DOE/PC-70512-11
Develop and Test
an Internally Cooled, Cabled Superconductor (ICCS) for Large Scale MHD Magnets
Semiannual Progress Report
Period from July 1, 1987 to December 31, 1987 J. R. Hale, A. M. Dawson, P.G. Marston
Plasma Fusion Center
Massachusetts Institute of Technology Cambridge, Massachusetts 02139, USA
This work was supported by the U.S. Department of Energy, Pittsburgh Energy Tech-nology Center, Pittsburgh, PA, 15236 under Contract No. DE-AC22-84PC70512. Repro-duction, translation, publication, use and disposal, in whole or part, by or for the United States Government is permitted.
NOTICE
This report was prepared as an account of work by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, appara-tus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trademark, manufacturer, or otherwise, does not necessarily constitute or imply its en-dorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
TABLE OF CONTENTS
Section Title Page No.
1.0 Introduction 1
2.0 Review of Technical Progress Prior to July 1, 1987 2
3.0 Summary of Current Work 3
4.0 Technical Progress, July 1 through December 31, 1987 3 4.1 Completion of Test Rig and Experimental Package 3
4.2 Instrumentation and Data Acquisition 6
4.2.1 Procedure 8 4.3 Results 9 4.3.1 Recalculations 9 4.3.2 Experimental Data 12 4.4 Conclusions 20 5.0 References 21
iii
List of Figures
Number
1
2
3
4
5
6
7
8
9
10
11
Title
Photograph of Test Coil Assembly
Photograph of Test Coil Attached to Mount
Schematic of Data Acquisition System
Worst Case I Energy Margin vs I/Ic for Samples A and B
Worst Case II Energy Margin vs I/Ic for Samples A and B
Experimental Data at Three Field Levels for Sample A
Experimental Data at Three Field Levels for Sample B
Visually Smoothed Data
-
Heater Energy vs I/I,
- for Samples A & B
Shot #81 Showing a Propagating Normal Zone
Shot #80 Showing the Initiation and Collapse
of a Normal Zone
Shot #110 Showing a Normal Zone that Collapses
after an Extended Time
Page No.
4
5
7
10
11
14
15
16
17
18
19
1.0 Introduction
A three-year program to develop and test an internally-cooled cabled superconductor
(ICCS) for large-scale MHD magnets is being conducted by MIT for the Pittsburgh Energy
Technology Center (PETC) under Contract DE-AC22-84PC70512. The program consists
of the following four tasks:
I. Design Requirements Definition
II. Analysis
III. Experiment
IV. Full-Scale Test
This report, covering the period from July 1 through December 31, 1987, discusses the
completion of Task III, the experimental phase of the program in which two prototypical
subscale superconductors were chosen, procured, wound into a test coil, tested and the
test data analysed and evaluated. Earlier work in Task III was described in report number
DOE/PC-70512-10 covering the period from January 1 through June 30, 1987. This report
focuses on the final test runs and the conclusions drawn from these experiments.
2.0 Review of Technical Progress Prior to June 30, 1987
Technical progress from the start of the program through June 1987 is reviewed briefly as a framework for the report of progress contained in sections 3.0 and 4.0.
To initiate the preconceptual magnet design, it was assumed that a typical retrofit MHD magnet would:
1) accommodate a supersonic MHD channel of about 35 MW, output, requiring a peak on-axis field of 4.5 tesla and,
2) operate at a design current in the neighborhood of 25 kA.
To facilitate the design process, it was assumed that the dimensions and construction of the conductor for the full-scale retrofit magnet would be the same as those used in the large D-shaped magnet built by Westinghouse Corporation for the Large Coil Program tokamak TF coil study.('-8 ) In this way advantage could be taken of the manufacturing technology that had been developed for that project. The MHD program chose to use NbTi rather than NB3Sn since the maximum required field for an MHD system is more
appro-priate to NbTi, a better understood and less expensive conductor than Nb3Sn, although it
had never been considered for use in an ICCS configuration in an MHD application where the requirements are significantly different than those for fusion.
The retrofit magnet's size and field strength were selected based on information ob-tained by surveying the MHD community('. A relatively high design current was chosen with the goal of minimizing overall system cost('). The selection of overall ICCS dimen-sions and construction methods was aimed at minimizing conductor development time and cost by using a conductor size for which tooling and production experience already exist.
An initial preconceptual design for a retrofit-scale magnet was generated that incor-porates a 600 rectangular-saddle-coil ICCS winding without substructure that will operate at a design current of 24 kA in a stainless-steel force-containment structure and cryostat.
A detailed computer analysis of the winding showed that maximum fields were about 7.2 T rather than the 6 T estimated. The winding was therefore modified to reduce the maximum field and to ensure stable operation. The resulting design had coils with increased thickness, increased end-turn bend radius and lower current density, resulting in a magnet preconceptual design that compared favorably with earlier designs in reliability, manufacturability, and cost effectiveness.
Once the preliminary design was completed it was necessary to revise and improve the preconceptual design and to provide greater detail. Electromagnetic analyses were reviewed and checked using alternate approaches. Pressure drop and frictional heating in
the conductor coolant circuit were reviewed as were a number of critical structural details
which proved to be in need of further analysis. A sound basis was established upon which
to base conductor design requirements and the experimental test program design.
3'
4The conductor design requirements were established(') and two candidate subscale
conductors were identified and their parameters defined. A test plan
4was developed,
subscale conductor for the long-sample tests was ordered and the test rig was designed.
Preliminary conductor bending tests were performed, a test heater was acquired and a
number of preliminary tests were performed on it. These included bending tests, soldering
tests and performance measurements.
5Analytical work on the subscale conductors continued in parallel, with a particular
emphasis on stability predictions.
3.0 Summary of Current Work
During this period, the test rig design, construction and assembly were completed.
Test parameters were solidified and instrumentation specifications and signal sources
de-cided. Instrumentation leads were attached to the test rig assembly which was then
mounted on an experimental probe assembly to allow it to be inserted in a
background-field-providing magnet system and to be helium cooled.
Tests were run, data were collected using a LeCroy 32-channel digitizer, and stored
on a MicroVax; the data were subsequently analysed. The data indicated that further
consideration of the two-in-one copper/superconductor cable configuration is warranted.
This is encouraging as this conductor will be considerably less expensive to manufacture
than the all-multifilamentary type ICCS used as the alternative subscale conductor.
4.0 Technical Progress, July 1 to December 31, 1987
4.1 Completion of Test Rig and Experimental Package
Figure 1 is a photograph of the experimental test coil after winding and
instrumen-tation were completed. The current and helium supply leads can be seen projecting from
the top of the coil and the fine wires coiled on the upper surface are the lead wires from
the instrumentation. Figure 2 shows the upper end of the coil attached to the mounting
fixture that allows the coil to be hung in the liquid helium Dewar that is inserted into the
background field coil in which the experiments are performed.
Figure 2
-
Photograph of the Test Coil Attached to the Mounting Unit
4.2 Instrumentation and Data Acquisition
The following ten analog signal sources were monitored: 1: The current through the pulse heater.
2: The voltage across the heater terminals.
3: The voltage generated by an iron-doped gold/constantan thermocouple; the thermo-couple was mounted directly on the conductor sheath at the center of the heated length.
4: The potential difference measured on the surface of the sheath at either end of the heated length.
5: The potential difference along one turn, measured on the surface of the sheath ap-proximately one turn "above" the heated length.
6: The potential difference along one turn, measured on the surface of the sheath ap-proximately one turn "below" the heated length.
7: The total potential difference between the ends of the sample under test. 8: The transport current in the sample.
9: The pressure inside the sheath, measured by a transducer mounted at one end of the two-in-hand winding.
10: The total potential difference between the ends of the other sample. The two samples were electrically in series, but only one at a time was subjected to energy perturba-tions; this channel monitored the voltage on the one not being pulsed.
Figure 3 shows a schematic diagram of the data acquisition setup. The ten analog signals were connected to the inputs of a LeCroy model 8212A 32-channel digitizer. The entire sequence of digitizing and storing the ten channels of data on the microVax's Winch-ester disk was handled by the MIT MDSt software package. Nine of the channels were also displayed on the video monitor a few seconds after completion of each digitize/store sequence. The experimental run was organized by shot number: each shot was manually initiated by discharging the capacitor into the heater. At the same instant, the data acqui-sition system began to convert and store the ten signals, and continued for two seconds, with a time resolution of 500 ps, yielding 4096 data points per channel for each shot.
E
0?~6Z cw*4cAS~Stj~g-,c E c JOIRST(IrfuuNo-z VejeqA-30 0 C £ C C C C C 4 C U 0 ~0 U, N C C C 0 C C 4 C C c C C C C C 4 4 4 4 C £ C *1~ C U U U c C 0 or U -C~o of U4 3..--o ~ - a.-a. L4 * U 'I.. a. ..-0 a a. 4 U 0 V c 0 0 ow-La. .H 0 f il. A L~S. -0 -44 W, C1 wU bo * ~ - h h h - & & -C 4 U, U UK
c c 'I cI
co Z r c c a ic 15 a CL L CL W ca L---i4.2.1 Procedure
Management of cryogenic and vacuum systems was carried out with no difficulty. Liquid helium level in the Dewar was monitored continuously with a conventional level sensor. Prior to the first shot, the ICCS was charged with pure helium gas, through a heat exchanger at 4.2 K, to a pressure of 3 atmospheres (about 44 psig). The intent was to begin every shot with this same initial pressure, but this was not always realized in practice.
Testing of each sample was carried out at three values of background field, 6 T, 7 T and 7.8 T.t The tests were conducted in the following manner: first, the background field was set at one of the three test values; then, these steps were carried out in a looping sequence:
I. Set the transport current to the desired value.
II. Charge the capacitor to the desired voltage; the choice was typically based on the results of prior shots.
III. Discharge the capacitor into the heater by triggered manually with a push-button, simultaneously triggering the digitizer.
IV. The sample voltage was monitored on an oscilloscope: in the event of rapidly rising voltage-indicative of a propagating normal zone-the current supply was manually interrupted to avoid damaging the sample.
V. Return to step I if more data are needed at the current field level, otherwise change background field to new value, and then go to step I.
During the run 67 shots were made, each one recording 2 seconds of digitized data from 10 signal sources. The stored data on Winchester disk was subsequently reduced and
analysed.
A measurement was made of the critical current of Sample A at each of the three test values of background field. Inasmuch as the current capacity of Sample A was less than that of Sample B and the two were in series, it was not possible to measure directly the critical current of B.
t
Values given here are for the central field at the midplane of the background field magnet (magnet 6B at the Francis Bitter National Magnet Laboratory). The actual values at the location of the samples were slightly different because the samples were wound on a 4.5 in. mandrel about 10 in. long: the field was higher by about 2-3% near the midplane bore OD where the perturbed length of sample was located, lower by 15-20% near the4.3 Results
4.3.1 Recalculations
Early in the course of the project, two sets of calculations had been made, representing two worst-case scenarios. However, the measured critical current for sample A exceeded the value claimed by the manufacturer; it was these latter values that had been used in calculations described in earlier reports. Because the actual values were significantly different, then, the calculations were repeated in order to provide a more realistic basis for comparison with the experimental results.
The first scenario was based on a computer model that had been developed to simulate transient stability phenomena, in which temperature rises following energy perturbations in the strands were calculated. The time scales for this model were rarely longer than 100 ms. For this scenario, the initial energy disturbance was deposited in the strands along the entire length of the conductor. The chief source of energy input in this case was joule heating. Figure 4 shows the result of these calculations; the energy pulse was 10 ms in duration.
The second worst-case scenario was different from the first in that all of the energy was taken to be supplied by the initial perturbation; this was done to simulate the case in which the energy is deposited at the surface of the sheath rather than in the strands, thereby - in the worst case - heating the interstitial helium preferentially. That is, the strand temperature could rise only by contact with the warming helium rather than by direct energy deposition or by contact with the warm sheath. This calculation, then, involved nothing more than computing the enthalpy of all the constituent materials. This result is shown in Figure 5.
The degree to which these calculated scenarios are representative of the physical ar-rangement of the constituent materials, and of the physical processes that occurred during the experimental testing was expected to be modest at best. For example, in calculating the Worst Case I, the perturbation was applied to the strands as a 10 ms inductive pulse over the whole length of the sample. The intent for the experiment was to apply the per-turbation to 30 cm of the a 4.6 m samples as a 10 ms current pulse to a heater soldered to the sheath. As thermocouple data showed later, however, it is likely that although the current pulse was 10 ms in duration, the actual pulse of heat energy at the sheath surface was more like 500 ms in duration; evidently, the ceramic insulation within the heater pro-vided a greater thermal barrier than expected. That is, the experimental conditions were such that close comparison of the results with the Worst Case I calculations is risky. It
MHD/ICCS Energy Margin -- Worst Case I 2000 Sample A Sample B 1500
1000
500<rr
7T0
0 0.2 0.4 0.6 0.8I/Ic
Figure 4 - The recalculated results for Worst Case I plotted for both A and B type conductor.
MHD/ICCS Calculated Enthalpy -- "Worst Case II"
0.2 0.4 0.6 0.8
Figure 5
-
The recalculated results for Worst Case II
plotted for both A and B type conductor.
11
2000
1500 1000500
0
Sample A 6 Sample B -8 7. 7.79T 0 1is worth noting that the test configuration is intended to simulate actual operating condi-tions and that the characteristic thermal behavior observed in the tests should be similar to that of an MHD magnet. This will be verified at proof-of-concept scale.
4.3.2 Experimental Data
The next two figures, 6 and 7, show all data points for both samples and three val-ues of background field; on each graph there is superimposed the curve representing the calculations for Worst Case II, that is, the enthalpy. The similarity between data and calculated enthalpy is quite good for both samples at 6 T nominal background field, less so for the other two values of field. The larger discrepancy at the higher fields may indicate either that the determination of critical current was more in error at the higher fields, or that joule heating made a larger contribution to the energy input at higher fields than at lower ones. Viewed as a whole, these six plots suggest that for the case of heat pulses of several hundred milliseconds duration applied to the surface of the ICCS sheath, most of the energy is absorbed by the interstitial helium, and that the temperature of the strands is raised to the current-sharing level chiefly by virtue of being in contact with this warming helium. Figure 8 is a plot of "smoothed" data for both samples, showing that when com-pared on the basis of I/Ic, the two conductors behaved comparably. This is a significant observation as it indicates that the two-in-one conductor should be considered further as an alternative to conventional all-multifilamentary strand ICCS.
However, the next series of sample A plots reveals that the situation is not that simple. The top trace in each plot is the thermocouple output, referenced to 0 K.t The middle trace is the resistance of the heated length of conductor, and the bottom trace is the total resistance of the sample. Superimposed on all three traces is the trace of transport current$, with the scale indicated on the right side axis.
Figures 9 and 10 show a pair of sequential shots, one of which led to a propagating normal zone (#81), and one that did not (#80). In #80, current-sharing commences at about 200 ms, but collapses after lasting for about 350 ms. A similar situation can be seen in the trace of shot #110 (Figure 11), although current-sharing in that case lasted longer. These traces show evidence that at least some of the strands are heated by contact with the sheath to a temperature above that of the interstitial helium, which is also being heated by contact with the sheath. But the source of heat - the sheath - reaches a temperature
t
40 IV = 4 K, 150 pV = 11.4 K, 250 tLV = 17.3 K.!
The decay in current apparent whenever current-sharing is underway is due to the fact that the power supply acts as a constant voltage source rather than as a constant current source.lower than the current-sharing temperature,Tc,, before the helium temperature rises to Tca, and therefore, the helium can continue to serve as a heat sink for the heated strands. In these two shots, the strands were able to cool to less than Tc,, despite the joule heating addition to their energy burden.
Snot #81 shows a case in which the strands, once current-sharing set in, were not cooled to less than Tc, before the interstitial helium's temperature was raised to T., by contact with sheath and the warming strands.
In order to measure the speed with which a normal zone would grow, voltage taps were attached to the sheath approximately three turns (24 inches) beyond the heated length in both directions. From limited data, the speed of propagation of the Tc, regime was calculated to be approximately 1.3 m/s. Additional data with voltage taps repositioned to be closer to the ends of the heated length would make it possible to verify this value.
-xx x Oxx X c x 4 'q. 0 (N 0 0 0CY 0 (~~ /fw)~~ xd .x 3 A~ u O x0 ,n 0 0l 4 a. '.1 0 0 0 (N 3 0o a 0 a 0 0n a U (X/rw) Ou3/A6Jau3 ;lOgH m c0 x -x *4 XX < x . a. c (0 a. A) Q) --0 ho .44 0 (N
(3o/() Xd~t~l3/X~au3 al*a 0 cm ~0 0 U C C C IN C C (~/r~) Xd C~u3Aju Ca~ 0. C 0 C IN ~0 0. IN 0 I C ( o /fw) XdIOLu3/A6J;u3 4al0 H a
15
so x0 x x ' x X X CIA a -x '3 X' X<< C -x -a 0 z x x ' 0 C C C IN C C 0C C 0 14 C PJ4MHD/ICCS Test Data -- Visually Smoothed
2000
Sample A Sample B ... 1500 6T U 7T 1000 Li 7.79T 0 5000
0 0.2 0.4 0.6 0.8I/Ic
Test; Shot: 81 Energy: 748.78 mj/cc; Field: 7.7900 T; Sample A
0.5 1.5
Figure 9 - Shot #81 Showing a Propagating Normal Zone
17
Iccs MHD/ 250 200 150 100 50 18 8 6 4 2 18 8 6 4 2 09
a-a (-3 I-C C -6 1000 800 600 400 2009oo
800 600 400 200Yomo
800 600 400 200 0 0. E E E 0 0 1 2MHD/ 250 200 150 100 50
18
8 6 4 2 8 6 4 2 0 0ICCS Test; Shot: 80 Energy: 535.06 mj/cc; Field: 7.7900 T; Sample A
0.5 1 1.5
time (sec)
Figure 10 - Shot #80 Showing the Initiation and Collapse of a Normal Zone
0.. 0 U C C a: 74 1000 800 600 400 200 9000 800 600 400 200 Yomo 800 600 400 200 0 2
MHD/ 250 200 150 100 50 8 6 4 2 8 6 4 2 0
ICCS Test; Shot: 110 Energy: 978.96 mj/cc; Fiel& 6.0000 T; Somple A
0 0.5 1 time (sec) 1.5 1000 800 600 400 200
9000
800 600 400 2009000
800 600 400 200 0 E 0 C-E 0~ E 0 2Figure 11
-
Shot #110 Showing a Normal Zone that Collapses after an Extended Time
19
a-0 U C: 764.4 Conclusions
Inasmuch as the effective duration of the energy pulse was greater than had been intended, no conclusions can be drawn regarding the comparative stability of these two conductors against short duration (1-10 millisecond) perturbations. The conclusions regarding their response to longer duration (a few hundred millisecond) heat inputs applied to the surface of the sheath are:
1) If operated at equivalent I/Ic, these two conductors will behave comparably with respect to their response to so-called 'long' duration heat perturbations at the sheath surface.
2) There is good thermal heat-sink action provided by the interstitial helium, enough so that the bulk of the energy needed to raise the temperature of the strands to Ta comes from the initial source of the perturbation.
3) Although the thermal contact between the strands and the sheath may be intermittent along a given conductor length, it is adequate to enable at least some of the strands to be heated by conduction from the inner surface of the sheath. This suggests that the electrical contact is also moderately good, although not necessarily equally good to all strands in any given length interval.
The general conclusion to be drawn from the tests is that it is reasonable to proceed with a proof-of-concept conductor design and test which will be based on the two-and-one subscale conductor. This work will be preceeded by further evaluation of an advanced magnet design concept for the retrofit-scale magnet which should ensure a cost-effective, reliable magnet system.
5.0 References
