Test of the EURATOM LCT coil (NbTi conductor) with forced flow He II cooling

Fusion Engineering and Design 45 (1999) 361 – 375 www.elsevier.com/locate/fusengdes

Test of the EURATOM LCT coil (NbTi conductor) with forced flow He II cooling M. Darweschsad a, S. Fink a, G. Friesinger a, A. Gru¨nhagen b, R. Heller a, A. Hofmann a, W. Herz a, P. Komarek a, W. Maurer a, G. No¨ther a, K. Rietzschel c, G. Schleinkofer d, M. Su¨ßer a, A. Ulbricht a,*, F. Wu¨chner a, G. Zahn a a Forschungszentrum Karlsruhe, Institut fu¨r Technische Physik, Postfach 3640, D-76021 Karlsruhe, Germany Hauptabteilung fu¨r Ingenieurtechnik, Forschungszentrum Karlsruhe, Institut fu¨r Technische Physik, Postfach 3640, D-76021 Karlsruhe, Germany c Hauptabteilung Prozeßdaten6erarbeitung und Elektronik, Forschungszentrum Karlsruhe, Institut fu¨r Technische Physik, Postfach 3640, D-76021 Karlsruhe, Germany d Hauptabteilung Versuchstechnik, Forschungszentrum Karlsruhe, Institut fu¨r Technische Physik, Postfach 3640, D-76021 Karlsruhe, Germany b

Received 12 January 1999; received in revised form 10 March 1999; accepted 11 March 1999

Abstract The EURATOM LCT coil, a D shaped NbTi coil, was tested in the TOSKA facility at Forschungszentrum Karlsruhe with pressurized forced flow He II cooling. A suitable cryogenic system based on an existing 1.8 K He II 300 W liquifier was designed for cooling the coil by pressurized forced flow He II circulated by pumps across heat exchangers immersed in a He II bath. Three types of pumps (piston, centrifugal, thermomechanical) were used. The cryogenic system worked well under usual operation and fault conditions. No significant differences were found between He I and He II forced flow cooling. The D-shaped coil was reinforced by stainless steel belts keeping the D shape in single coil operation mode. No backlash was found between the coil and its reinforcement structure in agreement with predictions of the finite element analysis. The coil achieved  11 T as predicted from single strand measurements and previous tests in the frame of the Large Coil Programme. The results are an encouraging step in using this technology in large superconducting magnet systems for magnetic confinement in fusion research. © 1999 Elsevier Science S.A. All rights reserved. Keywords: TOSKA; NbTi coil; He II cooling

1. Introduction * Corresponding author. Tel.: + 49-7247-823911; fax: +497247-822849. E-mail address: [email protected] (A. Ulbricht)

The 1.8 K cooling technique is indispensable for achieving the highest magnetic fields with

0920-3796/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved. PII: S 0 9 2 0 - 3 7 9 6 ( 9 9 ) 0 0 0 1 9 - 8

362

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

superconducting materials presently available for technical applications. The cooling technique was applied initially for cooling superconducting high radio frequencies niobium and lead cavities by operating a He II bath under subatmospheric conditions [1]. For superconducting magnet cooling by He II, serious dielectric insulation properties of the bath cooled magnet winding and the sudden breakdown of the He subatmosphere had to be overcome in case of fault conditions. This was successfully solved by making use of the high thermal conductivity of He II to separate a subatmospheric He II bath from an atmospheric subcooled bath across a heat exchanger [2]. The atmospheric subcooled bath was embedded thermally insulated in a atmospheric 4.2 K bath for removal of external losses and reduction of the losses at the 1.8 K cooling loop. Both baths were connected by a plug-like valve acting as a thermal barrier. This component acts as a relieve valve in the case of emergency, and also solves the safety requirements. Such a system was developed and used for the superconducting NbTi toroidal field coils of the French tokamak Tore Supra. It has been operated successfully for more than 10 years [3]. This cooling scheme is now a standard technology for most of the bath cooled highest field magnets for research and NMR spectroscopy operated around the world [4]. The challenging field levels (13 T) and operation temperature margins ( 2 K) for existing tokamak designs (NET, Next European Torus; ITER, International Thermonuclear Experimental Table 1 Some data of the EURATOM LCT coil Geometrical dimensions (envelope) Bore Rated current Average winding current density Rated maximum field Inductance Number of turns Rated dump voltage Cooling paths Test pressure Design mass flow

3.6×4.6×0.7 m 2.3×3.3 m 11.4 kA 25.7 MA m−2 8.11 T 1.58 H 588 2.5 kV 28 (length 250 m each) 2.5 MPa 150 g s−1

Reactor) has favoured the development of Nb3Sn conductors in the last 10 years [5,6]. The fabrication technique of such coils needs much more effort caused by the reaction heat treatment and brittelness of Nb3Sn compared to NbTi coils as experienced in ‘The Large Coil Task (LCT)’ [7]. Simultaneously in the LCT project, the reliable and predictable operation of large forced flow cooled NbTi coils was demonstrated. Therefore, the operation of NbTi coils with forced flow He II was investigated as a fallback for the case that the Nb3Sn route would show too many problems especially for the magnetic field range where both solutions could be applied in principle. For the NET, an alternate solution for the toroidal field coils with NbTi with He II cooling was studied [8]. The existence of a 300 W He II refrigerator and the EURATOM LCT coil at Forschungszentrum Karlsruhe was a challenge to explore the forced flow He II cooling technique combined with operation limits of a large NbTi magnet. The cooling technique has been based on the same principles described above for bath cooled coils with the change that a pump has to circulate the supercritical He II through the internally cooled superconducting cable [9,10]. An investigation of the operation of the LCT coil at a current of 20 kA resulted in the conclusion that only a reinforcement of the mechanical structure was necessary [11]. Therefore, a task was started within the European Fusion Technology Programme containing the development of components needed for He II forced flow cooling circuit and the operation of a large forced flow cooled fusion magnet (EURATOM LCT coil) in the He II temperature range.

2. The LCT EURATOM coil

2.1. Pre6ious operation conditions The EURATOM LCT coil was tested in LCT test facility as a single coil and in toroidal configuration (Table 1). It achieved electromagnetic boundaries derived from

the the the the

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

363

Fig. 1. An exploded view of the reinforcement structure of the EURATOM LCT coil.

short sample measurements of the strand material. The coil was operated by subcooled supercritical forced flow helium at a temperature of 3.8 K. The maximum field levels achieved at the winding were 9.0 T as a single coil and 9.2 T in the toroidal configuration as predicted from short sample strand measurements [7]. The EURATOM LCT coil was operated as a single coil without any additional support structure up to a current of 10 kA [12 – 16]. The weakest points in this load case were relatively short bolts in the straight section of the coil case flange. Using the structure of the central bucking post of the torus test facility of the ‘Large Coil Task’ as a reinforcement, the coil achieved a maximum current of 16 kA within its mechanical load limits [7]. In this case, maximum stresses in the range of 700 MPa occurred at the clamp insert hole at the top and bottom of the straight section of the D [17]. From this point of view, such a kind of reinforcement had no further potential for operating the coil above these currents. For the envisaged goal, operating the EURATOM LCT coil at a He II temperature of 1.8 K, a maximum possible current of up to 20 kA was expected. There-

fore, a new design of a reinforcement structure was needed for that condition.

2.2. Reinforcement structure An exploded view of the reinforcement structure is presented in Fig. 1. The basic idea of this design is that the two beams at the front and at the rear side of the LCT coil are surrounded by 12 steel belts along the height of the coil case. This reinforcement structure keeps the bending especially of the straight leg of the D in acceptable limits by tensioning the belts. The fishbone shaped rear beam, consisting of horizontal plates and the vertical plate rests on the flange of the LCT coil case. The vertical plate was fabricated by copy milling from an epoxy resin cast taken from the flange. Originally the LCT coil was not designed to have suitable space for force transmission in this area. Every horizontal plate of the backbone plate had to be individually adapted to the LCT coil case shape. The front beam was milled from a forged steel part. At the front side, the LCT coil has a high quality mating surface to the front beam. For all structural parts, the nitrogen strengthened steel 316 LN was used.

364

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

Each belt consists of two halves which are divided in the middle for easy assembling around the LCT coil. Each half belt of the front part is screwed together on the front and back side by bolts with counteracting threads. All belts were heated up to 50°C along 3 m on each side during the tightening of the bolts in order to generate a pretension of the belts when their temperature returned to room temperature. The measured horizontal deformation of the LCT coil was in good agreement with the FEM calculations performed for these pre-stressed belts [18].

2.3. Finite element model A detailed finite element model was used for the calculations [19]. The model was first constructed in the NASTRAN code and was later remodeled in ABAQUS (Size of the model: No. nodes, 5392; No. elements, 5592; Degrees of freedom, 17 832) [20]. The winding pack was modeled with orthotropic material properties and an azimuthal dependence of radial Young’s modulus. The latter became necessary to fit the results of the calculations to the first measurements [13,21]. The force transmission between the winding pack and the stainless steel coil case was modeled by gap elements.

2.4. Mechanical instrumentation The instrumentation of the EURATOM LCT coil was used as described in [7]. During testing of the coil in the TOSKA facility, the global deformation of the coil was measured across the apices in vertical and horizontal directions. An additional 30 strain gauges were installed onto the tension belts.

The LCT coil will be used later on in the W 7-X prototype coil test and the ITER TF model coil test will be used as a background field coil.

3.1. Cryogenic supply system To respond to future requirements, the cooling system of the TOSKA facility has to be extended according to the flow scheme in Fig. 3. The 2 kW refrigerator was connected to the He I control cryostat B 250 and to the vacuum vessel B300. The 500 W refrigerator was connected to the newly installed He II control cryostat B 1000 for the subatmospheric operation at 1.8 K. During the LCT coil test, it must be possible to supply the winding alternatively from the He I control cryostat with supercritical He or for 1.8 K operation from the He II cryostat with pressurized helium II. It was therefore foreseen to valve on and off the winding to the forced flow pressurized He II circuit of the control cryostat B 1000 or to the forced flow He I supercritical circuit of the control cryostat B 250. For the pressurized He II operation thermal barrier for the venting lines had to be developed and tested separately (Fig. 3). A cold storage medium pressure vessel (20 bar) was installed to collect the expelled helium gas to avoid a response of the safety valves and He losses. Three different pumps (3 cylinder piston, centrifugal and thermomechanical) were installed in the superfluid control cryostat. The whole system was controlled by programmable logic controllers (PLC). The circuits were operated by a control and visualization system (VXL) on a VAX station [23]. Both forced flow circuits were tested separately and in combination. It was demonstrated in the test of the LCT coil at 1.8 K that both the He I and He II cryogenic supply systems are now fully available.

3. The TOSKA facility

3.2. Electric supply system

The LCT coil was installed in the TOSKA facility after 9 years idle time (Fig. 2). The facility components cryogenic supply system, dump circuit of the LCT and data acquisition were installed, taking into operation and commissioned by the test of the LCT coil at 1.8 K [22].

The 50 kA, 30 V DC power supply was used as a current source for the LCT coil which was connected across a separation switch (ST) in the positive and negative busbars (Fig. 4). The dump resistor (R) was switched parallel to the coil (L). The short circuit path of the coil

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

consist of two arc chute breakers (S) in parallel with a pyro breaker (F) in series for having simple redundancy if one of the arc chute breakers does not close or open. The switching sequence is controlled by a PLC with a backup relay circuit for the safety discharge. In case of a dump trigger, the arc chute switches are closed and the power supply is switched in the inverter mode. The coil is now short circuited by the branch (S,F). Subse-

365

quently, the separation switches ST open and isolate the power supply from the dump circuit. After that, the arc chute breakers open and commutate the current in the dump resistor. The time constant was 12.5 s and the dump voltage 2.5 kV at 20 kA. The thyristors of the power supply were protected against voltage peaks by an overvoltage protection device (OV). The short circuit branch in parallel to the main short circuit with a current

Fig. 2. A view into the TOSKA vacuum vessel with the installed LCT coil. On the right side of the LCT coil is the space for later test coils (W 7-X prototype coil and ITER toroidal field model coil, TFMC).

366

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

Fig. 3. Flow diagram of the cooling system.

limiting resistor (RN) was installed for arc free opening of the separation switches. The delay time between trigger and breaker opening was  675 ms with a jitter of 20 ms, mainly determined by the opening time of the separation switches. The maximum switching power handled for extracting 300 MJ was 48 MW.

The channel number can be adapted to the requirements of the experiment. The LCT coil instrumentation was used as existing. The additional mechanical instrumentation is described in Section 2.4.

4. The 1.8 K test of the EURATOM LCT coil and results [22]

3.3. Instrumentation and data acquisition The operation was performed in two steps. The cryogenic supply system was equipped with sensors necessary for control and protection. All sensors were connected to the new data acquisition system (Fig. 5). The necessary data for controlling the facility were adapted in their engineering units in the data base ORACLE where they were available for further analysis procedures. For the operation, the data were collected by scanners with B 5 s repetition rate. Fast changes during quench or fast discharge were acquired by transient recorders with 1 MHz sampling rate. In the present state, the scanners have 610 channels and the transient recorders  32.

Fig. 4. Electrical power supply with dump circuit (LCT coil: R, Dump resistor; S, Arc chute breaker; F, Pyro breakers; ST, Separation switch; RN, Current limiting resistor; SN, Closing switch; OV, Overvoltage protection).

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

367

Fig. 5. The block diagram of the data acquisition, control and visualization system for the TOSKA facility. Table 2 Cooling conditions Operation mode

16 kA

19 kA

Standby

19 kA

Standby

Winding Temperature (K) Mass flow rate (g s−1) Heat load (W)

3.5 50 10

1.8 80 35

1.8 20 25

1.8 80 48

1.8 30 37

Case & structure Temperature (K) Mass flow rate (g s−1) Heat load (W)

3.5 50 100

3.5 70 90

3.5 40 90

4.5 26a 65

4.5 25a 75

Current lead Temperature (K) Mass flow rate (g s−1) Cold gas return (g s−1)

3.5 2×1.6 0

3.5 2×2.0 0

3.5 2×0.3 0

4.5 2×1.6 1.6/3.6

4.5 2×0.4 2×0.5

a The reason for the low mass flow rate of the case during 4.5 K operation was a disturbed He pump and the case was cooled directly by mass flow of the 2 kW refrigerator.

First, the winding and case were cooled at 3.5 K by the forced flow circuit of the control cryostat B 250. Later, for pressurized He II operation, the

winding was connected to the cooling circuit of the B 1000. The operation parameters are given in Table 2.

368

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

4.1. Standard operation

4.2. Pressurized He II operation

The LCT coil with its reinforcement (total mass now 60 t) was cooled down again during 290 h to its operation temperature with  1 K h − 1 and a temperature difference of B 40 K between inlet and all other temperature sensors. The leak rate at 4 K was less than 5× 10 − 6 mbar s − 1. The coil was ramped up with 20 A s − 1 in steps accompanied by dump tests up to its nominal current of 11.4 kA at an operation temperature of 4.6 K. The same procedure was repeated at 3.6 K. After a steady state operation at 16 kA, the ramp up was started with 5 A s − 1. The coil showed a spontaneous quench at 16.5 kA in full agreement with the current sharing data measured earlier (Fig. 6) [24]. The dump was delayed by 2.5 s caused by the impact of magnetic stray fields on relays of the quench detector. This was avoided for the following runs by a m-metal screening. The estimated hot spot temperature was 100 K. This was derived from the temperature of the expelled helium with  80 K. After a quench of three pancakes, the cryogenic system handled the expelled gaseous helium without He losses and the recooling took place in less than 2 h.

To proceed to a 1.8 K forced flow cooling operation, root blowers were started for cooling the control cryostat (B 1000) to 1.8 K. This is obtained with a suction pressure of 16 mbar. When the dewar was filled with liquid helium, the winding was connected to the 1.8 K loop and cooled down to its superfluid operation temperature while the coil case, reinforcement structure, and current leads remained on the He I cooling loop at a temperature of 3.5 or 4.5 K (Fig. 7). The heat load of the winding and also the outlet temperature of the first and last pancake indicate that most heat is coming from the current feedthroughs of the winding which were not designed for a temperature difference between winding and current lead. This also explains the higher heat input to the winding during the operation of the case and current leads at 4.5 K instead of 3.5 K. The outlet temperature and the heat load to a winding and case clearly show the additional eddy current losses during ramp up and down. A typical operation with pressurized He II is shown in Fig. 8. The following procedure consisted of ramping up to different current levels (11.4, 16.3, 17.3, 19.0 kA), performing an inverter mode discharge, and, in a second step, a manual triggered

Fig. 6. Current sharing diagram for The EURATOM LCT coil derived from measured current sharing data and extrapolation to the HE II range.

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

369

Fig. 7. Switching from He I to He II operation.

Fig. 8. Operation up to 19 kA and standby.

dump. The ramp rate was 20 A s − 1 up to 11.4 kA and above this level, 10 A s − 1. A ramp rate of 5 A s − 1 was used for ramping the coil into the current sharing region.

4.3. Electromagnetic properties The coil was ramped up in its current sharing region, clearly visible on the resistive voltage. The

370

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

quench started, as expected, in the center pancake. The inlet temperature was 1.89 K, the current 19.45 kA and the maximum field 10.8 T. This quenching current was reproducible within the measuring accuracy. A detailed analysis were done using the single strand measurements of each cable and applying a scaling law for NbTi. For each cable, a band of the critical currents was given. The stable operation was possible up to an averaged strand current at the lower boundary of the band. The averaged strand quench current was positioned within the band [24]. The operation limit tests in the LCT test facility, ORNL, Oak Ridge, USA were performed at the outermost double pancake DP 7 by induced quenches or current sharing measurements. The spontaneous quenches at 3.5 and 1.89 K took place in the center pancake DP 4. Both results are in fair agreement (Fig. 6). This confirms predictable operation limits for the NbTi cable technology. With regard to quench propagation, no differences were found between He II and He I [24]. The coil and the dump circuit withstood a dump voltage of 2.45 kV with a peak power of 48 MW.

4.3.1. The effect of large 6olume fringing fields on facility components The impact of large volume magnetic fringing fields on components in the experimental area was investigated. In the fringing fields of the LCT coil, the dump circuit used for the POLO project was operated together with the 30 kA power supply and a copper coil [25]. The current level of the LCT coil was increased in steps and the function

of the electrical system was tested. First faults were observed in the control circuits of the POLO dump circuit at current levels of 8–13 kA with a big hysteresis corresponding of field levels in the cabinets of  3 mT. A Reed relay was identified causing the fault. Changing of the orientation of the relays with respect to the magnetic fringing field direction solved the problem. At 16 kA current level, the current transformer of the 30 kA power started to saturate which led to emergency switch off of the power supply (fringing field level  5.5 mT in this area). An iron plate shield surrounding the transformer resulted in a shielding of  50% which was sufficient for fault free operation up to 19 kA LCT coil current level corresponding to a fringing field level of 6.5 mT [26]. Iron cored current transformers showed no impact from fringing fields up to levels of B20 mT. The fringing fields for future test configurations can be scaled according to Table 3 in first an approximation. A programmable logic controller (PLC) (product: Siemens Simatic S5-135) was prepared in hard- and software for being tested in the surrounding of magnetic fringing fields. The PLC was tested in a magnetic field up to 22 mT (vertical 20 mT, horizontal 9 mT) and remained operable in all its components [27].

4.4. Thermohydraulic properties [28] The handling of the 4.5/3.5 K cooling system together with the 2 kW refrigerator was free of problems during computer controlled cool down,

Table 3 The number of Ampere turns of different coil configurations which have been tested and will be tested in the TOSKA facility (Axis position: Vacuum vessel axis and coil axis) Coil, Coil, Configuration

Conductor current

Ampere turns

Axis position

LCT POLO LCT (1.8 K operation) LCT+W 7-X LCT+ITER TFMC ITER TFMC

10.0 kA 22.5 kA 19.5 kA LCT: 14 kA+W 7-X: 18.7 kA LCT: 16 kA+TFMC: 80 kA 80.0 kA

5.88 MA 1.25 MA 11.5 MA 10.5 MA 17.25 MA 7.84 MA

Þ

Þ Þ Þ Þ

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

371

Table 4 Thermohydraulic properties after a safety discharge (SD) without or with quench (Q) (peak values)a Operation mode

3.5 K, 16 kA (SD)

1.8 K, 19 kA (SD)

1.8 K, 19.6 kA (SDQ)

Winding Inlet temperature (K) Outlet temperature (K) Pressure (bar) Heat load (MJ)

14.7 8.8 5.3 0.38

11.7 11.0 8.8 1.6

26.5 17.4 26.6 2.0

Case Inlet temperature (K) Outlet temperature (K) Pressure (bar) Heat load (MJ)

12.5 18.4 5.6 1.1

17.5 21.0 18.1 1.3

24.3 21.7 13.4 1.65

a

SD, without quench; SDQ, with quench.

Fig. 9. Pressure drop across the LCT coil as function of the mass flow for He II operation.

the coil cooling with the secondary pump loop, and also during recooling after quench or dump. Stable He II cooling of the winding with the piston and fountain pumps was demonstrated and the disturbance of the mass flow after a dump up to 14 kA was negligible. At a higher current, the pumps were switched off after dump or quench. For a continuous and equal mass flow through the 28 cooling channels, a rate of 80 g s − 1 was used at 1.8 K because the observed mass flow

instability caused a low mass flow rate. This was caused by the partial opening of the gravity loaded check valves which were parallel to orifices by magnetic forces on weak ferromagnetic parts inside. No significant differences in pressure drop and heat losses between He I and He II cooling conditions after a dump or quench were observed (Tables 2–4). The pressure drop measurement as a function of mass flow is in agreement with the usual calculation formulas for He I (Fig. 9). The

372

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

shear viscosity had no remarkable contribution (B 20%) within the measuring accuracy [29]. The critical mass flux (g sm − 2) was exceeded for the given channel cross-section (m2) conductor [30]. Dumps up to 300 MJ of stored energy were easily handled by the cryogenic system without any helium gas release to the atmosphere. Each of the three pumps were able to recool the LCT coil and its reinforcement structure in 2 h. The installation of a cold gas return path at the cold end of the current leads of the LCT coil improved the adjustment of the operation parameters remarkably. At lower currents, the minus current lead showed an unstable operation behaviour. No explanation was found for this phenomena from the evaluation of the operation parameters. A piston-, a centrifugal- and a thermomechanical pump, were tested at 1.8 K operation and all achieved their expected parameters. No impact of magnetic fringing fields was found on the operation up to field levels of B 7 mT (19 kA LCT coil current) for the piston and centrifugal pump. For application, the pump has to be integrated in an optimized cooling circuit. The most suitable pump depends upon the properties of this circuit.

4.5. Mechanical properties and FEM calculations [31,32] The prestressed reinforcement structure was very effective and showed no backlash and hysteresis effects as observed in prior tests [13]. The equivalent stresses calculated from the strain gauge rosettes were in fair agreement within 25% with the calculations (Fig. 10). The compression of the winding pack in the straight section of the D shape is in full agreement with measurements performed in all single coil tests [31]. The gap measured between winding and case at the inner side of the straight part of the reinforced coil corresponds to measurements at the ORNL test facility where the LCT-coil was fixed to a bucking post (Fig. 11). This confirms the quality of the design, the proper installation and the force transmission from the coil to the reinforcement structure. The global deformation of the coil measured in horizontal and vertical directions across the apices also met the FEM results within the measuring accuracy (horizontal, 2.09 0.1 mm; vertical, 0.359 0.05 mm at a maximum current of 19.4 kA). The belt stresses varied from  400 MPa in the center down to  250 MPa in the upper and

Fig. 10. Comparison of the measured von Mises stresses on the coil case surface by strain gauge rosettes with the results of a FEM calculation. The position of the rosettes is represented in the graph.

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

373

Fig. 11. The gap between winding and coil case as well as the horizontal deformation of the coil bore for the different test modes of the EURATOM of the LCT coil.

lower belts. The deviation to the FEM calculations was  33% for the center belts. There are also differences between the front and rear side on the same belt in the range of 6%. Considering the complex construction of the reinforcement, the agreement is adequate. All described measurements were given at the maximum current. They were linear in current square, leading to an excellent force transmission without any backlash. Therefore, from the agreement of the global measurements with FEM calculations, it can be derived that the calculated maximum stresses really occurred in the coil (Winding: circumferential: s8 = 298 MPa; radial: sr = − 47 MPa; axial: sz = − 58 MPa; shear: sr8 = − 42 MPa; Case: von Mises: sequ =281 MPa). It can be concluded that for a D shaped toroidal field coil, the construction of a rigid reinforcement structure is possible to test the coil at its outermost operation limits.

5. Conclusions After 9 years idle time, it was confirmed that the EU-LCT coil with all its components kept its

full performance. Therefore, the coil is suitable to be used as a background field coil for the ITER TFMC test. For the first time, it was demonstrated that a coil of this size can be cooled with forced flow pressurized He II. Three types of pumps (piston, centrifugal and thermomechanical) were tested for generating forced flow pressurized He II. All pumps were suitable for the operation with He II. All operation conditions were successfully mastered. No substantial differences in thermodynamic and quench propagation were found between He I and He II. The cryogenic system was used as it existed and therefore not optimized for an economic operation heat load for the refrigerator. The current feedthroughs of the LCT coil were not designed to have a temperature gradient from 4.2 to 1.8 K. Another adapted design can reduce the heat load to the 1.8 K circuit. The clarification of the mass flow instabilities by a malfunction of check valves will lead to a reduction of the total mass flow rate of  50 g s − 1, which correspondents to lower pressure drop combined with a lower pumping power. This fact also reduces the heat load to the refrigerator.

374

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375

In summary, there is substantial potential for reducing the heat load of the refrigerator for an economic operation of a larger forced flow cooled magnet system with pressurized He II. The test demonstrates the engineering standards of the design principles of the EURATOM LCT coil with fully predictable operation limits. The construction of a suitable mechanical structure for the testing of a D shaped coil as a single coil was demonstrated [32]. Acknowledgements This work was performed within the Project Nuclear Fusion of the Forschungszentrum Karlsruhe (FZK) and was part of the European Technology Programme. The authors acknowledge the work of all people outside and inside FZK, as well as the European industry contribution to the success of the experiment. References [1] A. Citron, J. Halbritter, M. Kuntze, Entwicklungen auf dem Gebiet der Hochfrequenzsupraleitung, Kernforschungszentrum Karlsruhe, KFK-Ext.0376-05, Dezember 1976. [2] G. Bon Mardion, G. Claudet, J. C. Vallier, Superfluid helium bath for superconducting magnets, Proceedings of the Sixth ICEC, Grenoble, France, 11–14 May 1976, pp. 159 – 162 [3] J. Pamela, Ten years of operation and development of Tore Supra Soft, Proceedings of the 20th SOFT, Marseille, France, 7–11 September, 1998. [4] P. Komarek, Hochstromanwendung der Supraleitung, B.G. Teubner, Stuttgart, 1995. [5] The NET Team, Predesign report NET, Next European Torus, Fus. Eng. Des. 21 (1993) [6] N. Mitchell, ITER magnets design and R & D, Proceedings of the 20th SOFT, Marseille, France, 7–11 September, 1998 [7] D. S. Beard, W. Klose, S. Shimamoto, G. Vecsey, The IEA Large Coil Task, Fus. Eng. Des. 7 (1988). [8] D. Ciazynski, J.L. Duchateau, B. Jaeger, A. Trossian, B. Turck, Study of the TF coils of NET using a NbTi conductor cooled by forced flow of He II, Association EURATOM CEA, Cadarache, P/EM/90-07, April 1990 [9] A. Hofmann, A. Kahlil, Consideration on magnet design based on forced flow of He II in internally cooled cables, Proceedings of the 14th SOFT, Avignon, France, 8 – 12 September 1996, pp. 1811–1817

[10] G. Zahn, A. Hofmann, H. Bayer, H. Berndt, R. Doll, W. Herz, M. Su¨ßer, E. Turnwald, B. Vogelei, W. Wiedemann, Test of three different pumps for circulating He I and He II, Cryogenics 32 (1992) 100 – 103. [11] A. Hofmann, P. Komarek, W. Maurer, G. Ries, B. Rzezonka, H. Salzburger, Ch. Schnapper, A. Ulbricht, G. Zahn, Further use of the EURATOM LCT coil, Proceedings of the 15th SOFT, Utrecht, Netherlande, 19–23 September, 1988, pp. 1556 – 1802 [12] W. Herz, H. Katheder, H. Krauth, G. No¨ther, U. Padligur, K. Rietzschel, F. Spath, L. Siewerdt, M. Su¨ßer, A. Ulbricht, F. Wu¨chner, G. Zahn, Results of the European LCT coil in the TOSKA facility, Proceedings of the ASC San Diego, CA, USA, September 9 – 13, 1984, IEEE Trans. Magn., MAG-12 (1985) 249 – 252 [13] W. Herz, H. Katheder, H. Krauth G. No¨ther, U. Padligur, K. Rietzschel, F. Spath, L. Siewerdt, M. Su¨ßer, A. Ulbricht, F. Wu¨chner, G. Zahn, Testing of the European LCT Coil in the TOSKA Facility, Proceedings of the 13th SOFT, Varese, Italy, 24 – 28 September, 1984, pp 1487 – 92. [14] A. Maurer, A. Ulbricht, F. Wu¨chner, Effect of the azimuthal dependence of the radial Young’s modulus on the mechanical behaviour of the European LCT coil, Proceedings of the MT-9 Zu¨rich, Switzerland, 9 – 13 September, 1985, pp. 428 – 431. [15] H. Tsuji, S. Shimamoto, A. Ulbricht, et al., Experimental results of domestic testing of the pool-boiling cooled Japanese and the forced-flow cooled EURATOM LCT coil, Cryogenics 25 (1985) 537 – 551. [16] W. Herz, H. Katheder, H. Krauth G. No¨ther, U. Padligur, K. Rietzschel, F. Spath, L. Siewerdt, M. Su¨ßer, A. Ulbricht, F. Wu¨chner, G. Zahn, Design and process engineering of the Karlsruhe toroidal field test facility TOSKA, Proceedings of the ICEC-10, Helsinki, Finland, Butterworth, Guildford, 31 July – 3 August, 1984, pp. 83 – 86 [17] Siemens, LCT – EURATOM, Endgu¨ltiger Entwurf, Abschlußbericht o3/84 Band B: Berechnungen, S. 34 ff, 1984 [18] B. Rzezonka, Vorspannung in den Ba¨ndern der Versta¨rungsstruktur der LCT – Spule zur Reduzierung der Schubspannnung in der Wicklung, Technischer Bericht Ident-Nr. 32.10634.7, INTERATOM Bergisch-Gladbach 1, September 1991. [19] A. Maurer, Structural Analysis of the EURATOM Coil for the Large Coil Task, Proceeding of the MT-7, IEEE Trans. Magn., MAG-17 (5) (1981) 2043 – 2096. [20] B. Rzezonka, U. Hoeck, LCT/EURTOM-Spule Konvertierung des NASTRAN – FE – Modells in ein ABAQUS– FE – Modell, Technischer Bericht Ident-Nr. 32.10289.5, INTERATOM Bergisch-Gladbach, Mai 1990. [21] A. Gru¨nhagen, Festigkeits-Parameter-Studien zu TOSKA – Upgrade Prima¨rbericht 31.03.02 P 01A, HIT, Kernforschungszentrum, Karlsruhe, Oktober 1992, unpublished note. [22] M. Darweschsad, G. Dittrich, A. Gru¨nhagen, S. Fink, G. Friesinger, A. Go¨tz, R. Heller, W. Herz, A. Hofmann, P.

M. Darweschsad et al. / Fusion Engineering and Design 45 (1999) 361–375 Komarek, O. Langhans, W. Lehmann, W. Maurer, I. Meyer, G. No¨ther, G. Perinic, W. Ratajczak, H.P. Schittenhelm, G. Schleinkofer, K. Schweikert,, Ch. Sihler, E. Specht, H.J. Spiegel, M. Su¨ßer, A. Ulbricht, A. Vo¨lker, F. Wu¨chner, G. Zahn, V. Zwecker, Operation of the upgraded TOSKA facility and test results of the EU-LCT coil cooled with forced flow Helium II at supercritical pressure, Proceedings of the 19th SOFT, Lisbon, Portugal, 16 – 20 September, 1996 [23] A. Augenstein, H. Barthel, P. Gruber, R. Kaufmann, P. Klingenstein, U. Padligur, K. Rietzschel, G. Wuerz, H.P. Zinnecker, Data Acquisition, Control and Visualisation System for the Upgraded TOSKA Facility at FZK, Proceedings of the 19th SOFT, Lisbon, Portugal, 16 – 20 September, 1996 [24] R. Heller, M. Darweschsad, S. Fink, G. Friesinger, W. Herz, W. Maurer, G. No¨ther, G. Schleinkofer, M. Su¨ßer, A. Ulbricht, F. Wu¨chner, G. Zahn, Operation limits of the EURATOM LCT coil at He II temperature, Cryogenics 5 (1998) 519–523 [25] M. Darweschsad, G. Friesinger, R. Heller, M. Irmisch, H. Kathol, P. Komarek, W. Maurer, G. No¨ther, G. Schleinkofer, C. Schmidt, Ch. Sihler, M. Su¨ßer, A. Ulbricht, F. Wu¨chner, Development and test of the poloidal prototype coil POLO at the Forschungszentrum Karlsruhe, Fusion Eng. Des. 36 (1997) 227–250

.

375

[26] J. Zimmermann, A. Ulbricht, F. Wu¨chner, Streufelder beim Betrieb der LCT-Spule, Interner Bericht, Forschungszentrum Karlsruhe, Juli 1997 [27] J. Zimmermann, M. Su¨ßer, F. Wu¨chner, Test einer Siemens Simatic S5-135 im statischen Magnetfeld, Interner Bericht, Forschungszentrum Karlsruhe, August 1997 [28] G. Zahn, Experiences with 1.8 K forced flow HeII cooling of large superconducting magnets, Proceedings of the ICEC, UK, 27 13 – 17 July, 1998 [29] A. Hofmann, Some aspects of cryogenic pumps, in: B. Seeber (Ed.), Handbook of Applied Superconductivity, Institute of Physics Publishing, Bristol, Philadelphia, 1998 [30] A. Hofmann, Design of a fountain effect pump for operating the EURATOM LCT coil with forced flow He II, in: R.W. Fast (Ed.), Advances in Cryogenics Engineering, vol. 37, Part A, Plenum Press, New York, 1992 [31] A. Gru¨nhagen, B. Kneifel, FE-Analysen und Vergleich mit den Testergebnissen der versta¨rkten LCT-Spule, FZK Karlsruhe, Wiss. Berichte FZKA 5894, April 1997 [32] M. Darweschsad, G. Friesinger, A. Gru¨nhagen, R. Heller, W. Herz, B. Kneifel, P. Komarek, W. Maurer, G. No¨ther, S. Raff, M. Su¨ßer, A. Ulvbricht, F. Wu¨chner, G. Zahn, The test of the EURATOM LCT coil at the outermost limits at 1.8 K as example for single coil testing of D shaped toroidal field coils, Proceedings of the MT-15, Beijing, China, 20 – 24 October, 1997, pp. 373 – 376