in Applications Development
SYSTEMS/ROCHESTER GAS AND
This report covers only the nonproprietary ORNL contributions to the HTS Transformer Project for the FY 1996. Other than ORNL, the organizations participating in the project are the Intermagnetics General Corporation (IGC), Waukesha Electric Systems of General Signal Corporation (WES), and the Rochester Gas and Electric Company (RG&E). Rensselaer Polytechnic Institute provides consulting under subcontract to IGC. The objectives of the project are (1) to develop a low-cost, wind-after-react HTS suitable for use in transformers and other ac apparatus; (2) to determine the markets, technical and economic feasibility, and benefits to society of HTS power transformers of medium (30 MVA) to large rating; and (3) to design, build, and test an ~1-MVA single-phase HTS demonstration transformer. Significant progress in achieving the first two objectives has resulted in the decision to continue into the demonstration phase, which, together with the completion of key experiments, has been the major focus of efforts this year.
In FY 1996, the team
Design and Testing of Subsystems
Core, Tank, and Bushing Subsystem
Core, Tank, and Bushing Subsystem
It is highly desirable to use as many
components from existing transformer
technology as possible. With this in mind, we
decided to determine whether existing
transformer tank technology could be applied to
the HTS transformer. The resulting design for
the 1-MVA demonstration is externally very
similar to a full-scale commercial HTS tank and
resembles a conventional tank without a heat
exchanger. As a part of the development
process, a sub-sized tank was built at WES and
was successfully tested for weld integrity at
ORNL. The core section for the demonstration
unit is similar in size to a single phase of the
30-MVA design. This choice was made to allow
the use of the core for prototype testing of
several coil systems, to gain experience with
full-size components, and to allow use of
existing core technology at WES. A small core
cross section is presently undergoing tests at
ORNL for vacuum outgassing performance. In
addition, a full-scale bushing is undergoing tests
Cooling and Cold Mass Support
ORNL has lead responsibility for the 80 K
cold mass, liquid nitrogen, and cooling loop
subsystems. Initial design has been completed.
Final testing and integration awaits completion
of coil and core fabrication.
AC Loss Measurements in Coil
Two sets of ac loss tests were performed in
the ORNL variable-temperature cryostat on sub-sized transformer coil pairs provided by IGC.
The coils were fabricated from IGCs surface-coated BSCCO-2212 tape using winding
techniques based upon IGCs experience with
homopolar motor coils being developed for a
related Naval Research Laboratory program.
The ac loss measurements were made by
measuring the calibrated temperature rise in He
cooling gas (1545 K) that resulted from ac
current pulses. Fault-current performance was
also successfully completed on one-third height
near-design diameter coils at IGC.
TESTS OF THREE PROTOTYPE
During the initial phase of the CRADA between ORNL and Southwire Co. to develop HTS underground transmission cable, two 500-A-class and one 2000-A-class prototype cables were constructed. The cables and short samples of the Bi-2223/Ag HTS tapes were tested systematically at ORNL.
The cables were tested with both dc and ac
currents in liquid nitrogen. Both cables achieved
design currents; however, substantial
degradation in comparison with the short-sample critical currents (Ics) was observed. A
simple calorimetric technique was used to
measure the ac losses of the cables. A method of
utilizing the broad resistive transition of the
HTS cable was devised to calibrate the ac loss.
Different ac-loss behaviors were observed on
the insulated and uninsulated cables.
Short Sample Testing
|Fig. 2.1. Short sample of Ic along the length of the tape for use in cable 2.|
A series of short sample tests were performed on the Bi-2223/Ag HTS tapes acquired by Southwire Co. Seventy eight samples for the winding of the first cable and 11 samples for the winding of the second cable were measured. These 1-in.-long samples were tested in liquid nitrogen with up to 0.5-T magnetic field parallel and perpendicular to the wide face of the tape. Figure 2.1 shows the measured zero-field short sample Ics (at the 1-µV/cm criterion) along the length of the spool used to wind the second cable. Critical current varies significantly (by a factor of two) along the length of the tape. A mean Ic value of 20 A was measured (the end-to-end value was 17 A). Similarly, a mean Ic value of 19 A was measured for the tapes used to wind the first cable (the end-to-end value was 12 A). Apparently, damaged spots on a large spool were apt to be skipped when short samples (about 1-in. long) were taken.
Magnetic fields degrade the Bi-2223/Ag HTS tapes significantly at liquid nitrogen temperatures. At a background field value of 0.01 T, the present tapes showed an average of 10% degradation in Ic with field parallel to the wide face and 50 % degradation with field perpendicular to the wide face of the tape.
Bending tests were performed on selected samples of the HTS tapes. In a series of tests, I-V curves of 3-in.-long samples were measured before and after being wrapped side-by-side around a 1-in.-diam former. The samples from the lower-Ic spool showed an average degradation of 30%; those from the higher-Ic spool showed an average degradation of 53%.
Bending tests were also performed with
samples about 30-cm long by wrapping them
with lay angles of up to 30. Critical current
degradation between 40 and 50% of the 1-in.-long short sample values was observed.
Two prototype 500-A-class transmission cables were fabricated by Southwire using the 3.5 × 0.22 mm HTS tested tapes. The 1.2-m-long cables were made by spirally winding the tapes on a 22-mm (7/8-in.) copper former with lay angles of about 15.
|Fig. 2.2. Cable 1 assembled for both dc and ac current tests.|
For the first cable, no insulation was used to electrically separate the tapes. The ends of the tapes on the first layer were soldered onto the former. Successive layers were wound with alternating twist angles, and the ends were soldered to the previous layer. Seventy three tapes were wound in four layers in the first cable. Figure 2.2 shows a picture of the cable assembled and ready to be lowered into the test dewar. The main body of the cable was enclosed in a micarta pipe filled with wax to establish adiabatic conditions to measure the temperature rise (and thus the ac loss) of the cable.
The second cable was fabricated in a way similar to the first cable, except that Kapton tape was used between layers for insulation. Sixty six HTS tapes was used in the second cable.
The third Southwire cable was wound from
similar HTS tapes as were used in cables 1 and
2. The tapes were wound in 10 layers on a 1-in.-diam stainless steel former. Similar to layers in
cable 2, the successive layers were insulated
from each other with Kapton tape. Two hundred
HTS tapes were used in winding this cable.
DC Current Measurements
The electrical tests of the cables were
carried out in liquid nitrogen with the HTS cable
held upright in a 1.6-m-deep dewar.
|Fig. 2.3. CD I-V curves of cable 1 at the first cooldown.|
DC I-V of Cable 1
Four voltage taps were placed on the cable,
separated from each other by about 30 cm, and
were labeled as V1 to V4. Figure 2.3 shows the
I-V curves of the different sections of the cable
and of the whole cable (VTot). Gradual resistive
voltage rise was seen for currents starting at
about 400 A. All resistive voltage of the cable
came from the midsection (V23 ) at currents up to
650 A, caused by visible damage near the
middle of the cable. Nevertheless, the overall Ic
of 670 A at the 1-µV/cm criterion is higher than
the design value of 500 A.
|Fig. 2.4. I-V curves of outermost layer of cable 2 at three different bath temperatures.|
DC I-V of Cable 2
The layers of cable 2 were insulated from each other with Kapton tape, and separate current leads were brought out for each layer. Thus the cable could be tested as a whole or separately on individual layers. When the cable as a whole was tested, an Ic of 560 A was measured. Notice also that because of the broad resistive transition, both cable 1 and 2 can be operated stably at more than 1 kA.
During the test of the outermost layer of
cable 2, the liquid nitrogen bath was pumped to
lower temperatures. Figure 2.4 shows the I-V
curves of this layer at three different
temperatures of the liquid nitrogen bath. The Ic
o-f this layer increased from 149 to 186 A when
the bath temperature was lowered from 77 to
69 K. Thus an increase of about 25% in current-carrying capability can be achieved in the cable
by operating with subcooled liquid nitrogen (at
about 69 K).
DC Test of Cable 3
DC current test of the cable was performed
with a 2-kA power supply. Broad and smooth
resistive transition of the cable, similar to those
of the previous two cables, was observed. A
critical current of 1630 A was measured at the
1-µmV/cm criterion. At the power supply limit
of 2 kA, the cable produced an average resistive
voltage of 2.2 µV/cm.
|Fig. 2.5. U-E cyrves if cable 1 on successive thermal cycles.|
Thermal Cycle of Cable 1
After a few cycles of cooling down and
warming up of cable 1 for dc and ac current
measurements, a series of continuous thermal
cycle tests was performed. An I-V curve was
measured, and the cable was pulled out of the
liquid nitrogen bath. After it was warmed up to
room temperature in air, the sample was lowered
back down to the liquid nitrogen bath. Another
I-V curve was measured. Figure 2.5 shows a
series of these I-V curves at different thermal
cycles. Significant degradation was observed on
thermal cycling; however, the degradation
seems to level off after the fifth cycle. Critical
current of the cable decreased from 670 to
460 A after 10 thermal cycles (a 30 %
degradation). Power law fitting of the I-V curves
between 0.2 to 2 µV/cm also shows a decrease
of n-value from 3.5 to 2.6.
Comparison of Short Sample and
The measured Ic per tape of cables 1 and
2 averages about 8.8 A. This is significantly
lower than the average short sample value of
19.5 A measured on 1-in.-long short samples. As
is described in the series of short sample
measurements, several mechanisms can
contribute to the degradation of the cable Ic.
Short sample Ic measurements can be misleading
because they can skip bad spots in the long
lengths of the tape. Mechanical strain similar to
that applied in winding the cable can degrade
the Ic by about 50%. This can come from just
handling the long lengths of the tape and from
the bending applied in the cabling. The
magnetic-field degradation by the cable self-field is well known. In addition, thermal-cycling
degradation was also observed.
AC Current Measurements
Cables 1 and 2 were tested with 60-Hz ac currents up to 600 A rms. Steady rms voltages were observed at all test currents. A calorimetric technique was adopted to measure the ac loss of the cable at the applied ac currents. As is shown in Fig. 2.2, a micarta pipe filled with wax was used to thermally isolate the cable from the liquid nitrogen bath. A Cromel-Constantan thermocouple was attached to the middle of the cable to measure its temperature rise against liquid nitrogen at the same depth of the bath; temperature rise (T) of up to 0.3 K was observed.
To calibrate T against the power-loss rate, we used the broad resistive transition feature of the HTS cable itself. The cable was charged and held at a dc current above its Ic, where a resistive voltage can be measured. The T of the cable was also measured under this dc current. The dc EI product gave the average power loss for the measured T. This technique was found to be more responsive than the heater wires tried on cable 1. Because Ic is not uniform along the length of the cable, power generation is not uniform, but this is true for both dc power and ac loss. Therefore, the calibration technique is a good simulation of the ac loss. In addition, Joule heating at the cable ends did not contribute to the measured temperature rise because the ends of the cable were immersed in liquid nitrogen. This was verified by the observation that in the dc current calibration runs no temperature rise was observed until the current greatly exceeded Ic.
|Fig. 2.6. The ac losses of both cable 1 and 2 in reference to their DE I-E curve. For the ac current, rms currents is plotted.|
Figure 2.6 shows the measured ac losses of
the two cables as a function of the rms current.
Also shown in Fig. 2.6 for reference respective
dc I-E curves for two cables. Cable 1, which was
not insulated, behaved like a cryoresistive
conductor, showing power loss at all ac
currents. Similar behavior was reported by
Gannon et al.1 Cable 2 which was insulated,
showed no measurable ac loss until about 300 A
rms, where the cable also started to show
measurable dc resistive voltage. An average ac
loss of about 0.2 W/m was measured at 400 A
rms. Analysis of the loss data indicated that the
measured loss is governed by the power law
behavior of the HTS tape in the resistive
Cable 3 was also subjected to ac current
tests. Currents measuring up to 2.2 kA rms were
charged and held for about 10 min. Steady
voltage (mostly inductive voltage) was observed
over the cable in each case.
Two prototype 500-A-class HTS cables have been designed, constructed, and tested. Both cables achieved dc critical currents greater than the design value of 500 A. Furthermore, because of the broad resistive transition, they can be operated stably at more than 1 kA. A third cable was tested successfully to >2000 A ac current. Comparison of the cable Ic and the short sample values indicated a degradation of about 55%. Several mechanisms were identified as the probable cause of the degradation. Mechanical strain from handling the long lengths of the tape and from bending applied in winding the cable is thought to be the biggest source of degradation.
A calorimetric technique was used to
measure ac losses of the cables. A scheme of
utilizing the broad resistive transition of the
HTS cable was successfully used to calibrate the
loss rate. Loss measurements made on the cables
with 60-Hz ac currents showed that insulation
between the tapes is effective in reducing the ac
loss of the cable.