum
0.1MPa
vacuum
vacuum
0.1MPa
vacuum
0.1MPa
0.1MPa
0.1MPa
vacuum
0.2 MPa
300K
300K
523K
300K
367K
Pressure
Inside vacuum
vessel
Pressure
Outside
vacuum vessel
Pressure in
interlayer
temperature
Halo current load
EM load
no
no
no
no
and eddy current
load
Peak Stress
82.5 MPa
82.8 MPa
561 MPa
48.8 MPa
492 MPa
The vacuum vessel is considered as thin wall vessel. To check its stability is necessary. The loads
affect stability including dead weight, atmospheric pressure, borated water pressure and EM loads.
Atmospheric pressure and borated water pressure is uniform and EM load is non-uniform. Peak
EM load caused by eddy current in toroidal direction is 3.95kg/cm2, in normal direction is
3.9kg/cm2, in poloidal direction is 3.9kg/cm2, and peak EM load caused by halo current is 2kg/cm2.
Analysis considered all load conditions described in paragraph 2.2.
Linear buckling analysis indicates the double wall structure is safe enough for stability. Table4
shows different thickness of vacuum vessel wall can with stand different buckling load.
Table 4 Buckling load with different thicknes
Model cases
Wall thickness (mm)
Buckling load(MPa)
1
8
1.26
2
9
2.09
3
12
3.45
4
15
6.41
5
20
14.49
The vacuum vessel should bake-out to 250ºC to degas helium and hydrogen in the structure
material. Because vacuum vessel is double wall structure, hot nitrogen will be employed to go
through the channel, which made up by vacuum vessel inner shell, outer shell and ribs for the
vacuum vessel bake-out. Considering the vacuum vessel symmetric one octant was select for hot
gas recycling unit. A thermal analysis model of one octant has been set up and analysis was
accomplished. Fig4 shows the temperature distribution when the highest temperature on the model
is 550K. The average temperature on the octant is 536K and the maximal temperature difference is
27K. This is satisfactory the requirements of vacuum vessel bake-out. Table5 shows the parameters
of hot nitrogen gas and temperature situation. During plasma operation the vacuum vessel will be
maintain 90~95ºC. Boride water recycling system is possible to be used to adjust temperature of
vacuum vessel
Fig4 Temperature distribution during bake-out
Table 5 Parameters of hot nitrogen
Inlet
Inlet
Outlet
temperature
pressure
pressure
550K
0.2MPa
0.18MPa
5.4 Cryostat with Thermal Shields
A cryostat vessel (CV) with a thermal shield of 80 K houses the TF and PF assembly along
with the vacuum vessel and supports all loads from them.
The cryostat vessel of HT-7U is a cylindrical configuration which is divided into a cap
structure, a middle ring and a base structure. The cryostat and thermal shields is made of 316L
stainless steel.
The thermal shield comprises the internal shield (ITS) covering the vacuum vessel, the
external shield (ETS) covering the walls of the cryostat (bottom, cylinder and upper head), the
vacuum port shields (PTS) enclosing the vacuum vessel port ducts. ITS has a noncircular
cross-section and shields the thermal radiation from the vacuum vessel. ETS consists of three parts
including an upper structure, a middle ring and a lower structure with support structures for
shielding the thermal radiation from the cryostat vessel walls. The thermal shield is made of a
sandwich structure with square tubes welded on, which encloses all of the superconducting
magnets of HT-7U.
ITS has to closely follow the shape of the VV, for space reasons, and is therefore of a
segmented, toroidal design. It is composed by 14 of 22.5° sectors and 4 of 11.25° sectors. The
sectors are connected each other to form a D shape cross-section torus to enclose the vacuum vessel.
ETS is divided into three parts: upper cap, middle cylinder and bottom platform. Each part of ETS
consist of 8 octants. ITS and ETS is connected each other by the 16 horizontal PTS and 32 up and
down vertical PTS to form a rigid self-supporting structure under its own gravitational and thermal
loads and is attached to the bottom of cryostat by 4 inboard and 8 outboard supports. The inboard
and outboard supports are stainless steel multi-plate type structure, it is allowing radial movements
whereas, are used to fix the toroidal position of the thermal shields. For the reason of the space
limited, a considerable effort has been expended on keeping the design of the structure as slim as
possible. Therefore, the structure of the thermal shield is made of double-wall panels, sandwich
structure consist of two stainless steel panels and weld quadrate cooling pipe in between see Figure
4., for strength reasons and additionally for reducing the radiant heat loads on the magnet structures
without overly complicating the cooling tube layout, by interception of heat loads from panels
facing the vacuum vessel and keeping the panels facing the coils relatively cold. The main
parameters of the thermal shields are listed in Table 1.
Unit
VVTS
CTS
VPTS
Surface area
m2
76
166
66
Total mass
ton
4
12
4
Out diameter
m
2948
7.122
Inner diameter
m
2.28
Height
m
5.662
Wall thickness
mm
25
25/40*
25
Panel thickness
mm
3
3/5*
3
Cooling pipe dimension
mm
19x19x2
Maximal distance between
mm
200
two pipes
Number of cooling channel
16
24
Cooling channel length
m
32.6
He gas mass flow rate
g/s
110
Maximal Pressure drop
bar
0.4
16
To minimize the heat load received from the warm surfaces and to reduce the heat load
radiated to the 4K surfaces, the thermal shield panels are polished and the multi-layer insulation
will be used for ETS. While the thermal shield performs no safety function, their repair or
replacement, particularly of the ITS, would involve dismantling major parts of the TF magnet, VV
and other in-cryostat components. Further enhancements against failure are obtained by having
electrical breaks incorporated in ITS sector joints, reducing the electromechanical loads on the
structure.
Analysis of the thermal loads is very important for ensuring that the thermal shield can
handle the loads both locally, without excessive stress and thermal traction, and globally to
interface with the cryogenic plant within the specified boundary conditions, and in particular, to
ensure that the cryogenic plant can handle the heat loads in a cost-effective manner. The main
results of thermal load analysis for both the “hot” (100° C) and cold (room temperature) vacuum
vessel wall plasma operation state and VV baking state (250° C), together with the associated
thermal shield thermal hydraulic data, are summarized in Table 2.
Table 2.
Unit
Thermal radiation load
kW
Maximum outlet He temperature
K
Average panel temperature
K
Maximum panel temperature
K
ITS+PTS
ETS
300° K
373° K
523° K
300° K
11.7
28
106
15.2
110
90
151/139
95
(1)
Values for panel of the ITS towards VV and TF coil respectively.
(2)
The surface emissivity is assumed 0.2 for heat load and temperature estimation.
There is limited cooling capability due to fixed helium compressor and turbine expander in the
refrigerator. Pressured helium gas from the main cryogenic plant, with an inlet temperature and
pressure of 57K and 0.52MPa respectively, is used to cool the thermal shield system. The total
helium mass flow rate for all thermal shields is limited not more than 110 g/s and the total pressure
drops are limited within 40 kPa.
Therefore, cooling circuit is designed as follow: the 16 ITS segment and 16 up and horizontal
VPTS are connected in parallel and formed first group (32 parallel channel) and the 8 octant of
upper cap, 8 octant of middle cylinder and 8 octant of bottom platform of ETS plus bottom PTS are
connected in parallel and formed second group (24 parallel channel). The two groups are connected
in serious. Pressured helium gas from refrigerator is used to cool the ITS and PTS group at first.
The outlet temperature of helium gas is too high after cooling of ITS and PTS, when the vacuum
vessel wall is working in hot and baking state and has to be recooled by a liquid nitrogen
pre-cooled heat exchanger again, and then can be used to cool down the ETS. The cooling circuit
schema is shown in Fig 5. The heat loads to cold mass (radiation from surface mainly) are
somewhat higher than earlier estimates for VV hot wall operation and it makes the average
temperature of panel higher and will increase the heat load on the magnet system. Furthermore, the
large temperature gradients in the panels lead to considerable stress. Therefore, a thin low
emissivity layer of silver coating covered on ITS inner surface are being studied to reduce these
loads. Figure 6 shows the temperature distribution in ITS at VV hot wall operation. Now, a
prototype of 1/16 VVTS, up and bottom vertical and horizontal PTS and 1/8 of ETS bottom
platform have been made to check the manufacture feasibility. It shows that ETS bottom platform
has considerable dimension deflection after weld and has to be reinforced.
5.5 Divertor and In-verssel Components
The HT-7U superconducting tokamak incorporates a double null divertor that maybe operated in
a single or double null mode. Provisions are incorporated to provide for the capable of radiative
divertor operation to reduce the peak heat flux to the divertor target plate. Particle pumping is
provided through vertical ports to control impurity and the plasma density.
The divertor is designed for steady-state thermal operation. HT-7U pulse lengths will be dozens of sec. on the first phase
(about 5 years) and dozens of sec. to 1000 sec. on the second phase. The materials used for the dicvertors are mostly SS-316L
for the structure and cooling water folds, Cu-Cr-Zr alloy for heat sink and coolant tubes and normal graphite for the plasma
facing material during first phase.
The divertor of HT-7U consists of outer target plate, inner target plates, domes and baffle
plates in both the top and bottom of the vacuum vessel. Each part of the divertor was separated to
16 segments toroidally (22.5 sector). This was done to allow installation/remove of the divertors
within the space limitation of the plasma chamber. Individual cooling system of each sector is
connected to cooling water fold through removable fitting.
DIVERTOR OUTER PLATE
DOME
PASSIVE PLATE
DIVERTOR
INNER PLATE
FEED BAKE COI
MOVABLE LIMITER
LARGE VOLUM
PLASMA CONFIGRATION
BIG ELONGATION
PLASMA CONFIGRATION
CIRCLE PLASMA CONFIGRATION
cross-section of HT-7U PFC
Design of the HT-7U tokamak divertor is based on long pulses and steady state operation. The final goal of the pulse
lengths of HT-7U will be 1000 sec. Analysis indicates the peak heat fluxes of the plasma facing surface of outer target is
2.05MW/m2 when operating in normal divertor mode, and 1.00MW/m2 when operating in radiative divertor mode (The input
power up to 6.5 MW, that is the Max. input power of HT-7U of present plan for first phase). On first phase the material of
plasma facing tiles limited the temperature of plasma facing surface. It must be below 600C. When the temperature is above
600C, amount of carbon sublimated from the tiles give to main plasma, as impurity will be unacceptable.
The general arrangement of the divertor system shows as Fig. 1. W-shaped divertors provide a closed gas box. Particle
pumping is provided through 16 upper vertical ports and 16 lower vertical ports and provide total 70m3/s pumping speed for
both top and bottom divertor to control the plasma density. Each divertor system is comprised of 16 toroidal sectors of 22.5
respectively.
Divertor modules are required to be relatively well seal to prevent neutral particle migration behind the modules and
reentry into the main plasma. This required gaps of each sector of the divertors adjusted to proper width and secondary particle
shield will be used when it is believed necessary. Adequate pumping speed of divertor particle pumping system is necessary for
control plasma density.
An adequate cooling system has been designed to remove heat load come from plasma. This insures temperature of plasma
facing surface below the allowable level (800C). If the temperature were above 800C graphite of plasma facing tiles would
release unacceptable amount of carbon as impurity to main plasma.
Radiation divertor operation may require that a target gas to be introduced in the throat area of the divertor. This is
accomplished with independently controlled gas feed lines, which are routed to supply gas through several gas supplying tubes
and formed a high radiative region to mitigate high heat flux concentrating on small area of plasma facing surface of the
divertors.
Each module must be removable and can be reinstalling in plasma chamber, all of the graphite tiles are designed able to be
replaced in plasma chamber. The size of each module ensures it can be taken out and in vacuum vessel through maintaining
port. Coolant system of each model is possible to be removed from and reconnected to water main fold.
Module edge-to-edge alignment is critical to limit peak heating of plasma facing surface of the targets. This alignment
must be within 1mm. It will be achieve by the fabricated accurate of heat sink and the alignment of each sector of heat sink. Heat
sink will be installed and aligned to the measured field in order to distribute the intense heat flux uniformly.
Taken into account of the outgassing rate of graphite tile. Bake-out temperature of divertor must be above 350C. Hot
nitrogen through cooling tube is used to heat all PFCs.
Vacuum vessel of HT-7U has major radius of 1.7m, overall inside height of 2.45m and radial
width of 1.4m. The “X” point coordinates in the radial direction is around 1.55m and in the “Y”
direction is around 0.8m(origin is at the center of the machine). From Fig .1 it can be known the
cross-section of the divertors. The W-shaped divertor consists of inclined targets, domes in the
private flux regain and baffle plates connected to target plates. The structures of the divertors are
segmented toroidally. The inclined target, domes and baffles are divertor separated to 16 segments
respectively in toroidal direction.
The coolant tubes are 12mm o.d. and 10mm i.d. Thermal performance in area of highest heat flux the distance of the
coolant tubes are reduced to minimum. The coolant tubes are made of Cu-Cr-Zr alloy and brazed to heat sink, which was made
of the same material of the tubes. All coolant tubes are paralleled and connected to secondary cooling water folds by welded.
Secondary cooling water folds are connected to main fold by removable fitting. Plasma facing surface consists of graphite tiles,
which are bolted to heat sink. Between the graphite tiles and heat sink flexible graphite foil with very thin thickness is used to
improve thermal conduct. It will give a better thermal contact of graphite and heat sink. Fig 2 shows the different layers of the
divertor. Each segment is bolted to SS-316L supports and the supports welded to vacuum vessel. All the supports and heat sinks
are designed proper to against electric-magnetic force caused by plasma disruption and halo current. Divertor pumping system
for upper and lower divertor is through upper and lower vertical ports to control the plasma density. Total pumping speed is
70m3/s. Fore passive plates are located at upper part, lower part and outer part, inner part of the
vacuum vessel. At outer part and inner part, upper plate and lower plate is connect as saddle loop
respectively.
Fast feed back coil is arranged inside the vacuum vessel, located at both upper and lower part
of the vessel. Each coil consists of two turns, and the upper coil is connected to lower coil make up
saddle coil.
On the first phase the input power will not more than 6.5MW. A kind of normal graphite tile are used for plasma facing tiles is
reliable. On the second phase when plasma heating system and plasma current driving system is update. C-C composite
material should be used as plasma facing tiles. According to length of coolant tubes, friction factor and loss coefficient relations.
Pressure drop in coolant tubes with 12mm o.d. and 10mm i.d. is 0.01MPa at a flow velocity of 1.36m/s when input power is up
to 6.5MW and operated in radiative divertor mode.
The divertor operating temperature is above 150C and peak heat flux strike to divertor target plates is shown in table 1
(normal operation) and table 2 (radative operation). Inlet temperature of coolant is 150C and temperature rise is 40C. Under
this situation temperature of plasma facing surface will not above 600C. Table 3 and table 4 shows the normal operation and
radiative operation at different peak heat flux the temperature of plasma facing surface, pressure drop and mass flow rate of
coolant respectively.
Table 1
Input Power
Peak Heat Flux(MW/m2)
(MW)
(outer target)
1.11
0.01
2.31
0.05
3.65
0.41
5.29
1.16
6.59
2.05
Table 2
Input Power
Peak Heat Flux(MW/m2)
(MW)
(outer target)
1.11
0.01
2.31
0.05
3.65
0.41
5.29
1.16
6.59
1.00
Table3 (normal operation)
Peak heat
Temperature
Pressure
Mass
flux
of plasma
drop
flow rate
(MW/m2)
facing
(MPa/m)
(kg/s)
surface (C)
0.01
246
3x10-7
1.5x10-3
0.05
277.2
5x10-6
7.5x10-3
0.41
384
2.3x10-3
6.2x10-2
1.16
557
1.4x10-2
0.17
768
-2
0.31
2.05
3.4x10
Table4 (radiative operation)
Peak heat
Temperature
Pressure
Mass
flux
of plasma
drop
flow rate
facing
(MPa/m)
(kg/s)
2
(MW/m )
surface (C)
0.01
246
3x10-7
1.5x10-3
0.05
277.2
5x10-6
7.5x10-3
0.41
384
2.3x10-3
6.2x10-2
1.16
557
1.4x10-2
0.17
1.00
520
1x10-2
0.15
It is known from the tables. When input power is lower then 5.29MW. Normal operating model is reliable. When input power
is up to 6.59MW radiative operation must be started.
5.8
Tokamak Assembly and Maintenance
For the configuration of the HT-7U device, the key to the settlement of the assembly
procedure is selection of the different subassembly ways of the TF magnets, the VV body and the
inner TRS.
A sophisticated optical metrology system (SOMS) will be adopted to setup a reference
coordinate system with the assembly base. During the engineering design of the HT-7U device, not
only the assembly datum plane of each main part should be defined, but also the specified points of
the measured target on each main part should be defined as the measured datum points based. SOMS
can also be used to inspect and measure the size and the contour of each important part of the HT-7U
device during the fabrication phase. With actual size and shape of each main part, the actual 3D
modules can be made by CAD software. So, an assembly procedure can be simulated by computer to
predict the assembly precision before the real assembly is carried out.
The assembly procedure of the HT-7U device is constituted of three sub-procedures that are
the supports assembly procedure, the torus subassembly procedure of TF with VV and the inner TS
and the final assembly procedure. At the beginning of the assembly, a virtual coordinate system will
be defined with respect to the assembly base. Several fiducial targets on the wall and the basement of
the test hall should be defined with reference of the coordinate system by SOMS.
During the phase of the supports assembly procedure, the base of CV will be installed on the
device support platform. Because a transition port for inlet and outlet of current bars and helium pipes
are trapped on the bottom of HT-7U, it should be installed to the central port of the CV base.
Following is the installation and position of the supports for VV, the cold mass and TRS. Fig. 2
shows the supports of the cold mass, TRS and VV on the base of CV. After alignment and adjustment
for all supports, the base part of TRS will be installed on the supports of TRS. And then four lower
support ports of the inner TRS will be installed on the TRS base. A circle support, which is spliced
into two halves by insulation structures, will be installed on the lower supports of the cold mass.
Since three lower PF coils are attached under the TF tore assembly, they should be stored on
the base of TRS temporary, seeing Fig. 3. Also, for the installation of the TF quadrant subassemblies
a temporary support should be installed in the centre of the CV base.
During the torus subassembly procedure of TF with VV and the inner TS, Each TF coil and
each 1/16 VV and the inner TS will be installed and connected to be a whole torus.
Three upper PF coils will be lowered from the top of the TF assembly down to position. By a
system of screw jacks or hydraulic jacks, three lower PF coils will be lifted and installed at position.
The subassembly of the central solenoid (CS) will be finished on the subassembly platform. The
subassembly of CS will be hanged into the hole formed by wedged straight legs of the TF cases using
a hanging tool and the bridge crane and adjusted to the position by screw supports connected between
the TF case and the subassembly of CS. Before next procedure, the treatment of joints of conductor
and cooling channel must be finished.
The installed superconducting magnets with the VV body and the inner TRS is showed in Fig.
5. All the PF coils including CS, the VV body and the inner TRS are adjusted and positioned with
respect to the virtual coordinate system by SOMS and verified with the position of the TF torus.
The middle ring and the top of the outer TRS will be installed to position and cover all the
cold mass. The entire upper vertical shield ports and the horizontal shield ports will be installed from
the outside of TRS. The cylindrical section and the cap of CV will be installed in the same way as the
outer TRS. To handling the middle ring and the top of the outer TRS and the cylindrical section and
the cap of CV, a special handling tool has to be designed, on which the large component can be
adjusted to a vertical orientation. Finally, twelve upper ports, twelve horizontal ports and residual
four lower ports will be inserted into their corresponding shield ports of TRS from the outside of CV.
After connecting all the ports with VV and CV, the assembly procedure of the main parts of the
HT-7U device is finished, seeing Fig. 6.
5.11 CICC Design
Several CICC’s with different configurations have been designed during the past three
years. Six third stage sub-cable samples with different configurations and surface-treatments had
been designed, fabricated and tested to compare their cryogenic properties. After analyses on the
test results in detail, a new design of the conductors has been issued and is planned to be used in
the PF&TF coils
[7]
. The conductor has the configuration of (2SC+2Cu)345 with a central
copper cable, as shown in Fig. 8. Four short samples in two sample-assemblies have been tested at
Centre de Recherches en Physique des Plasmas (CRPP) in Switzerland in September 2001.
Sample-assembly-1 consists of one TF conductor and one PF conductor. Two of them have the
same void fraction of ~35.5% and are coated with tin alloy on the surface of all superconducting
(SC) strands and copper strands. On the 3rd stage sub-cables of the PF conductor, a stainless steel
foil is wrapped for reducing AC losses. Sample-assembly-2 consists of two PF conductors in which
the surface of all SC and copper strands is coated with nickel and the void fraction is ~35.5%. One
of the PF conductors in sample-assembly-2 has stainless steel foil wrapped on the 3rd stage
sub-cables. The design parameters of the sample-assembly-1 are listed in Table IV. DC
performances, AC losses and stability are the essential data to be tested. Current sharing
temperatures and quench temperatures with different background fields and currents, critical
currents and quench currents with different background fields at 4.5K have been tested. AC losses
have also been tested using the calorimetric and magnetization methods. Although the critical
currents tested at different helium temperatures are 15-20% lower than the calculated values,
shown in Fig. 9, and the current sharing temperatures with different background fields are also
0.2K lower than calculated values, the tested results are still within the acceptable range. Because
the tested time-constants of the TF and PF conductors are much lower than the calculated values
and the AC losses at the operating current are lower than those without current, the AC losses
results should be analysed further. The stability tests indicate that the TF and PF conductors are
capable of withstanding verifying field with the designed Iop/Ic. But the PF conductor has shown
higher stability to verifying field with same Iop/Ic as the TF conductor, which means that the
wrapped stainless steel foil on the 3rd stage sub-cable was effective in reducing AC losses.