superconducting fault current limiter with integrat …
TRANSCRIPT
SUPERCONDUCTING FAULT CURRENT LIMITER WITH INTEGRAT ED
VACUUM INTERRUPTER
A thesis submitted to The University of Manchester for the degree of
Doctor of Philosophy
in the Faculty of Engineering and Physical Sciences
2012
Xiaoze Pei
SCHOOL OF ELECTRICAL AND ELECTRONIC ENGINEERING
Content
2
Content
List of Figures..................................................................................................................7
List of Tables .................................................................................................................14
Nomenclature.................................................................................................................15
List of Abbreviations ....................................................................................................19
Abstract..........................................................................................................................21
Declaration.....................................................................................................................23
Copyright Statement.....................................................................................................24
Acknowledgments .........................................................................................................25
The Author.....................................................................................................................26
1 Introduction.......................................................................................................27
1.1 Background .....................................................................................................27
1.2 Project motivation and context........................................................................29
1.3 Organisation of thesis......................................................................................32
2 Literature Review .............................................................................................34
2.1 Superconductivity ...........................................................................................34
2.1.1 Elementary properties of superconductors..........................................34
2.1.2 Materials..............................................................................................39
2.1.3 Summary .............................................................................................42
2.2 Fault current limitation....................................................................................42
2.2.1 Introduction.........................................................................................42
2.2.2 Operating principles of fault current limitation ..................................43
2.2.3 Application of fault current limiters....................................................44
2.2.4 Conventional methods.........................................................................45
2.2.5 Non-superconducting fault current limiters ........................................46
2.2.6 Superconducting fault current limiters................................................50
2.2.7 Summary .............................................................................................62
Content
3
2.3 Vacuum interrupter .........................................................................................64
2.3.1 Introduction.........................................................................................64
2.3.2 Vacuum interrupter structure ..............................................................65
2.3.3 Vacuum arc [94]..................................................................................66
2.3.4 Vacuum breakdown ............................................................................67
2.3.5 Current breaking in a vacuum.............................................................69
2.3.6 Practical design for high current levels...............................................70
2.3.7 Main application fields........................................................................71
2.3.8 Summary .............................................................................................72
2.4 Vacuum interrupter actuator ...........................................................................73
2.4.1 Spring actuator ....................................................................................73
2.4.2 Solenoid actuator.................................................................................73
2.4.3 Permanent magnetic actuator ..............................................................74
2.4.4 Voice-coil type actuator ......................................................................75
2.4.5 Summary .............................................................................................76
2.5 Conclusions.....................................................................................................77
3 SFCL Coil and Experimental Test Rig...........................................................78
3.1 Introduction.....................................................................................................78
3.2 Coil former......................................................................................................78
3.3 Coil manufacturing process.............................................................................80
3.3.1 MgB2 wire current connections and winding process.........................80
3.3.2 MgB2 wire heat treatment ...................................................................83
3.4 Instrumentation ...............................................................................................84
3.4.1 Voltage signals ....................................................................................84
3.4.2 Current signals ....................................................................................85
3.4.3 Temperature signals ............................................................................85
3.5 Control ............................................................................................................87
3.5.1 High-current test circuit ......................................................................87
3.5.2 Low-current test circuit .......................................................................88
3.5.3 LabVIEW control programme ............................................................89
3.5.4 Cryostat ...............................................................................................90
3.6 Conclusions.....................................................................................................93
4 Experimental Investigation of SFCL Coils.....................................................94
Content
4
4.1 Introduction.....................................................................................................94
4.2 Single-strand SFCL Coil .................................................................................95
4.2.1 Instrumentation ...................................................................................95
4.2.2 Calibration of BAS16 diode temperature sensors...............................97
4.2.3 Temperature profile.............................................................................98
4.2.4 Frequency sweep.................................................................................99
4.2.5 Quench tests ......................................................................................100
4.2.6 Long duration quench tests ...............................................................109
4.2.7 Simulated fault test............................................................................113
4.2.8 Temperature rise test .........................................................................113
4.2.9 Summary ...........................................................................................114
4.3 Three-strand SFCL Coil................................................................................115
4.3.1 Instrumentation .................................................................................115
4.3.2 Temperature profile...........................................................................116
4.3.3 Frequency sweep...............................................................................117
4.3.4 Current sharing test ...........................................................................119
4.3.5 Quench tests ......................................................................................120
4.3.6 Long duration quench test.................................................................125
4.3.7 Summary ...........................................................................................127
4.4 Conclusions...................................................................................................128
5 Modelling of SFCL Coil .................................................................................129
5.1 Introduction...................................................................................................129
5.2 MATLAB model...........................................................................................130
5.2.1 Model description..............................................................................130
5.2.2 Comparison with short cycle quench tests ........................................134
5.2.3 Comparison with long duration quench tests ....................................136
5.2.4 Summary ...........................................................................................139
5.3 Finite element thermal model........................................................................140
5.3.1 Model description..............................................................................140
5.3.2 Results and discussion ......................................................................144
5.3.3 Summary ...........................................................................................148
5.4 Prediction for three-second fault test ............................................................148
5.5 Conclusions...................................................................................................150
Content
5
6 Operating Actuator for Vacuum Interrupter .......... ....................................151
6.1 Introduction...................................................................................................151
6.2 Analytical model ...........................................................................................153
6.2.1 Model description..............................................................................153
6.2.2 Magnetic field in the airgap ..............................................................155
6.2.3 Effect of the actuator coil current on the electromagnetic force.......158
6.3 Finite element model.....................................................................................160
6.3.1 Model description..............................................................................160
6.3.2 Magnetic field distribution in the airgap...........................................162
6.3.3 Effect of the actuator coil current on the electromagnetic force.......167
6.3.4 Magnetic latch...................................................................................171
6.4 Design of the full operating actuator.............................................................172
6.4.1 Contact popping ................................................................................172
6.4.2 Contact bounce and rebound.............................................................173
6.4.3 Contact welding ................................................................................174
6.5 Design of prototype actuator and interrupter ................................................175
6.6 Construction of prototype operating actuator ...............................................178
6.6.1 Actuator stationary part.....................................................................178
6.6.2 Actuator moving part ........................................................................179
6.6.3 Complete prototype...........................................................................180
6.7 Conclusions...................................................................................................181
7 Design of the Actuator Control Circuit ........................................................182
7.1 Introduction...................................................................................................182
7.2 Control circuit ...............................................................................................182
7.2.1 Topology selection............................................................................182
7.2.2 Components selection .......................................................................183
7.3 Trigger signal and MOSFET drive circuit ....................................................185
7.3.1 Trigger signal circuit .........................................................................185
7.3.2 MOSFET drive circuit.......................................................................188
7.4 Conclusions...................................................................................................191
8 Experimental Investigation of an SFCL Coil with Integrated Vacuum
Interrupter ...................................................................................................................192
8.1 Introduction...................................................................................................192
Content
6
8.2 Testing of the vacuum interrupter operating actuator ...................................192
8.2.1 Actuator magnetic field and static force ...........................................192
8.2.2 Opening the vacuum interrupter at atmospheric pressure.................194
8.2.3 Opening the vacuum interrupter in a vacuum...................................198
8.2.4 Comparison of the tests at atmospheric pressure and in the vacuum202
8.3 Single-strand SFCL coil with and without the vacuum interrupter ..............203
8.3.1 Test rig diagram ................................................................................204
8.3.2 Quench test........................................................................................204
8.3.3 Simulated fault test............................................................................207
8.4 Conclusions...................................................................................................210
9 Conclusions and Further Research...............................................................211
9.1 Conclusions...................................................................................................211
9.2 Further research.............................................................................................214
References ....................................................................................................................215
Appendix A – Materials Data.....................................................................................228
Appendix B – Components for Operating Actuator................................................242
Total word count of thesis: 53732
List of Figures
7
List of Figures
Figure 1.1 Methods of fault current limitation [3].........................................................28
Figure 1.2 Potential applications of FCL devices [3] ....................................................28
Figure 1.3 Proposed resistive SFCL with vacuum interrupter.......................................31
Figure 2.1 Comparison of resistivity of normal conductors and a typical superconductor
[23] ..................................................................................................................................35
Figure 2.2 Exclusion of the magnetic field in a superconductor [25]............................36
Figure 2.3 Applied magnetic field: Type I (left) and Type II (right) superconductor [25]
.........................................................................................................................................37
Figure 2.4 Typical temperature profile of a Type II superconductor [25] .....................37
Figure 2.5 Configuration of SuperPower 2G HTS wire (wire type SCS4050) [30] ......40
Figure 2.6 Typical fault current waveforms with and without fault current limiting ....43
Figure 2.7 Solid-state FCL with a turn-off IGCT ..........................................................47
Figure 2.8 Solid-state FCL with thyristors.....................................................................47
Figure 2.9 Hybrid switching FCL ..................................................................................48
Figure 2.10 Thyristor-controlled resonant FCL.............................................................49
Figure 2.11 Thyristor-controlled series-parallel resonant FCL .....................................49
Figure 2.12 Thyristor-controlled series resonant FCL...................................................50
Figure 2.13 Resistive SFCL [2] .....................................................................................51
Figure 2.14 Bridge type SFCL.......................................................................................55
Figure 2.15 DC biased iron core SFCL..........................................................................57
Figure 2.16 Shielded iron core SFCL ............................................................................59
Figure 2.17 Fault current controller SFCL.....................................................................60
Figure 2.18 Flux-lock type SFCL ..................................................................................61
Figure 2.19 Typical structure of a vacuum interrupter [101].........................................65
Figure 2.20 Typical structure of a permanent magnetic actuator...................................74
Figure 3.1 Former specification (side view) [17]...........................................................79
Figure 3.2 Prototype former...........................................................................................80
Figure 3.3 Cryostat interior ............................................................................................81
Figure 3.4 Clamp used to connect wire to copper braid ................................................82
List of Figures
8
Figure 3.5 Top copper braid current connection............................................................82
Figure 3.6 Coil after heat treatment (left) and soldered joints (right) ............................83
Figure 3.7 Coil ready to be installed in the cryostat (left) and installed in the cryostat
(right)...............................................................................................................................84
Figure 3.8 Temperature monitor (top) and temperature controller (bottom) .................86
Figure 3.9 Cryocon S700 silicon diode temperature probes ..........................................86
Figure 3.10 BAS16 diode temperature sensor soldered onto the MgB2 wire ................87
Figure 3.11 High-current test circuit schematic [15, 17] ...............................................88
Figure 3.12 Low-current test circuit schematic [15, 17]................................................89
Figure 3.13 Screenshot of LabVIEW control programme.............................................90
Figure 3.14 Cryostat system: vacuum cryostat vessel with vacuum pump set (left),
vacuum pump (top middle), gauge (bottom middle) and cryocooler compressor (right)
.........................................................................................................................................91
Figure 3.15 Cryostat top plate........................................................................................92
Figure 4.1 Schematic showing locations of normal voltage taps (left), detailed voltage
taps and BAS16 diode temperature sensors (right).........................................................96
Figure 4.2 Temperature diode calibration curves when cooling down (1st) and warming
up (2nd) ...........................................................................................................................98
Figure 4.3 Temperature profile of the coil .....................................................................99
Figure 4.4 Impedance of the coil with varying frequency at 25K ...............................100
Figure 4.5 Total coil impedance with varying frequency at 25K ................................100
Figure 4.6 Coil response at 34K with a potential peak current of 311A......................102
Figure 4.7 Coil response at 34K with a potential peak current of 311A, highlighting the
point of quench..............................................................................................................102
Figure 4.8 Coil temperature response at 34K with a potential peak current of 311A..103
Figure 4.9 Coil response at 34K with a potential peak current of 372A......................104
Figure 4.10 Coil response at 34K with a potential peak current of 372A, highlighting
the point of quench........................................................................................................104
Figure 4.11 Coil temperature response at 34K with a potential peak current of 372A105
Figure 4.12 Coil response at 32K with a potential peak current of 622A....................105
Figure 4.13 Coil response at 32K with a potential peak current of 622A, highlighting
the point of quench........................................................................................................106
Figure 4.14 Coil temperature response at 32K with a potential peak current of 622A106
Figure 4.15 Coil response at 30K with a potential peak current of 700A....................107
List of Figures
9
Figure 4.16 Coil response at 30K with a potential peak current of 700A, highlighting
the point of quench........................................................................................................107
Figure 4.17 Coil temperature response at 30K with a potential peak current of 700A108
Figure 4.18 Estimated quench currents versus temperature.........................................108
Figure 4.19 Coil current response during a ten-cycle quench test with a potential peak
current of 372A .............................................................................................................110
Figure 4.20 Coil voltage response during a ten-cycle quench test with a potential peak
current of 372A .............................................................................................................110
Figure 4.21 Coil first turn detailed voltage response during a ten-cycle quench test with
a potential peak current of 372A...................................................................................111
Figure 4.22 Coil temperature response during a ten-cycle quench test with a potential
peak current of 372A.....................................................................................................111
Figure 4.23 Coil current response during a fifty-cycle quench test with a potential peak
current of 372A .............................................................................................................112
Figure 4.24 Coil temperature response during a fifty-cycle quench test with a potential
peak current of 372A.....................................................................................................112
Figure 4.25 Coil current response to simulated fault at 34K with a potential peak
current of 316A .............................................................................................................113
Figure 4.26 Coil temperature response during continuous 200Apeak current test at 30K
for one hour ...................................................................................................................114
Figure 4.27 Picture of the three-strand coil (Courtesy of Hyper Tech) .......................115
Figure 4.28 Schematic showing locations of the voltage taps .....................................116
Figure 4.29 Temperature profile of the coil .................................................................117
Figure 4.30 Impedance of the coil with varying frequency at 25K .............................117
Figure 4.31 Total coil impedance with varying frequency at 25K ..............................118
Figure 4.32 Coil response with current flow at 30K....................................................119
Figure 4.33 Coil response at 34K with a potential peak current of 249A....................120
Figure 4.34 Each coil strand voltage response of the first and fifth turns at 34K with a
potential peak current of 249A......................................................................................121
Figure 4.35 Coil response at 32K with a potential peak current of 467A....................121
Figure 4.36 Each coil strand voltage response of the first and fifth turns at 32K with a
potential peak current of 467A......................................................................................122
Figure 4.37 Coil response at 30K with a potential peak current of 529A....................123
List of Figures
10
Figure 4.38 Each coil strand voltage response of the first and fifth turns at 30K with a
potential peak current of 529A......................................................................................123
Figure 4.39 Each coil strand voltage response of the fifth turn at 30K with a potential
peak current of 529A.....................................................................................................124
Figure 4.40 Estimated quench currents versus temperature.........................................125
Figure 4.41 Coil current response during a ten-cycle quench test with a potential peak
current of 249A .............................................................................................................126
Figure 4.42 Coil voltage response during a ten-cycle quench test with a potential peak
current of 249A .............................................................................................................126
Figure 4.43 Each coil strand voltage response of the first and fifth turns during a ten-
cycle quench test with a potential peak current of 249A ..............................................127
Figure 5.1 Results comparison for a fault at 34K with a potential peak current of 372A
.......................................................................................................................................135
Figure 5.2 Results comparison for a fault at 32K with a potential peak current of 622A
.......................................................................................................................................135
Figure 5.3 Results comparison for a fault at 30K with a potential peak current of 700A
.......................................................................................................................................136
Figure 5.4 Results comparison for a ten-cycle fault at 34K with a potential peak current
of 372A .........................................................................................................................137
Figure 5.5 Results comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A .............................................................................................................137
Figure 5.6 Results comparison for a ten-cycle fault at 34K with a potential peak current
of 372A (considering the heat dissipated into the nitrogen) .........................................138
Figure 5.7 Results comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A (considering the heat dissipated into the nitrogen).............................138
Figure 5.8 Temperature comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A with adiabatic and nitrogen boundary conditions...............................139
Figure 5.9 Flux2D FE model of the geometry .............................................................141
Figure 5.10 Mesh of the Flux2D FE model, showing detail of the coil in the slot......141
Figure 5.11 Instantaneous power loss density in the coil during a one-second fault at
34K with a potential peak current of 372A...................................................................143
Figure 5.12 Time variation of temperature in the centre of the coil for varying distances
from the former .............................................................................................................145
List of Figures
11
Figure 5.13 Temperature profile vertically down the former through the centre of the
coil after a one-second fault, with a gap of 1.5mm from the former ............................145
Figure 5.14 Temperature profile after a one-second fault, with a gap of 1.5mm from the
former............................................................................................................................146
Figure 5.15 Temperature response comparison for a one-second fault at 34K with a
potential peak current of 372A from the MATLAB model and the FE model.............147
Figure 5.16 Temperature profile through the centre of the coil for a one-second fault at
34K with a potential peak current of 372A from the MATLAB model and the FE model
.......................................................................................................................................148
Figure 5.17 Modelled current response for a three-second fault at 34K with a potential
peak current of 372A.....................................................................................................149
Figure 5.18 Temperature response comparison for a three-second fault at 34 K with a
potential peak current of 327A from the MATLAB model and the FE model.............149
Figure 6.1 DVS10CB vacuum interrupter ...................................................................151
Figure 6.2 Geometry of the actuator (the two steel walls in the front are not shown).154
Figure 6.3 Simplified model of the magnetic circuit ...................................................154
Figure 6.4 Actuator permanent magnet operating points on the 2nd quadrant B-H
characteristic (demagnetisation curve) at 20oC.............................................................155
Figure 6.5 Geometry of the actuator displayed in Vector Fields Opera: full view
without steel walls at the front (left) and plan view (right)...........................................160
Figure 6.6 Geometry of the actuator displayed in Vector Fields Opera (only a quarter is
shown using model symmetry) .....................................................................................161
Figure 6.7 N48 permanent magnet normal demagnetisation curve at 20ºC [143].......161
Figure 6.8 EN1A mild steel B-H curve [147] ..............................................................162
Figure 6.9 3-D plot of the flux density distribution produced by the magnets ............163
Figure 6.10 Flux density distribution in the airgap produced by the magnets.............163
Figure 6.11 Cross-sectional view of the flux density distribution with vectors produced
by the magnets...............................................................................................................164
Figure 6.12 Flux density distribution in the airgap produced by the magnets (along the
path noted in Figure 6.11) .............................................................................................165
Figure 6.13 Plan view of the flux density distribution with vectors produced by the
magnets .........................................................................................................................166
Figure 6.14 Flux density distribution in the airgap produced by the magnets (along the
path noted in Figure 6.13) .............................................................................................166
List of Figures
12
Figure 6.15 3-D plot of the flux density distribution produced by the magnets and the
actuator coil carrying 50A in the anti-clockwise direction ...........................................168
Figure 6.16 Flux density distribution in the airgap produced by the magnets and the
actuator coil carrying 50A in the anti-clockwise direction (along the vertical path)....168
Figure 6.17 3-D plot of the flux density distribution produced by the magnets and the
actuator coil carrying 50A in the clockwise direction ..................................................169
Figure 6.18 Flux density distribution in the airgap produced by the magnets and the
actuator coil carrying 50A in the clockwise direction ..................................................169
Figure 6.19 Actuator force versus current characteristics............................................170
Figure 6.20 Geometry of the actuator with magnetic latches (the two steel walls and
latches at the front are not shown) (left) and cross-sectional view (right)....................171
Figure 6.21 3-D structure of the operating actuator.....................................................176
Figure 6.22 Cross-sectional view of the vacuum interrupter actuator (all dimensions in
mm) ...............................................................................................................................176
Figure 6.23 Plan view of the vacuum interrupter actuator (all dimensions in mm).....177
Figure 6.24 Plan view of the stopper and carbon fibre plate with latch steel (all
dimensions in mm)........................................................................................................177
Figure 6.25 N48 Nd-Fe-B permanent magnet..............................................................178
Figure 6.26 Steel frame with permanent magnets........................................................179
Figure 6.27 Actuator coil on the fibreglass tube..........................................................179
Figure 6.28 Operating actuator with vacuum interrupter.............................................180
Figure 7.1 Schematic diagram of the actuator control circuit ......................................183
Figure 7.2 Schematic diagram of the precision full-wave rectifier [154] ....................186
Figure 7.3 Schematic diagram of the ‘open’ signal circuit for the vacuum interrupter187
Figure 7.4 Schematic diagram of the ‘close’ signal circuit for the vacuum interrupter
.......................................................................................................................................188
Figure 7.5 Schematic diagram of the MOSFET drive circuit ......................................189
Figure 7.6 Trigger signal and MOSFET drive circuit..................................................190
Figure 7.7 Vacuum interrupter with its actuator and control circuit............................190
Figure 8.1 Actuator opening and closing force versus current ....................................194
Figure 8.2 Opening operation of the vacuum interrupter with a capacitor voltage of
100V at atmospheric pressure .......................................................................................196
Figure 8.3 Opening operation of the vacuum interrupter with different capacitor
voltages at atmospheric pressure...................................................................................197
List of Figures
13
Figure 8.4 Operating actuator in the vacuum chamber (left) and external view of the
vacuum chamber (right) ................................................................................................199
Figure 8.5 External connectors for the vacuum interrupter and its actuator................199
Figure 8.6 Opening of the vacuum interrupter with different capacitor voltages in the
vacuum chamber ...........................................................................................................200
Figure 8.7 Opening of the vacuum interrupter with different capacitor voltages in the
vacuum chamber with the current duration increased to 15ms.....................................201
Figure 8.8 Comparison of the opening of the vacuum interrupter at atmospheric
pressure and in the vacuum...........................................................................................202
Figure 8.9 Schematic of the high-current test circuit with the vacuum interrupter .....204
Figure 8.10 Quench test with a potential peak current of 324A with and without the
vacuum interrupter ........................................................................................................206
Figure 8.11 Temperature rise of the coil during a quench test with a potential peak
current of 324A with and without the vacuum interrupter............................................207
Figure 8.12 Simulated fault test without the vacuum interrupter.................................208
Figure 8.13 Simulated fault test with the vacuum interrupter......................................209
List of Tables
14
List of Tables
Table 2.1 Summary of active SFCL projects .................................................................63
Table 4.1 MgB2 wire specification.................................................................................95
Table 5.1 Temperature response comparison with fault time variation.......................147
Table 6.1 DVS10CB vacuum interrupter specification...............................................151
Table A.1 Thermal conductivity of alumina................................................................228
Table A.2 Specific heat capacity of alumina ...............................................................229
Table A.3 Density of alumina ......................................................................................231
Table A.4 Thermal conductivity of copper ..................................................................231
Table A.5 Specific heat capacity of copper .................................................................232
Table A.6 Density of copper ........................................................................................233
Table A.7 Resistivity of copper ...................................................................................233
Table A.8 Thermal conductivity of magnesium diboride ............................................234
Table A.9 Specific heat capacity of magnesium diboride............................................234
Table A.10 Density of magnesium diboride ................................................................235
Table A.11 Thermal conductivity of monel.................................................................236
Table A.12 Specific heat capacity of monel ................................................................237
Table A.13 Density of monel .......................................................................................237
Table A.14 Thermal conductivity of niobium..............................................................237
Table A.15 Specific heat capacity of niobium.............................................................238
Table A.16 Density of niobium....................................................................................239
Table A.17 Thermal conductivity of nitrogen..............................................................240
Table A.18 Specific heat capacity of nitrogen.............................................................240
Table A.19 Density of nitrogen....................................................................................241
Table A.20 Thermal conductivity of polystyrene ........................................................241
Table A.21 Specific heat capacity of polystyrene........................................................241
Table A.22 Density of polystyrene ..............................................................................241
Table B.1 Summary of components for prototype operating actuator.........................242
Nomenclature
15
Nomenclature
Symbol Meaning SI Units
Aac Cross-sectional area of the actuator coil m2
Acb Cross-sectional area of copper in the copper braid m2
Ag Area of the airgap m2
Am Permanent magnet arc surface area m2
Apeak Peak current A
Arms Root mean square current A
B Magnetic flux density T
Bc Critical magnetic flux density T
Bg Magnetic flux density in the airgap T
Bm Permanent magnet magnetic flux density T
Br Remanence in the permanent magnet T
Cp Specific heat capacity J/kg·K
d Distance between two electrodes m
di
dt Rate of change of current A/s
Adi
dt
Rate of change of current in strand A of the three-strand SFCL
coil A/s
Bdi
dt
Rate of change of current in strand B of the three-strand SFCL
coil A/s
Cdi
dt
Rate of change of current in strand C of the three-strand SFCL
coil A/s
sdi
dt Rate of change of current in the high-current test circuit A/s
dv
dt Rate of change of voltage V/s
E Back-emf V
F Force on the actuator coil N
Fc Actuator closing force N
Nomenclature
16
Fo Actuator opening force N
Fr Actuator reluctance force N
g Power loss density W/m3
gc Power generated in the SFCL coil per unit volume W/m3
gcb Power loss density in the copper braid W/m3
gcm Power loss density in the SFCL coil (calculated from
measurement) W/m3
Hc Critical magnetic field strength A/m
Hc1 Lower critical magnetic field strength A/m
Hc2 Upper critical magnetic field strength A/m
Hg Airgap magnetic field strength A/m
Hm Permanent magnet magnetic field strength A/m
I Current in the actuator coil A
IA Current in strand A of the three-strand SFCL coil A
IB Current in strand B of the three-strand SFCL coil A
Icm Measured current passing through the SFCL coil A
IC Current in strand C of the three-strand SFCL coil A
Is High-current test circuit current A
Iq Quench current A
Jc Critical current density A/mm2
k Thermal conductivity W/m·K
K Spring stiffness N/mm
Kp Permeance coefficient H/m
l Total length of the actuator coil m
lmf Length of the actuator coil in the magnetic field m
L Inductance of the actuator coil H
LA Self-inductance of strand A of the three-strand SFCL coil H
LB Self-inductance of strand B of the three-strand SFCL coil H
Lc High-current test circuit inductance H
Lcb Length of the copper braid m
LC Self-inductance of strand C of the three-strand SFCL coil H
lg Length of the airgap m
lm Length of the permanent magnet m
Nomenclature
17
M Mutual inductance H
MAB Mutual inductance between strand A and strand B H
MAC Mutual inductance between strand A and strand C H
MBA Mutual inductance between strand B and strand A H
MBC Mutual inductance between strand B and strand C H
MCA Mutual inductance between strand C and strand A H
MCB Mutual inductance between strand C and strand B H
M1 Mass of the moving part kg
M2 Mass of the stationary part kg
N Turns of the actuator coil -
P Pressure Pa
r Distance from the centre along the radial direction m
R Resistance of the actuator coil Ω
RA Resistance of strand A of the three-strand SFCL coil Ω
RB Resistance of strand B of the three-strand SFCL coil Ω
Rc High-current test circuit resistance Ω
RC Resistance of strand C of the three-strand SFCL coil Ω
Rcb Resistance of the copper braid Ω
Rfcl SFCL coil resistance Ω
t Time s
T Temperature K
Tc Critical temperature K
v SFCL coil volume m3
vc Velocity m/s
vcb Volume of the copper braid m3
vcm Measured voltage across the SFCL coil V
v10 Initial velocity of the moving contact before collision m/s
V Voltage of supply for the actuator coil V
VA Voltage across strand A of the three-strand SFCL coil V
VB Voltage across strand B of the three-strand SFCL coil V
Vc Voltage across the capacitor V
VC Voltage across strand C of the three-strand SFCL coil V
Vs Voltage of the high-current test circuit V
Nomenclature
18
V0 Voltage across the capacitor at time t = 0 V
x Distance along the SFCL coil length m
z Distance from the centre along Z-axis direction m
ρ Resistivity of copper Ω·m
ρd Density kg/m3
ρ0 Normal state resistivity of a superconductor Ω·m
µ0 Permeability of free space (4π×10-7) H/m
µm Relative permeability of the permanent magnet -
List of Abbreviations
19
List of Abbreviations
1G First generation
2G Second generation
1-D One-dimensional
2-D Two-dimensional
3-D Three-dimensional
AC Alternating current
Al 2O3 Alumina
AMF Axial magnetic field
AMSC American Superconductor
ASL Applied Superconducting Limited
BIL Basic impulse level
BSCCO Bismuth strontium calcium copper oxide
CAS Chinese Academy of Sciences
CIEE Chinese Institute of Electrical Engineering
CTFF Continuous tube forming and filling
CuCr Copper chrome
DC Direct current
DoE Department of Energy
emf Electromotive force
EPR Electromagnetic repulsion plate
EU European Union
FE Finite element
FCL Fault current limiter
HTS High temperature superconductor
IBAD Ion beam assisted deposition
IGCT Integrated gate commutated thyristor
KEPCO Korea Electric Power Corporation
KEPRI Korea Electric Power Research Institute
List of Abbreviations
20
LBCO Lanthanum barium copper oxide
LSIS LS Industrial Systems
LTS Low temperature superconductor
LVDT Linear variable differential transformer
MCP Melt, cast and processed
MFCL Matrix fault current limiter
MFP Mean free path
MgB2 Magnesium diboride
MOCVD Metal organic chemical vapour deposition
MOD Metal-organic deposition
MOSFET Metal-oxide-semiconductor field-effect transistor
MRI Magnetic resonance imaging
NbTi Niobium titanium
Nd-Fe-B Neodymium-iron-boron
OD Outer diameter
PIT Powder-in-tube
PLD Pulsed laser deposition
RABiTS Rolling assisted biaxial textured substrates
RMF Radial magnetic field
RMS Root mean square
SCE Southern California Edison
SFCL Superconducting fault current limiter
SF6 Sulfur hexafluoride
SRBL Synthetic resin bonded lamination
TRV Transient recovery voltage
U.K. United Kingdom
U.S. United States
U.S.A. United States of America
VI Vacuum Interrupter
YBCO Yttrium barium copper oxide
Abstract
21
Abstract
Fault current levels in land-based power systems are generally rising because of the
increase in renewable generation capacity. Once the fault current level exceeds the
capacity of the existing protection equipment, expensive upgrades become necessary. In
order to avoid excessively expensive equipment upgrades, many fault current limitation
techniques have been investigated.
This thesis presents the work conducted on the design, manufacture and testing of a
resistive superconducting fault current limiter (SFCL) with an integrated fast-acting
vacuum interrupter. The practical application of magnesium diboride (MgB2) in round
wire form was also investigated.
A single-strand MgB2 SFCL coil was investigated and demonstrated repeatable and
reliable current-limiting action. In practical power system applications, the development
of SFCLs needs a considerable scale-up of the current-carrying capability of the MgB2
wire samples. One option is to use parallel wires in order to carry current levels in the
kA range. The behaviour of a prototype three-strand MgB2 SFCL coil was assessed,
which showed that each of the three wire strands shared the current approximately
equally and demonstrated reliable and repeatable behaviour during testing. The MgB2
SFCL coil with multiple wire strands in parallel shows considerable potential as a
practical method for scaling-up the current levels required for power system
applications.
One of the significant operational issues for resistive SFCLs is the temperature recovery
time after a fault is cleared. A vacuum interrupter was integrated therefore into the
SFCL system to quickly remove the superconducting coil from the circuit during a fault
condition and allow the superconducting coil to recover whilst a bypass resistor acted as
a current limiting resistor. A fast-acting actuator and its control circuit were designed
and manufactured to control the operation of the vacuum interrupter. The SFCL with a
prototype vacuum interrupter was successfully tested to validate the design process. The
Abstract
22
energy dissipated in the superconducting coil was significantly reduced by the fast
operation of the vacuum interrupter and the recovery time significantly reduced. This
research demonstrates the potential of a cost-effective and compact SFCL for the power
system applications.
Declaration
23
Declaration
No portion of the work referred to in this thesis has been submitted in support of an
application for another degree or qualification of this or any other university or institute
of learning.
Copyright Statement
24
Copyright Statement
i. The author of this thesis (including any appendices and/or schedules to this
thesis) owns any copyright in it (the “Copyright”) and s/he has given The
University of Manchester the right to use such Copyright for any administrative,
promotional, educational and/or teaching purposes.
ii. Copies of this thesis, either in full or in extracts, may be made only in
accordance with the regulations of the John Rylands University Library of
Manchester. Details of these regulations may be obtained from the Librarian.
The page must form part of any such copies made.
iii. The ownership of any patents, designs, trade marks and any and all other
intellectual property rights except for the Copyright (the “Intellectual Property
Rights”) and any reproductions of copyright works, for example graphs and
tables (“Reproductions”), which may be described in this thesis, may not be
owned by the author and may be owned by third parties. Such Intellectual
Property Rights and Reproductions cannot and must not be made available for
use without prior written permission of the owner(s) of the relevant Intellectual
Property Rights and/or Reproductions.
iv. Further information on the conditions under which disclosure, publication and
commercialisation of this thesis, the Copyright and any Intellectual Property
and/or Reproductions described in it may take place is available in the
University IP Policy (see
http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-
property.pdf), in any relevant Thesis restriction declarations deposited in the
University Library, The University Library’s regulations (see
http://www.manchester.ac.uk/library/aboutus/regulations) and in The
University’s policy on presentation of Theses.
Acknowledgments
25
Acknowledgments
I am sincerely grateful to my academic supervisor, Professor Sandy Smith, for his
invaluable guidance and support during my Ph.D. research. I would therefore like to
take this opportunity to express my gratefulness and appreciation for his encouragement
and great patience.
I would like to thank the Dorothy Hodgkin Postgraduate Award, EPSRC and Rolls-
Royce Plc. for the financial funding of my Ph.D. research. Thanks are also due to Mark
Husband from Rolls-Royce for his guidance during the work.
I would like to thank all the academic staff, research associates and Ph.D. candidates of
the Power Conversion Group. A special thanks to Dr Paul Tuohy for his help
proofreading this thesis.
Thanks are also due to Malcolm Bailey and Paul Shaw from the School of Electrical
and Electronic Engineering’s electronics and mechanical workshops for manufacturing
some of the equipment used in the experimental test.
Finally, I would like to thank my family for their blessings and support, especially to
my husband Xianwu Zeng for his encouragement and accompanying me in the U.K.
The Author
26
The Author
Xiaoze Pei obtained her BEng (Hons) in Electrical Engineering from Beijing Jiaotong
University in 2006. Thereafter, she received her MEng (Hons) in Power Electronics and
Power Drive from Beijing Jiaotong University in 2008. The work described in this
thesis, which started in 2008, constitutes the author’s first major research project. The
following paper has been published:
Xiaoze Pei, Alexander C. Smith, Mark Husband, and Matthew Rindfleisch,
“Experimental Tests on a Superconducting Fault Current Limiter using Three-
strand MgB2 Wire”, in 22nd International Conference on Magnet Technology,
Marseille, France, 2011, pp 1-5, [Accepted by IEEE Transactions on Applied
Superconductivity].
Chapter 1: Introduction
27
1 Introduction
1.1 Background
The generation and transmission capacity of electric power systems is increasing to
meet growing demands [1, 2]. In addition, the interconnections within the power grids
increase to improve power quality and reliability. Distributed renewable energy
generation such as solar energy and wind energy, are also being connected into power
systems. All these connected systems cause the capacity of the power systems to
increase, which leads to the potential of increasing fault current levels. Increasing fault
current levels in transmission and distribution systems have become a serious problem
for land-based power systems.
The rising fault current levels will ultimately require a large amount of equipment in the
existing power systems such as circuit breakers and transformers, to be replaced. In
order to avoid large scale and expensive equipment upgrades, cost-effective fault
current limitation techniques are being examined [3].
In the past several decades, a variety of fault current limitation techniques have been
investigated, summarised in Figure 1.1 [3]. Generally, they fall into two broad
categories: ‘permanent impedance increase’ and ‘condition based impedance increase’
[3]. The first category includes techniques such as introducing high impedance
transformers and current limiting reactors, which add extra impedance during both
normal and fault conditions. Amongst the latter category, novel fault current limiters
(FCLs) using superconductors and/or semiconductors have been studied [2]. FCLs
introduce negligible impedance during normal operation and are almost invisible to the
power system. Once a fault occurs, however, the impedance of the FCLs increases and
reduces the fault current to a level that is acceptable to the existing protection equipment
in the system. Potential applications for FCLs in power networks are shown in Figure
1.2.
Chapter 1: Introduction
28
Figure 1.1 Methods of fault current limitation [3]
Figure 1.2 Potential applications of FCL devices [3]
The discovery of high temperature superconductors (HTS) in the 1980’s significantly
improved the potential for the practical application of superconducting fault current
Chapter 1: Introduction
29
limiters (SFCLs). SFCLs utilise the characteristics of the superconducting materials to
limit the fault current. Many SFCL design concepts have been evaluated [2, 4-7];
resistive SFCLs have the simplest and most compact structure [4, 8-11]. Bismuth
strontium calcium copper oxide (BSCCO) and yttrium barium copper oxide (YBCO)
have been widely researched for SFCLs applications. More recently, magnesium
diboride (MgB2) in simple round wire form has been tested and shown to have
significant potential as a low-cost resistive SFCL [12-15]. The critical temperature of
MgB2 is 39K and, for this thesis study, the MgB2 is operating in solid nitrogen (<63K).
The most serious problem with resistive SFCLs is overheating during a fault and the
long recovery time of the superconducting material. This research proposes therefore to
integrate a vacuum interrupter with a resistive SFCL to overcome this problem.
1.2 Project motivation and context
Resistive SFCLs have been previously investigated because of their simple structure
and making direct use of the non-linear characteristics of the superconducting materials.
MgB2 has been manufactured in round wire form which makes coil manufacture easier.
MgB2 also provides the potential to be a cost-effective and economic material for SFCL
applications because the raw materials are cheaper than BSCCO and the manufacturing
process is simpler than YBCO coated conductors [16]. A prototype single-strand MgB2
SFCL coil was manufactured and tested by Oliver [17] and the coil demonstrated
successful operation via an experimental test rig was built for the test of the prototype
coil. In addition, Oliver [17] implemented a finite element (FE) model for the thermal
design of the alumina former and a MATLAB model to simulate the behaviour of the
SFCL coil [15, 17]. This thesis builds on the research work reported by Oliver [17]
expanding the experimental studies, thermal modelling and simulation. Additionally,
this research also integrates vacuum interrupter mechanism.
As with the work of Oliver [17] a resistive SFCL using round MgB2 wire as the
superconducting material was initially investigated in this thesis. The wire had
improved manufacturing and mechanical properties to that used by Oliver.
Chapter 1: Introduction
30
For typical distribution system voltages of 11kV or 33kV, for example, the development
of an MgB2 SFCL needs considerable scale-up in term of the current-carrying capability
of the MgB2 wire. The two obvious options to achieve this are to increase the wire size
and use a multifilament core and/or use multiple wire strands. Large wire sizes in
monofilament superconductors may result in a condition referred to as ‘flux jumping’, a
phenomena that can lead to instability such that the material moves from a
superconducting state to a normal conducting state. Thus the wire diameter is limited
[18]. Typical distribution level currents of 1250Arms, for example, mean it is unlikely
that a single wire SFCL will be feasible [19]. In practical power system applications,
parallel wires will likely be required in order to carry current levels in the kA range. The
full quench current level will be reduced; however, if the individual wire strands (and
connections) are not closely identical in terms of their quench current level and
impedance. Ensuring each strand of wire carries the same current is therefore an
important issue for multi-strand wire solutions. The behaviour of a three-strand MgB2
SFCL coil is assessed following the investigation of a single-strand SFCL coil.
The major disadvantage of a resistive SFCL is that the superconducting material heats
up rapidly after quenching during a fault condition and may take several minutes to cool
down and recover after the fault is cleared. The purpose of this research project is
therefore to explore the potential of integrating a vacuum interrupter into the resistive
SFCL directly. The cryostat used to cool the superconducting material typically uses a
vacuum around the cold chamber as super thermal insulation. The objective is to mount
a vacuum interrupter into the vacuum chamber of the cryostat in series with the
superconducting coil. A bypass resistor is then connected in parallel with both of them
but external to the cryostat. The schematic circuit of the proposed resistive SFCL with
an integrated vacuum interrupter is shown in Figure 1.3. During normal operation, the
impedance of the vacuum interrupter and superconducting coil remain essentially zero
such that the bypass resistor conducts virtually no current. When a fault occurs, the
superconducting coil begins to quench and starts to develop a resistance. The voltage
that appears across the superconducting coil then increases and can be used as the
trigger signal for the vacuum interrupter. The vacuum interrupter needs to be designed
to open quickly and isolate the superconducting coil from the system. The bypass
resistor acts as the fault current limiting resistance after the vacuum interrupter opens.
Chapter 1: Introduction
31
The temperature rise of the superconducting coil is reduced as a result and additionally
the coil starts to recover whilst the bypass resistor continues to limit the fault current.
The bypass resistor would be in the circuit during the fault for typically a few hundred
milliseconds before the back-up circuit breaker isolates the fault. The bypass resistor is
therefore designed and rated for short-term operation which makes it compact and
relatively low-cost. The fast-acting vacuum interrupter will reduce the energy dissipated
in the superconducting coil and allows for a faster recovery time compared to a basic
resistive SFCL. It is demonstrated in this thesis that it is possible to operate the vacuum
interrupter within a half-cycle of the onset of the fault current and also possible to
reconnect the SFCL coil within seconds rather than minutes.
Figure 1.3 Proposed resistive SFCL with vacuum interrupter
A fast operating actuator for the vacuum interrupter is a key specification requirement
for this system and hence a main feature of this research project. The characteristics of
existing vacuum interrupters and actuator technologies are reviewed. A ‘voice-coil’ type
actuator with magnetic latches is proposed for this application. The actuator is expected
to provide fast opening and closing of the vacuum interrupter whilst magnetic latches
can hold the actuator in the open or closed position.
It is worth noting the differences between the proposed research and a hybrid SFCL. In
a hybrid SFCL, an interrupter is driven by the fault current through a parallel coil [20,
21]. The operating of the interrupter is therefore strongly affected by the fault current
level. When the fault current is low, the opening stroke of the interrupter is small and
the current in the superconducting coil may fail to interrupt because of insufficient
Chapter 1: Introduction
32
clearance distance between the contacts [22]. In addition, there is a risk of the
interrupter reclosing before the fault current is cleared. In this research project, the
vacuum interrupter is operated by a separate fast-acting actuator whose operation is
independent from the fault current level. If successful, this concept would represent a
technical step forward in realising a practical low-cost SFCL solution.
1.3 Organisation of thesis
Chapter 2 explains the characteristics of the superconducting materials related to this
research project. Three superconducting materials utilised in current SFCL research
projects are also covered. A review of fault current limitation theory, conventional
practical methods of fault current limitation, novel FCLs including non-superconducting
and superconducting FCLs, is reported. A summary of the characteristics of the vacuum
interrupter and its applications are also presented. In addition, four different structures
of vacuum interrupter actuators are compared.
Chapter 3 describes the manufacturing process of the MgB2 SFCL coil undertaken
during the study, which includes the current connections, winding the coil and the heat
treatment process. The instrumentation of voltage, current and temperature signals on
the prototype coil is introduced. High-current and low-current experimental test circuits
and a LabVIEW control system are also presented.
Chapter 4 presents the experimental investigation of a single-strand and three-strand
MgB2 SFCL coil. The three-strand MgB2 SFCL coil is evaluated to explore the potential
of increasing the current carrying capacity using multi-strand wires. The quench
behaviour of both SFCL coils is also evaluated.
Chapter 5 presents the work carried out to expand and improve the MATLAB and finite
element thermal model of the single-strand MgB2 SFCL coil presented by Oliver [17].
The development of the MATLAB model is to help understand the SFCL current-
limiting process and inform future design work. The FE thermal model is built to
calculate the temperature distribution around the coil and surrounding cryogen. The
results of both the MATLAB and FE models are compared with the experimental test
results to verify the accuracy of the models. The two models are then used to predict a
Chapter 1: Introduction
33
three-second quench test to check whether the SFCL coil can withstand a fault of this
duration.
Chapter 6 describes the design and manufacture of an operating actuator for the vacuum
interrupter. A ‘voice-coil’ type actuator is used due to its simple structure and quick
response time. An analytical model and FE model of the actuator are built to estimate
the magnetic field in the airgap and resultant force produced on the coil. The magnetic
latches in both the open and closed positions are included in the FE model. The design
of the high performance operating actuator for the vacuum interrupter investigates many
problems such as contact popping, welding, bounce and rebound.
Chapter 7 describes the design and manufacture of an actuator control circuit. A full-
bridge DC-DC converter is utilised because it allows bidirectional current flow. The
‘open’ and ‘close’ signals are produced by a trigger signal circuit which is used to
control the operation of MOSFETs using a drive circuit.
Chapter 8 presents the experimental test results of an MgB2 SFCL coil with an
integrated vacuum interrupter. The vacuum interrupter operating actuator is initially
tested at atmospheric pressure before placing it into a separate vacuum chamber.
Thereafter, the vacuum interrupter is connected into the SFCL system with a bypass
resistor in parallel. The experimental test results with and without the vacuum
interrupter are compared.
Chapter 9 summarises the conclusions drawn from the work reported in this thesis and
outlines additional work that could be conducted in the future.
Chapter 2: Literature Review
34
2 Literature Review
2.1 Superconductivity
Superconductivity is a phenomenon that occurs in certain materials below a critical
temperature, magnetic field strength and current density that exhibit zero DC resistivity.
In 1911, Heike Kamerlingh Onnes found that pure mercury exhibited zero resistance
when the temperature was reduced below 4K, which was the first discovery of
superconductivity [23]. A superconducting alloy composed of niobium and titanium
(NbTi) was discovered in the 1960’s and is still one of the popular superconducting
materials due to its high critical magnetic field strength and current density levels [18].
Before 1986, all the superconducting materials discovered exhibited superconductivity
below 25K and were termed low temperature superconductors (LTS). In 1986, Karl
Müller and Johannes Bednorz achieved superconductivity at around 30K in lanthanum
barium copper oxide (LBCO), which was the first high temperature superconductor
(HTS) [24]. Today, in high temperature superconductors the critical temperature of
YBCO is 90K whilst the critical temperature of BSCCO is 105K. This is important
because they exhibit superconductivity in liquid nitrogen which is cheaper and much
easier to produce than liquid helium.
2.1.1 Elementary properties of superconductors
2.1.1.1 Zero DC resistivity
The resistivity of a normal conductor and a typical superconductor is shown in Figure
2.1. In a normal conductor the current is carried by the electrons which are constantly
colliding with the atoms. During each collision some of the energy is dissipated and
converted into heat, which causes electrical resistivity in the conductor. When the
temperature decreases, the thermal vibrations of the atoms reduces and the electrons are
less frequently scattered by atoms, reducing the resistivity of the conductor. The
resistivity of a pure metal should approach zero as the temperature reduces towards 0K.
However, in practice a metal cannot be perfectly pure and will contain some impurities.
Chapter 2: Literature Review
35
The electrons are scattered by these impurities which is independent of temperature. As
a result, there is a residual resistivity at 0K for impure metals [23].
The temperature at which a superconductor exhibits zero resistivity is called its critical
temperature. Above the critical temperature, the resistivity of a superconductor is
similar to a normal metallic conductor. Zero resistivity below the critical temperature
can be explained by ‘cooper pairs’, which formed part of BCS theory developed by
John Bardeen, Leon Cooper and John Schrieffer [24]. A ‘cooper pair’ is two electrons
that are bound together at low temperature. The velocity of the ‘cooper pair’ is low at
low temperature and therefore there is insufficient energy to break the pairing. The DC
resistivity is zero.
Res
isti
vit
y
Impu
re m
etal
Pure m
etal
Res
isti
vit
y
Temperature0
Super
cond
ucto
r
Tc
Figure 2.1 Comparison of resistivity of normal conductors and a typical superconductor
[23]
2.1.1.2 Meissner effect
A magnetic field would be fully excluded from the inside of a superconductor if it is
placed into a magnetic field of relatively low strength below the critical magnetic field
strength, Hc. This perfect diamagnetism in a superconductor, as shown in Figure 2.2
(left), is known as the ‘Meissner effect’. As the magnetic field strength increases above
Hc, the superconducting state will break down, which means that the magnetic field can
fully penetrate the superconductor.
Generally speaking, superconductors can be divided into Type I and Type II
superconductors. Type I superconductors are usually made from a single metallic
element and exhibit perfect diamagnetism below their critical magnetic field strength.
Chapter 2: Literature Review
36
Type II superconductors are always alloy or compounds and they are able to remain in
the superconducting state at a higher applied magnetic field strength and current density
levels, which makes commercial applications possible.
Figure 2.2 Exclusion of the magnetic field in a superconductor [25]
Type I superconductors must expel all the applied magnetic field to stay in the
superconducting state, which means they can only be placed in a magnetic field of
relatively low strength. Figure 2.3 (left) shows that a Type I superconductor has only
two states: the superconducting state and normal state.
A Type I superconductor exhibits zero resistivity and zero internal magnetic flux
density when the applied external magnetic field strength, H, is below the critical
magnetic field strength, Hc. Once the applied magnetic field strength is higher than Hc, a
Type I superconductor would completely break down; the magnetic field can then
penetrate the superconductor and it would become a normal conductor. A Type I
superconductor has relatively low current carrying capacity because a transport current
will generate its own internal magnetic field and this seriously limits its application.
In Figure 2.3 (right), it can be seen that a Type II superconductor has three states: the
superconducting state, mixed state (vortex state) and normal state. When the magnetic
field strength is below the lower critical field strength, Hc1, it is in the superconducting
state with zero resistivity and internal magnetic field. Hc1 for a Type II superconductor
is equivalent to Hc for a Type I superconductor. When the applied magnetic field
strength increases between Hc1 and the upper magnetic field strength, Hc2, the
Chapter 2: Literature Review
37
superconductor is in the mixed state, which allows some penetration of magnetic field.
In the mixed state, magnetic field penetrates the superconductor via a series of magnetic
flux filaments that pass through the material. The filaments are surrounded by
circulating super currents which screen the magnetic field from the rest of the
superconductor. These filaments together with the super currents are termed vortices.
Once the magnetic field strength increases above Hc2, the superconductor becomes a
normal conductor.
Ap
pli
ed m
agn
etic
fie
ld
Figure 2.3 Applied magnetic field: Type I (left) and Type II (right) superconductor [25]
2.1.1.3 Critical temperature
The temperature profile of a Type II superconductor can be divided into three regions:
the normal region, transition region and superconducting region, as shown in Figure 2.4.
20ρ
Res
isti
vit
y
Figure 2.4 Typical temperature profile of a Type II superconductor [25]
Chapter 2: Literature Review
38
In the normal region, the resistivity of a superconductor is similar to a normal metallic
conductor. In the transition region, a superconductor transforms from the normal
conducting state to the superconducting state. Finally, in the superconducting state, a
superconductor exhibits zero DC resistivity. The normal state resistivity, ρo, is defined
as the resistivity at a temperature where the onset of superconductivity begins [25]. The
critical temperature is usually defined where the resistivity is half its normal state
resistivity (ρo/2).
2.1.1.4 Critical current density
The critical current density, Jc, is the third fundamental property of a superconductor
together with the critical magnetic field strength and critical temperature. The critical
current density is defined as the current density level which if exceeded will cause the
superconductor to lose its superconductivity. For a Type I superconductor, the critical
current density commonly can be taken as the level of current density which develops
the critical magnetic field strength on its surface.
A Type II superconductor behaves like a Type I superconductor for the magnetic field
strength produced by the transport current below the lower critical field strength, Hc1.
When the magnetic field strength produced by the transport current increases above Hc1,
a Type II superconductor allows magnetic field penetration via magnetic flux filaments.
As mentioned in section 2.1.1.2, these filaments are surrounded by circulating super
currents. Since the energy of the superconductor is not utilised in expelling magnetic
field, the critical current is generally much higher in a Type II superconductor than a
Type I. The increase of magnetic field penetration continues until the magnetic field
strength increases above Hc2 and the superconductor completely transitions to the
normal conducting state.
2.1.1.5 AC losses
Superconducting materials exhibit zero resistivity only when a DC current exists in it
because all the current is carried by super-electrons. However, it does not exhibit zero
resistivity to AC current because of alternating magnetic field and inertial mass of the
super-electrons. These AC losses translate into a resistivity. The AC resistivity may be
several orders of magnitude smaller than the normal state resistivity, ρo, but it is finite
and increases with frequency. For land-based power system frequency, AC losses are
Chapter 2: Literature Review
39
largely due to the hysteretic motion of the vortices, i.e. flux creep, hysteresis and related
magnetic phenomena in a superconductor [23, 25].
2.1.2 Materials
Amongst all the superconducting materials, BSCCO, YBCO and MgB2 are investigated
and utilised in SFCL applications. They are discussed in the following sections.
2.1.2.1 BSCCO
BSCCO with a critical temperature of 105K was discovered in 1988 by Maeda and his
colleagues at the National Research Institute for Metals in Japan [26]. BSCCO can be
manufactured into both bulk and tape form. BSCCO tape, which has been commercially
available since the late 1990’s, is also known as first generation (1G) wire.
BSCCO tapes are manufactured using a powder-in-tube (PIT) process. Precursor
BSCCO powders are filled into a silver or silver alloy tube which is drawn down to a
small diameter monocore wire. In order to manufacture multifilament BSCCO tape, a
specific number of monocore wires are stacked closely into a second silver or silver
alloy tube, which is again extruded down in diameter and rolled into a flat tape. The
tape is then reacted at high temperature to form a dense and aligned multifilamentary
tape and to increase the current carrying capacity. This design is relatively expensive
because a large amount of silver or silver alloy is used [27].
2.1.2.2 YBCO
YBCO was discovered to have a critical temperature of 90K in 1987. YBCO can be
manufactured in either thin film form or as a coated conductor. The thin film of YBCO
is created on a substrate using a pulsed laser deposition (PLD) technique [28]. However,
YBCO thin film has two problems which limit its commercial application. Firstly,
YBCO is very sensitive to grain boundaries. If the grain boundary angle is greater than
5°, the performance of YBCO thin film will be seriously affected. Secondly, YBCO thin
film is very brittle, which is difficult to wind into coils, for example.
An YBCO coated conductor, which comprises multiple coatings on a base substrate, is
also known as second generation (2G) wire. Companies that currently produce 1G HTS
wire are migrating to 2G HTS wire. SuperPower Inc. manufactures 2G wire using an
Chapter 2: Literature Review
40
ion beam assisted deposition (IBAD)/metal organic chemical vapour deposition
(MOCVD) method [29-31]. SuperPower can manufacture YBCO tape in excess of
500m with a critical current between 200 and 275A/cm-width [30]. The configuration of
the SuperPower 2G HTS wire type (SCS4050) is shown in Figure 2.5. The architecture
of 2G wire is totally different from 1G wire.
Figure 2.5 Configuration of SuperPower 2G HTS wire (wire type SCS4050) [30]
American Superconductor (AMSC) 2G wire is manufactured using a metal-organic
deposition (MOD)/rolling assisted biaxial textured substrates (RABiTS) method [9, 32].
This process produces a 4cm wide tape that is subsequently cut to a width of typically
4.4mm or 4.8mm. Currently AMSC manufactures YBCO tape in nominal lengths of 80
to 100m, with a critical current of 250A/cm-width in self-field at 77K [33]. AMSC
states that the 2G HTS wire they manufacture can carry over 100 times the electrical
current of copper wire of the same size, which significantly increases the current
capacity of power cables, for example [34].
Companies currently manufacturing HTS wire include Superconductor Technologies
Inc., Bruker Energy & Supercon Technologies, Inc., Zenergy Power plc., Nexans
Superconductors, Fujikura, Sumitomo, Furukawa and Showa.
Chapter 2: Literature Review
41
YBCO tape provides improved performance in a magnetic field and improved
mechanical properties with potentially lower cost through using cheaper raw materials.
Additionally, the large surface area of the tapes may enable them to cool down faster
after a quench (transition to the normal state). The manufacturing process however is
extremely complicated and this may make it difficult to be commercially competitive.
The tapes also have a low resistivity when quenched due to the presence of a copper
stabiliser, which means that long lengths are required to develop the desired resistance
in FCL applications. Furthermore, a tape form superconductor makes them less
attractive for winding into coils and may lead to larger AC losses than round wires.
2.1.2.3 Magnesium diboride
The discovery of MgB2 which exhibits superconductivity below 39K has attracted great
research interest when it was discovered in 2001 [35]. MgB2 can be made into round or
square wire form and also flat tape. The wire form can be chosen to match the
requirement of a specific application, making coil manufacture easier.
Companies manufacturing MgB2 wires include Columbus Superconductor, Hyper Tech
Research, Inc. and Hitachi Research Laboratory. Columbus Superconductor produces
MgB2 wires using an ex-situ PIT process, which is a similar manufacturing process to
1G wire [36]. The mixture of the magnesium and boron powder is heat treated at 900°C
for 1 hour to form MgB2 powder which is then filled into a metal tube. The tube is
reduced in diameter and sintered between 800 and 1000°C. Columbus Superconductor
is manufacturing MgB2 wires for tests in magnetic resonance imaging (MRI) and FCL
applications.
Hyper Tech developed and patented the continuous tube forming and filling (CTFF)
process [37]. A metal strip is formed into a U-shape and filled continuously with a
mixture of magnesium and boron powder. The U-shape tube is then closed and reduced
in diameter by wire drawing. The wire is then heated to form MgB2 superconductor:
called an in-situ process [16, 36-38]. The advantage of the CTFF process is that it can
make continuous long lengths of wire compared to the PIT process, which requires
larger and larger tubes to obtain longer lengths of wire. Hyper Tech produces a
Chapter 2: Literature Review
42
commercial multifilament wire consisting of 18 MgB2 filaments with a niobium barrier,
an inner copper monofilament sheath and a monel outer sheath [16].
The advantages of MgB2 wire are the use of cheaper raw materials compared with
BSCCO, a simpler manufacturing process compared with YBCO and reasonable
cooling costs compared with low temperature superconductors. In addition, MgB2 wires
can be manufactured with a variety of resistive sheaths. For a fault current limiter
application, for example, a short length of MgB2 wire can be used to insert a desired
resistance into the circuit by using a high resistivity sheath material such as stainless
steel. A large amount of work has been carried out on the fault current-limiting
properties of various MgB2 wires. All of these wire samples have demonstrated good
current-limiting properties with a fast transition from the superconducting to the normal
state. MgB2 wire therefore has become one of the strong candidates for the development
of SFCLs [12, 13, 15, 39].
2.1.3 Summary
The basic characteristics of superconductors including zero DC resistivity, critical
magnetic field strength, critical temperature, critical current density and AC losses have
been explained. Three superconducting materials namely BSCCO, YBCO and MgB2 are
commonly used for SFCL applications. Their raw materials, manufacturing processes
and properties have been summarised.
2.2 Fault current limitation
2.2.1 Introduction
Many factors can cause a fault in a power system. The fault current level can be
relatively large, which may damage equipment in the power system and even cause
permanent failure. Power systems have to be designed to withstand mechanical and
thermal stresses during a fault. Power system protection devices detect fault conditions
and operate circuit breakers and other devices to limit the damage. Today, fault current
levels in land-based distribution systems are of increasing concern because they are
generally rising due to the increasing capacity of connected distributed generation.
Increasing fault current levels will require expensive network investment in upgrading
Chapter 2: Literature Review
43
equipment such as circuit breakers and transformers. There is a growing need therefore
for fault current limiting devices embedded into electrical networks to avoid a large
scale and expensive upgrade of existing switchgear. FCLs are expected to reduce fault
current levels without adding additional impedance during normal operation. The
capital cost of purchasing and installing FCLs must be less than the cost of upgrading
the existing equipment before they can be attractive for commercial applications.
2.2.2 Operating principles of fault current limitation
After a short circuit occurs in a power network, the fault current will increase rapidly.
The rate of current rise depends on the source voltage, source impedance and fault
phase angle. A typical prospective short circuit current is shown in Figure 2.6.
Prospective fault current
Limited fault current
Nominal current
Time
Current
Figure 2.6 Typical fault current waveforms with and without fault current limiting
The fault current would eventually be interrupted by a conventional circuit breaker. If
the first peak of the fault current is higher than the rating of the device in the power
network, it is possible that damage may occur. The simplest way to reduce the fault
current is to increase the source impedance; however, this ultimately results in
additional voltage drop, reactive power and potentially a high transient recovery voltage
(TRV).
A fault current limiter must be inserted directly into the system in order to reduce the
first several peaks of fault current level under this situation. An ideal limited current
Chapter 2: Literature Review
44
waveform, as shown in Figure 2.6, is initially non-sinusoidal because the network
parameter is changed by inserting non-linear impedance.
The requirements for FCLs are listed as follows [2, 3]:
• Minimise impedance during the normal state. The resistance of the FCL during
normal state would produce extra heating and losses whilst the inductance would
induce voltage drop.
• Fast response. The FCL is able to detect the fault and reduce the fault current
below the required level before the first fault current peak.
• Quick and automatic recovery. It is desirable to reclose onto the power system as
soon as possible after a fault is cleared.
• Fail safe. The FCL would still limit the fault current even if it fails.
• Compact structure, light weight and low cost.
2.2.3 Application of fault current limiters
Possible applications for FCLs in power transmission and power distribution networks
are discussed in the following sections.
2.2.3.1 Transmission level voltage
Power transmission is a high voltage application which operates above 72kV. At the
transmission level voltage, FCLs are desirable for:
• Sub grid coupling. The distributed grids are often divided into sub grids to
manage the short circuit current. They are supplied by higher voltage levels via
separate transformers. These sub grids can be coupled together with FCLs,
providing higher power quality, fewer losses and less voltage drop.
• Busbar coupling. It is common to reduce the fault current by splitting of the
busbar. However, busbars coupled with FCLs are an excellent solution to reduce
fault currents to a reasonable level whilst inserting negligible impedance under
normal conditions.
Chapter 2: Literature Review
45
2.2.3.2 Distribution level voltage
Power distribution is a medium voltage level application with voltage ratings from
approximately 6kV to 72kV. FCLs can be used in the following locations:
• Busbar coupling and transformer feeder. Busbars coupled with FCLs would
increase the short circuit capacity. An FCL also can be placed in each
transformer feeder to increase the short circuit capacity so that the transformer
can be designed with lower impedance.
• Coupling of dispersed generation. With the development of wind and solar
power, there are increased numbers of connections of distributed generation.
They can be connected to a higher voltage grid via separate transformers if there
is no extra short circuit capacity. However, an FCL can be used between a
distribution generation system and a distribution grid to avoid the need for an
expensive transformer.
• Generator feeder. Increased generation capacity of generators would increase the
short circuit current; an FCL in a generator feeder location would be more
practical instead of upgrading an old substation.
• Power plant auxiliaries. These auxiliaries are close to power plants; therefore, it
is possible to have a high short circuit current. Installing FCLs will reduce the
rating of switchgear and the cost in these locations.
2.2.4 Conventional methods
There are two general categories of techniques to reduce short circuit currents:
permanent impedance increase and condition based impedance increase [2, 3].
The permanent impedance increase technique increases the impedance for both normal
and fault conditions, which includes splitting into sub grids, splitting of busbar,
increasing the voltage level, introducing high impedance transformers and current
limiting reactors. The power grid is split into small sub grids to reduce the power
capacity of each sub grid. The potential fault current level therefore would be reduced.
Busbars can be split into substations to reduce the fault current by reducing the
interconnections between power networks. However, these two methods may reduce the
stability and reliability of power networks. Increasing the voltage level will reduce the
Chapter 2: Literature Review
46
current that flows in the network for a given power capacity. However, this would lead
to increased system costs due to high voltage operation. High impedance transformers
and current limiting reactors are introduced to limit the fault currents by adding extra
impedance into the system. However, the inserted impedance will increase the voltage
drop across these devices, which will reduce the power quality of the system. All these
methods are not ideal since they reduce the performance of the system.
The condition based impedance increase technique only inserts impedance into the
network when a fault occurs. There are three traditional ways to achieve this: FCL
circuit breakers, fuses and Is-limiters. FCL circuit breakers made of baffle plates in
series are hard to design above 1kV because the arc voltage drop between the two plates
is only tens of volts [40]. Fuses and Is-limiters are widely used up to medium voltage
levels of 36kV. However, they have to be replaced manually after each fault operation.
Compared with these traditional methods, the new concepts of using superconductors
and/or semiconductors have the advantages of fast response time and automatic
recovery. They can also be designed for higher voltage levels.
2.2.5 Non-superconducting fault current limiters
2.2.5.1 Solid-state FCL
Solid-state FCLs perform fault current limiting by controlling semiconductors. The
schematic circuit of a solid-state FCL using a diode bridge circuit and an integrated gate
commutated thyristor (IGCT) is shown in Figure 2.7 [2, 41]. The operating principle is
as follows: during normal operation, AC current is commutated by the diode rectifier
and flows through the IGCT in one direction. Once a fault occurs, the IGCT usually
turns off in less than 1ms and the current is automatically diverted into a limiting
resistor. This topology only uses one turn-off IGCT. However, the current flows
through three semiconductors (two diodes and one IGCT) in series during normal
operation which causes on-state losses in the semiconductors.
A solid-state FCL without a turn-off device is shown in Figure 2.8 [2, 42]. During
normal operation, current flows through the main thyristors T1 and T2 in turn. When a
fault occurs, auxiliary thyristors T1a and T2a are usually triggered within 1ms, which
Chapter 2: Literature Review
47
results in the pre-charged capacitors C1 and C2 starting to discharge. This discharge
current reduces the current through the main thyristors T1 and T2, and when the current
reduces to zero they are turned off. The fault current is then commutated into C1 and C2
and eventually into the limiting resistors R1 and R2 after C1 and C2 are fully
discharged. Thyristors T1a and T2a are turned off at the next natural current zero-
crossing. This design uses cheap thyristors instead of an IGCT and has low on-state
losses compared with the previous topology because current only flows through one
thyristor during normal operation. However, capacitors C1 and C2 must be sufficiently
large to give the main thyristors enough time to recover and regain their forward voltage
blocking ability.
Figure 2.7 Solid-state FCL with a turn-off IGCT
Vac
Load
Circuit
breaker
+
T2
T2a
C2 L2
D2 R2
+
C1
D1R1
L1 T1a
T1
Figure 2.8 Solid-state FCL with thyristors
Chapter 2: Literature Review
48
2.2.5.2 Hybrid switching FCL
A hybrid switching FCL combines fast-acting mechanical switches with semiconductors
to reduce the number of semiconductor devices [2, 43], and is shown in Figure 2.9.
Figure 2.9 Hybrid switching FCL
During normal operation, all switches are closed and the current flows through the main
contact S1 because it has the lowest impedance with negligible losses. In the event of a
fault, S1 is opened (within 0.15ms) forming an arc across the contacts and the current is
diverted into IGCTs G1 and G2. G1 and G2 are turned off thereafter forcing the current
into a current limiting resistor R. S2 then opens without arcing to isolate G1 and G2
from the transient recovery voltage. This topology eliminates the on-state losses of the
semiconductors by placing a mechanical switch S1 in parallel. However, this system is
very complex and it is difficult to scale up to transmission level voltages.
2.2.5.3 Thyristor-controlled resonant FCL
A thyristor-controlled resonant FCL inserts a resonant inductor and capacitor or an
inductor to limit the fault current. Figure 2.10 shows the schematic circuit of a thyristor-
controlled resonant FCL [44, 45]. During normal operation, thyristors T1 and T2 are not
conducting and capacitor C provides series compensation in the power network. When a
fault occurs, thyristors T1 and T2 are triggered and inductor L is inserted into the
network in parallel with the capacitor. The parallel resonant LC circuit reduces the fault
current. This circuit may cause transient overvoltage, which is one of the concerns
relating to power quality.
Chapter 2: Literature Review
49
Figure 2.10 Thyristor-controlled resonant FCL
A series-parallel resonant FCL is one of the modifications of the resonant FCL, and is
shown in Figure 2.11 [44]. Inductor L1 and capacitor C are tuned to resonate at the
supply frequency. During normal operation, thyristors T1 and T2 are in the off state and
the voltage drop across the FCL is negligible. When a fault occurs, thyristors T1 and T2
are fired and the parallel resonant circuit (inductor L2 and capacitor C) in series with
inductor L1 will reduce the fault current. This circuit performs differently from the
original resonant FCL because it is initially a serial circuit.
Figure 2.11 Thyristor-controlled series-parallel resonant FCL
A series resonant FCL shown in Figure 2.12 is another modification of the resonant
FCL [44, 46]. Inductor L and capacitor C are tuned to resonate at the supply frequency.
During normal operation, the thyristors are not conducting and the voltage drop across
the FCL is negligible. In the event of a fault, the thyristors are turned on, which short-
circuits the capacitor C, leaving the series inductor L to limit the fault current. These
two modifications of the resonant FCL demonstrate improved power quality and
removal of high frequency oscillations from the bus voltage [44].
Chapter 2: Literature Review
50
Vac
C
T2
T1Circuit
breakerL
Load
Figure 2.12 Thyristor-controlled series resonant FCL
There are other non-superconducting FCLs such as polymer PTC resistor FCLs [47, 48]
and magnetic FCLs [49], but they will not be discussed here.
2.2.6 Superconducting fault current limiters
An SFCL is able to operate closer to an ideal FCL due to its non-linear characteristics.
Many different structures of SFCLs have been proposed and investigated after the
discovery of HTS. The various types are discussed in the following sections.
2.2.6.1 Resistive SFCL [2]
A resistive SFCL is the simplest and most compact SFCL design, which directly utilises
the characteristics of superconductors. The schematic circuit of a resistive SFCL is
shown in Figure 2.13. A resistor or inductor is normally placed in parallel with the
superconducting element to avoid overvoltage if the resistance of the superconductor
increases too rapidly after a fault occurs.
During normal operation, the superconducting element carries all the current and has
negligible resistance. Once a fault occurs, the current through the superconductor
increases quickly. When the current density in the superconductor exceeds the critical
current density level, the superconductor quenches and develops resistance, which limits
the current. The resistance of the superconductor increases quickly and therefore a high
percentage of the fault current is diverted into the parallel resistor or inductor, which
helps to limit the fault current. The fault will be completely cleared by the circuit
breaker usually within 200ms.
Chapter 2: Literature Review
51
Advantages:
• Compact structure, simple design and light weight.
• The superconductor automatically quenches after a fault occurs and a trigger
mechanism is therefore not essential.
• It is intrinsically fail safe. The superconductor will burn out if it fails.
Disadvantages:
• AC losses occur in the superconductor during normal operation and Joule losses
are dissipated in the superconductor after it quenches.
• Hot spot problems. It is impossible to guarantee all parts of the superconductor
quench at the same time. The weakest points will start to quench first and may
develop into hot spots.
• Long recovery time. Due to the heat dissipated and temperature rise of the
superconductor, it may take several seconds to several minutes to recover.
• Current leads are needed between external ambient conditions and the cryogenic
interface, which produce extra losses and may introduce a thermal insulation
problem.
Figure 2.13 Resistive SFCL [2]
ACCEL/Nexans lead a project called CURL 10/110 to develop resistive SFCLs using
BSCCO 2212 bulk [4, 50, 51]. A BSCCO 2212 bifilar coil with a bypass metallic
conductor was manufactured using a melt, cast and processed (MCP) technique [50]. In
Chapter 2: Literature Review
52
2004, Nexans installed a CURL 10 resistive SFCL in the grid for RWE in Germany for
testing and evaluation [51]. This system demonstrated effective current limitation
during the period of field testing [4]. From 2005, Nexans started to design a resistive
SFCL for high voltage power system applications. A concept design of a single phase
110kV/1850A resistive SFCL using BSCCO 2212 tubes with magnetic field assisted
quench was carried out [52].
Thereafter, Nexans designed and manufactured three resistive SFCLs for commercial
applications. Nexans manufactured a 12kV/2MVA resistive SFCL and Applied
Superconductivity Limited (ASL) designed the closed-loop cryogenic system. This
device was installed at Bamber Bridge, U.K. in 2009, which was the first commercial
system worldwide [53, 54]. In the same year, a 12kV/16MVA resistive SFCL was
installed in the auxiliary power supply at Boxberg, Germany, which was the first SFCL
installed in a power station [53, 54]. Another 12kV/9MVA resistive SFCL is presently
undergoing installation in Ainsworth Lane, U.K. Nexans also started to develop the
SFCL with YBCO coated conductors under the financial support of a German
government project called ENSYSTROB. A resistive 12kV/800A SFCL with YBCO
tape was successfully tested at the high power test lab in Berlin in September 2011 and
has been subsequently installed at the Boxberg power plant for field tests [8, 53, 55]. As
part of an European Union (EU) program called ECCOFLOW, a resistive 24kV/1005A
SFCL, which is designed for two applications, is presently under development [53, 55].
SuperPower developed a matrix fault current limiter (MFCL) using MCP BSCCO 2212
tube supplied by Nexans for transmission level voltages of 138kV and above [11, 56]. A
trigger matrix is connected in series with a current-limiting matrix. The MFCL uses the
trigger matrix to produce a magnetic field to quench the current-limiting matrix when a
fault occurs, which prevents hot spot problems. After the prototype was successfully
tested, SuperPower started to investigate a pure resistive SFCL with a YBCO coated
conductor [57]. SuperPower demonstrated that it is possible to develop this for both
distribution and transmission level voltages using a modular design. For example, a
distribution level SFCL of 11-15kV/800-2000A needs three SFCL modules, whilst a
transmission level SFCL of 138kV/1700A needs 14 SFCL modules [58].
Chapter 2: Literature Review
53
CESI Ricerca in Italy has manufactured and tested a prototype 500kVA resistive SFCL
using multifilamentary Ni-sheathed MgB2 tapes supplied by Columbus Superconductor.
The prototype showed good current-limiting characteristics with no degradation after
repeated fault tests [59]. In 2010, a 9kV/4MVA SFCL using BSCCO 2223 was
developed and installed at the outgoing feeders of the San Dionigi substation in north
Italy for field testing. The three-phase SFCL has demonstrated a successful short circuit
test [10]. A 9kV/15MVA SFCL for the incoming feeders using a YBCO coated
conductor is currently under development. The design and simulation results suggest
that it is feasible [60].
Siemens initially investigated a resistive SFCL using YBCO thin films [61]. In 2005,
Siemens and AMSC formed a partnership to investigate YBCO coated conductor for
resistive SFCLs [62]. A single phase 7.5kV/300A SFCL using AMSC 344S 2G HTS
wire was developed and successfully tested. This SFCL utilised a bifilar coils design to
cancel the magnetic field, which then reduced the inductance [9]. A project to develop a
138kV SFCL ‘SuperLimterTM’ for a demonstration project at Southern California
Edison (SCE) power network has been sponsored by the U.S. Department of Energy
(DoE) [1]. A hybrid resistive SFCL using modular superconducting elements has been
proposed. The superconducting modules are connected in series with a fast operating
switch and a reactor is connected in parallel with the modules and the switch [1, 20].
The superconducting modules reduce the fault current for the first three cycles. After
three cycles of fault current, the fast operating switch removes the superconducting
elements from the circuit and the fault current is limited by the reactor. This method
reduces the amount of heating in the superconducting modules [54]. A single-phase
prototype has been tested successfully; however, the three-phase system has been
cancelled due to lack of funding [55].
Tokyo Denki University and the Japanese National Institute of Advanced Industrial
Science and Technology have been investigating an SFCL using a vacuum interrupter
driven by an electromagnetic repulsion switch [63, 64]. A vacuum interrupter is
connected in series with the superconducting element and a coil is connected in parallel
with both of them. When a fault occurs, the superconducting element starts to quench
and the fault current transfers into the coil. The coil has two functions: firstly, it acts as
Chapter 2: Literature Review
54
a limiting coil; and secondly, it drives the electromagnetic repulsion plate (EPR) which
is fixed onto a moveable contact to open the vacuum interrupter quickly. This design
can reduce Joule losses dissipated in the superconducting element and hence reduce its
recovery time. A prototype SFCL using YBCO thin film was constructed and examined,
interrupting the fault current of the superconductor within a half-cycle after the
modification of the parallel coil [21]. The EPR would return to the closed position when
the fault current was completely interrupted by a back-up circuit breaker. An
electromagnetic repulsion switch for operating with the reclosing scheme, which
consists of an electromagnetic repulsion plate, cylinder, check-valve and intake vent,
was investigated. The reclosing time of the vacuum interrupter can be adjusted by the
intake vent. The recovery characteristics of AMSC 344S was examined and showed that
the recovery time increases as the voltage of the superconducting element increases.
The prototype with an electromagnetic repulsion switch demonstrated expected
adjustable reclosing behaviour [65, 66].
Hyundai and Yonsei University in Korea developed a 13.2kV/630A resistive SFCL
using AMSC 344S YBCO coated conductor. The superconducting coil effectively
limited the short circuit fault current without damage [67]. The Korea Electric Power
Research Institute (KEPRI) and LS Industrial Systems (LSIS) are leading a project to
develop a 22.9kV/630A hybrid resistive SFCL [68]. This hybrid resistive SFCL consists
of a superconducting coil, a fast operating switch, a driving coil and a current limiting
reactor [69]. In the event of a fault, the HTS becomes resistive and the current flows
through the driving coil, which produces a force to open the fast switch. The current is
subsequently diverted into the reactor to reduce the fault current. The hybrid resistive
SFCL reduces the amount of superconductors and hence the cost and volume of the
SFCL [70]. The SFCL was constructed and preliminary tested before placing it in the
Gochang Power Testing Centre of Korea Electric Power Corporation (KEPCO) grid in
November 2009 [70]. The SFCL is still under long-term operation and testing. The fault
and reclosing tests have been successfully conducted [71]. The protection parameters of
the network have been recalculated taking the SFCL into consideration. Short circuit
tests prove that an SFCL can be used in addition to the power system protection by the
adjustment of the parameters of the protection devices [72]. The final target is to
develop a 22.9kV/3kA hybrid resistive SFCL for general applications [73].
Chapter 2: Literature Review
55
2.2.6.2 Bridge type SFCL [2, 5, 74]
A bridge type SFCL is made up of a bridge rectifier, a DC voltage source and a limiting
coil, as shown in Figure 2.14. The level of DC current through the limiting coil, which
is supplied by the DC voltage source, is designed to be higher than the peak nominal
current of the power system. During normal operation, the DC voltage source keeps all
four diodes forward biased. The inductor therefore is bypassed from the power system.
When a fault occurs, if the AC current increases higher than the DC current, diodes D3
and D4 or D1 and D2 would be reverse biased in the positive or negative half-cycle.
The limiting inductor will therefore be connected into the system to limit the fault
current.
The inductor does not have to be made of superconducting material, but
superconducting material can be used to minimise the losses. In addition, during normal
conditions, the inductor only carries DC current, which makes a superconductor an ideal
choice. Thyristors can be used to replace the diodes, so it is possible for them to turn off
the current at the next current zero-crossing after a fault occurs.
Figure 2.14 Bridge type SFCL
Advantages:
• No AC losses in the superconducting coil because it is operating with DC
current.
• Fast recovery after the fault clears because the coil remains in the
superconducting state during the fault.
• The trigger current level can be adjusted by the DC current source.
Chapter 2: Literature Review
56
• Does not require a room temperature/cryogenic interface in the power line.
Disadvantages:
• AC losses in the semiconductors are relatively high.
• No fail safe mechanism. If one of the semiconductors fails and creates a short
circuit, the SFCL cannot limit the fault current.
The bridge type SFCL shown in Figure 2.14 was proposed by Boenig [5, 74]. The first
prototype 2.4kV/150A SFCL used BSSCO 2223 tape and was developed in 1993 [75].
After the success of the first prototype, another prototype 15kV/1.2kA SFCL was built
using thyristors instead of diodes. The thyristors can control the magnitude of the fault
current and interrupt the current completely. During high voltage testing in Southern
California Edison Centre Substation, a voltage insulation failure occurred in an auxiliary
piece of equipment and a three-phase test was therefore no longer possible [76].
The Chinese Institute of Electrical Engineering (CIEE) and the Chinese Academy of
Sciences (CAS) designed a 10.5kV/1.5kA bridge type SFCL using AMSC BSSCO 2223
tape [77, 78]. Four resistor/IGCT pairs were connected in series with a superconducting
coil instead of a traditional DC power supply for the bridge type SFCL. During normal
operation, the current flows through the bridge and the IGCT switches are turned on. In
the event of a fault, the first peak current is limited by the HTS coil. When the current in
the HTS coil reaches a pre-set level, the IGCTs are switched off, and the resistors are
inserted into the circuit. The fault current is then limited by the resistors and the HTS
coil [78]. After successful tests in the lab, the SFCL was installed in the Gaoxi
substation of Hunan in 2005 and subsequently demonstrated long-term reliable
operation in the real grid for one year. The SFCL reduced the fault current below 635A
successfully in a three-phase-to-ground short circuit test [78, 79].
2.2.6.3 DC biased iron core SFCL [2]
A DC biased iron core SFCL is shown in Figure 2.15. It consists of two iron coils,
which are driven into saturation by a DC supply. During normal operation, the iron
cores are fully saturated because the AC current is much lower than the DC current. The
Chapter 2: Literature Review
57
inductances of L1 and L2 are small because they are similar to air core inductors. Once
a fault occurs, the increased AC current will drive coil L1 or L2 out of saturation in a
positive or negative half-cycle and the operating region of the core returns back to the
high permeability region. This causes the inductances of the coils to increase, which
reduces the fault current level.
Advantages:
• Inherently fail safe.
• Fast recovery after the fault clears because the coil remains in the
superconducting state during the fault.
• No AC losses in the superconducting coil because it is operating with DC
current.
• Does not require a room temperature/cryogenic interface in the power line.
Disadvantages:
• The device is very bulky and heavy due to the iron core.
• AC losses in the primary windings and iron cores.
Figure 2.15 DC biased iron core SFCL
Zenergy Power has been investigating a DC biased iron core SFCL utilising BSCCO.
Under the support of the U.S. DoE, the first 15kV/1.2kA SFCL was installed and
successfully tested on the grid controlled by Southern California Edison since March
2009 [80]. The configuration of the iron core was assessed and several improvements
Chapter 2: Literature Review
58
were undertaken, which led to a lighter and more compact design [54]. After successful
testing of the demonstrator SFCL, Zenergy Power built a 12kV/1250A SFCL for ASL,
which is presently installed in a CE Electric U.K. substation near Scunthorpe [19, 55]. A
138kV/1.3kA SFCL using the same concept was under development and was expected
to be installed in AEP Tidd substation belonging to American Electric Power in early
2012 [54]. However, the project was terminated in September 2011 due to several
administrative crises [55].
Innopower designed and tested a single-phase and a three-phase 380V prototype DC
saturated iron core SFCL in 2004 and 2005 respectively [6]. Innopower then developed
a 35kV/90MVA SFCL using the same structure [81, 82]. The SFCL consists of six
rectangular cores and a central core, which are all placed in a single cryostat. During
normal operation, a DC magnetisation coil drives the iron core into saturation. When a
fault occurs, the current in the DC magnetisation coil is removed by a switch. The
energy in the iron core is dissipated in an energy release circuit and the iron core returns
to its high permeability operating region. The DC magnetisation coil may be subjected
to a high back electromotive force (emf) due to the opening of the switch; a protective
circuit is introduced therefore to suppress the induced voltage [83, 84]. The SFCL was
installed in the Puji substation operated by the Southern China power grid in 2007 for
live grid operation [82, 85, 86]. The SFCL demonstrated encouraging current-limiting
performance. The artificial imposed fault test in the live grid was in good agreement
with the design expectations [87, 88].
2.2.6.4 Shielded iron core SFCL [2]
Figure 2.16 shows the scheme of a shielded iron core SFCL, which is made up of a
primary winding around an iron core with a superconducting cylinder in between. This
SFCL is also called an inductive SFCL because its structure is similar to a transformer
with a short circuit secondary winding. During normal operation, the current in the
superconducting cylinder is lower than its critical current and it screens all the flux from
the iron core. The impedance of the device, which consists of the resistance of the
primary winding and the stray inductance, is very low. In the event of a fault, the
current in the superconducting cylinder exceeds the critical current and the cylinder
starts to develop a resistance. The magnetic flux penetrates into the iron core, so the
Chapter 2: Literature Review
59
inductance of the primary winding increases. The equivalent impedance of the device
becomes the inductance of the primary winding and the referred cylinder resistance to
the primary in parallel.
Superconducting cylinder
Primary winding Cryostat
Iron
core
Figure 2.16 Shielded iron core SFCL
Advantages:
• Fail safe. If the superconducting cylinder burns out, the magnetic flux will
penetrate into the iron core, which increases the inductance of the primary
winding to limit the fault current level.
• Does not require current leads for the room temperature/cryogenic interface.
Disadvantages:
• Long recovery time. Due to the heat dissipated and temperature rise of the
superconducting cylinder, it may take several seconds to several minutes to
recover.
• Large size and heavy weight, which limits its scalability.
• AC losses in the primary winding.
Chapter 2: Literature Review
60
2.2.6.5 Fault current controller SFCL [2]
A fault current controller SFCL consists of two anti-series connected thyristors, with
each thyristor connected with a superconducting coil inductor in parallel, as shown in
Figure 2.17.
Figure 2.17 Fault current controller SFCL
The thyristors are triggered at their respective peak load current and then constant DC
currents circulate through the parallel thyristors and inductors. The triggering signals are
removed after the thyristors are turned on. During normal operation, the AC current is
lower than the constant DC current, so that the thyristors are conducting and short-
circuit the inductors. When a fault occurs, the AC current rises above the constant DC
current. When the current is in the positive half-cycle, the current through thyristor T2
will go across zero and then T2 will turn off. Inductor L2 is inserted into the circuit to
reduce the fault current. Thyristor T1 will subsequently turn off at the negative half-
cycle and insert inductor L1 into the circuit.
Advantage:
• Fast recovery after the fault clears because the coil remains in the
superconducting state during the fault.
Disadvantages:
• AC losses in the semiconductors and superconducting coil.
• No fail safe mechanism. If one of the semiconductors fails and creates a short
circuit, the SFCL cannot limit the fault current.
Chapter 2: Literature Review
61
2.2.6.6 Flux-lock type SFCL[7, 89]
A flux-lock type SFCL consists of a flux-lock reactor with a superconducting coil L3
and a magnetic field coil. A schematic circuit of a flux-lock type SFCL is shown in
Figure 2.18. During normal operation, the superconducting element is in the
superconducting state, so the voltage across it is zero. The magnetic flux linkage
through the iron core is constant in a DC mode and therefore the voltages across these
three coils are zero, and the impedance of the SFCL is negligible. Once a fault occurs,
the superconducting element loses its superconductivity and develops a resistance,
which reduces the fault current. The magnetic flux then varies in the iron core and the
induced voltage across the coils changes. The current flows in the magnetic field coil
and the external magnetic field is applied to the superconducting element, which causes
the resistance of the superconducting element to increase faster and relatively evenly
along its length.
Figure 2.18 Flux-lock type SFCL
Advantages:
• During a fault condition, the magnetic field coil applies the magnetic field to the
superconducting element to make it quench equally, which prevents hot spot
problems in the superconducting element.
• The superconducting element is isolated from the power line.
Disadvantages:
• The device is very bulky and heavy due to the iron core.
Chapter 2: Literature Review
62
• AC losses in the three windings.
• Long recovery time. Due to the heat dissipated and temperature rise of the
superconductor, it may take several seconds to several minutes to recover.
Some modified versions of the flux-lock type SFCL have been investigated such as
using two co-axial air core coils as an alternate [90]. A flux-lock type SFCL with two
triggering current levels produces more effective current-limiting using a second
superconducting element in case the initial transient component of the fault current is
large [91].
A summary of the main SFCL projects which are believed to be currently active is
provided in Table 2.1.
2.2.7 Summary
As the capacity of power networks increases, fault current limitation devices are of great
interest. The operating principle and applications of fault current limitation devices were
briefly introduced. The conventional methods of fault current limitation were then
summarised. The topologies, operating principles and projects of non-superconducting
FCLs and SFCLs were discussed. The advantages and disadvantages of different
topologies were also summarised.
Chapter 2: Literature Review
63
Table 2.1 Summary of active SFCL projects
Country Lead Company Type Material Phase Test Rating Test Year Ref
Germany ACCEL/Nexans Resistive BSCCO 2212 bulk 3-phase 12kV, 600A 2004 [4]
Germany/U.K. Nexans/ASL Resistive BSCCO 2212 bulk 3-phase 12kV, 100A 2009 [53]
Germany Nexans Resistive BSCCO 2212 bulk 3-phase 12kV, 800A 2009 [53]
Germany/U.K. Nexans/ASL Resistive BSCCO 2212 bulk 3-phase 12kV, 400A 2011 [53]
Germany Nexans Resistive YBCO tape 3-phase 12kV, 800A 2011 [53]
EU Nexans Resistive YBCO tape 3-phase 24kV, 1005A 2012 [53]
Germany/U.S.A. Siemens/AMSC Resistive YBCO tape 1-phase 7.5kV, 300A 2007 [9]
Germany/U.S.A. Siemens/AMSC Hybrid Resistive YBCO tape 1-phase 138kV, 1.2kA 2011 [55]
Italy CESI Ricerca Resistive BSCCO 2223 tape 3-phase 3.2kV, 220A 2005 [60]
Italy CESI Ricerca Resistive MgB2 tape 1-phase 397V, 96A 2006 [59]
Italy ERSE Resistive BSCCO 2223 tape 3-phase 9kV, 250A 2010 [10]
Italy ERSE Resistive YBCO tape 3-phase 9kV, 1kA 2012 [10]
U.S.A. SuperPower Resistive BSCCO 2212 bulk 1-phase 8.6kV, 800A 2004 [11]
U.S.A. Zenergy DC biased iron core BSCCO 2223 tape 3-phase 12kV, 1.2kA 2009 [80]
U.S.A. Zenergy DC biased iron core BSCCO 2223 tape 3-phase 12kV, 1.2kA 2011 [19]
China CAS Diode bridge BSCCO 2223 tape 3-phase 10.5kV, 1.5kA 2005 [78]
China Innopower DC biased iron core BSCCO 2223 tape 3-phase 35kV, 90MVA 2007 [92]
China Innopower DC biased iron core BSCCO 2223 tape 3-phase 220kV, 300MVA 2010 [10]
Korea Hyundai Resistive YBCO tape 1-phase 13.2kV, 630A 2007 [67]
Korea KEPRI/LSIS Hybrid resistive YBCO thin film 3-phase 22.9kV, 630A 2007 [69]
Japan Toshiba Resistive YBCO tape 3-phase 6.6kV, 72A 2008 [93]
Chapter 2: Literature Review
64
2.3 Vacuum interrupter
2.3.1 Introduction
Air, oil, SF6 and vacuum are commonly used in circuit breakers. For medium voltage
applications, SF6 and vacuum have already replaced the previous air and oil solutions
and have become the most widely used types of circuit breaker. Vacuum as an
insulation medium for circuit breakers was first reported in 1926 by Royal Sorensen and
Mendenhall [94]. In 1962, General Electric developed the first power vacuum
interrupter capable of interrupting 12.5kA at 15.5kV. In the U.K., a vacuum interrupter
capable of interrupting 15.3kA at 13.2kV was tested in 1967. In the following year,
33kV vacuum interrupters were manufactured in both the U.S.A. and the U.K. Currently,
ABB commercially supplies a vacuum circuit breaker, the VD4, rated up to 12kV/4kA.
The rated breaking current is up to 63kA [95]. Schneider produces the Powersub™
vacuum substation circuit breaker which is rated at 15-38kV/600-4000A. This vacuum
interrupter can break a current up to 40kA and withstand a basic impulse level (BIL) of
110-150kV [96]. A 126kV single break vacuum interrupter with an axial magnetic field
contact is presently under research in China [97-99].
Vacuum interrupters are capable of carrying and breaking currents under both normal
and fault conditions. They are widely used in medium voltage applications because of
the following advantages [100]:
• Maintenance free design and high reliability.
• Self-contained and environment compatible. There is no risk of explosion or fire
and external effects during breaking.
• Compact and small volume due to high dielectric strength of vacuum.
• Comparatively low erosion of contacts leads to long electrical life span.
• Short contact stroke, which allows simple operating actuator and requires
relatively low mechanical energy to operate.
• Is capable of breaking a fault current with a high rate of change of current with
time (di/dt) and withstand transient recovery voltage because of the fast
dielectric recovery capability.
Chapter 2: Literature Review
65
2.3.2 Vacuum interrupter structure
The structure of a vacuum interrupter is shown in Figure 2.19. It is usually made up of a
pair of contacts, bellows, shields and a vacuum enclosure [94, 101]. When the current
rating of the vacuum interrupter is lower than a few kA, cylindrical butt contacts are
commonly used [94, 101]. For higher current ratings, spiral contacts or contrate contacts,
which can force the arc to rotate, are usually preferred. Most vacuum interrupter
contacts are made of copper chrome (CuCr).
Metal bellows are usually used in order to allow the movement of one of the contacts
inside the vacuum. One end of the bellows is connected to one end plate of the vacuum
interrupter. The other end is attached to the shank of the moving contact.
Shields are placed around the contacts and over the bellows. They are used to prevent
the metal vapour of the arc from condensing onto the vacuum enclosure and bellows.
Various materials such as stainless steel, nickel, nickel/cobalt iron alloys and copper
have been used for the shields.
Figure 2.19 Typical structure of a vacuum interrupter [101]
Chapter 2: Literature Review
66
The vacuum enclosure includes two end plates and insulation between them. The end
plates are usually made of stainless steel. The insulation between the two end plates in
early manufactured vacuum interrupters was made of glass. Nowadays, ceramic
materials such as alumina are generally the most commonly used material for insulation.
2.3.3 Vacuum arc [94]
When the current carrying contacts separate, a metal vapour arc is developed in the gap
between the two contacts. The diffuse mode and the constricted mode are the two main
forms of the vacuum arc [94].
When the current is relatively low, from a few amps to a few kA, the vacuum arc is in
the diffuse mode. A cathode spot is a small area that emits vaporised cathode contact
material. One or several spots, separated from each other, are spread on the cathode
contact. The number of cathode spots is determined by the magnitude of the current and
the contact materials. In a half-cycle of alternating current, the number of cathode spots
increases when the current rises and then decreases when the current reduces. When the
current approaches zero, only one cathode spot is left. The cathode spots are a conical
shape with the peak at the cathode and in continuous motion across the contact surface.
The cathode spots carry high current in small surface areas, which produce extremely
high electric fields and local temperatures and hence lead to field effect emission. The
gap between the contacts is filled with a neutral plasma column, which is made up of
metal vapour electrons and ions emitted from the cathode contact. Most of the metal
vapour is condensed on the anode, except some which escapes from the contact and
freezes on the shields [102].
The vacuum arc voltage is low in the diffuse mode, which is concentrated across the
cathode region. The arc voltage increases slowly as the current increases. The voltage
drop of the plasma column increases with the length of the gap and the current level.
The voltage drop of the plasma column is however much lower than the cathode voltage
drop, and therefore the length of the gap has little influence on the arc voltage. The
vacuum arc has a positive resistance characteristic, which allows several cathode spots
exist at the same time. The erosion of contacts is low in the diffuse mode because most
Chapter 2: Literature Review
67
of the emission electrons and ions are collected and recombined on the anode. In
addition, the polarity of the cathode and anode contacts is alternating due to AC current.
As the current increases, the energy dissipated in the vacuum interrupter increases
because both the current and arc voltage are increasing. This increases the heat flux to
the contacts and the shields. Once an anode spot establishes due to the focusing of the
current, the vacuum arc is transitioned into the constricted mode. The current level at
which the transition occurs depends on the material and size of the contacts. During the
transition from the diffuse mode to the constricted mode, the arc voltage increases with
a significant noise component. After the anode spot develops, the arc voltage drops and
becomes stable. In the constricted mode, the current is concentrated in a limited surface
area on the contacts, which causes localised overheating and considerable vaporisation.
The plasma in the gap then contracts into one column. The anode spot is playing an
active role in the discharge. The high metal vapour density causes the pressure of the
gap to be close to atmospheric pressure and therefore the vacuum arc in the constricted
mode shows similar characteristics to a high pressure arc. The constricted mode of the
vacuum arc produces considerable erosion of both contacts; one or two orders greater
than the diffuse arc.
2.3.4 Vacuum breakdown
Breakdown in a vacuum interrupter often occurs following arc interruption. It is hoped
that the dielectric strength of the contact gap is recovering, so that it can hold off the
transient recovery voltage, which the circuit is trying to apply across the contact gap. If
there is no electric field in the gap, the atoms or molecules of a gas are in random
motion and are frequently colliding with each other. These collisions are elastic because
the energy is preserved during the collisions. A small number of charged particles are in
random motion and they may recombine when they collide. When an electric field is
applied, the charged particles receive a force from the electric field and they accelerate,
so that their kinetic energy increases. The average distance between collisions is defined
as the mean free path (MFP). As the strength of the electric field increases, the charged
particles obtain more and more energy. It is possible that with the long mean free path
the charged particles accumulate enough energy and will ionise atoms when they
collide. An electron will be removed from the atom during a collision, leaving a positive
Chapter 2: Literature Review
68
ion behind. The electrons produced from the collisions will accelerate and ionise more
atoms, which can produce a large amount of electrons. This process is referred to as
Townsend avalanche, which may eventually lead to breakdown in gases. There are two
essential conditions for breakdown in gases: firstly, there needs to be enough atoms in
the gap to be ionised; and secondly, the mean free path needs to be long enough for the
charged particles to accumulate enough energy for ionisation.
In practical applications, circuit breakers are designed to circumvent one of the
conditions necessary to increase the breakdown voltage. For example, an SF6 interrupter
operates on the right side (high pd) of Paschen’s curve [103], where p is the pressure in
the interrupter and d is the distance between the two electrodes. The pressure of the SF6
gas is high and hence a large number of particles are in constant collision with each
other. They cannot gain sufficient energy to ionise atoms during collision however
because of the short mean free path. Therefore, the higher the pressure in the SF6
interrupter, the higher the breakdown voltage.
The vacuum interrupter is another example which operates on the left side (low pd) of
Paschen’s curve. The population of charged particles and atoms in the gap are very
limited. The charged particles have the long mean free path and can acquire enough
energy, but they may not collide with other atoms before they condense on the contacts.
Paschen’s curve shows that the dielectric withstand increases rapidly when the pd is
lower than 10-2Pa·m. The pressure of the vacuum interrupter is usually lower than
10-3mbar and the distance between the contacts can be designed from 1cm to 10cm.
Long path breakdown is possible for the vacuum interrupter because of the increase in
pd. It is necessary therefore to monitor the vacuum level in the vacuum interrupter in
order to ensure its reliability. Permanent monitoring of the vacuum level is not possible
because it has to be removed to do so.
The contact surface conditions of a vacuum interrupter can seriously affect the dielectric
withstand of the vacuum interrupter. The contacts of the vacuum interrupter experience
welding when closing and suffer from arcing when pulled apart whilst conducting
current. Most of the metal vapour and particles condense on the contacts when operating.
Chapter 2: Literature Review
69
The surface conditions of the contacts therefore change after operation, which may
significantly affect the ability to support a high dielectric strength between the contacts.
2.3.5 Current breaking in a vacuum
A vacuum interrupter is normally carrying current in the closed position during most of
its life time. It is designed to interrupt the current on command. The current interruption
is in fact rapid removal of the plasma which is conducting the current in the arc. The
plasma consists of metal vapour, metal ions, electrons, gas molecules, metal droplets
and particles. Most of the metal vapour are ionised and when they arrive at the anode,
they recombine with electrons and become metal atoms. The cooling of cathode spots,
which depends on the thermal conductivity of the contact material and atom evaporation,
is very rapid. Un-ionised metal vapour is not directly influenced by the electric field and
will eventually condense onto a cool surface. The production of the metal vapour relies
on the rate of erosion from the cathode. It has been proved that the metal vapour exists
for about 1µs in a 1cm contact gap [94]. The rate of current decrease at supply
frequency is much slower than the extinction of the metal vapour. A vacuum interrupter
therefore can easily break the arc in the diffuse mode and rapidly recover its high
voltage blocking capacity.
In the diffuse mode, when a vacuum interrupter breaks current at supply frequency, the
number of cathode spots decreases as the current decreases. When the current
approaches zero, the energy is not enough to maintain a high temperature at the root of
the arc and the arc becomes unstable. The last spot extinguishes and the current is
interrupted abruptly, which leads to a current chopping phenomena. Current chopping
may cause overvoltage because the current in the inductive load cannot stop suddenly
and there is no free-wheel path through a vacuum interrupter for the current after
chopping. The current is then commutated into the capacitance of the network, which
causes overvoltage.
When the arc is in the constricted mode, the current interruption is changed by the
anode spot, which continues to emit metal vapour after a current zero. In addition, the
temperature of the anode spot is relatively high and the positive ions attack it. It is
possible to build the cathode spot on the previous anode when the current alternates to
Chapter 2: Literature Review
70
the next half-cycle. The removal of the constricted arc is much slower than the diffuse
arc; the rate of change of voltage with time (dv/dt) capability of a vacuum interrupter
therefore is much lower in the constricted mode.
In conclusion, the diffuse arc is easier to interrupt than the constricted arc, which
suggests that it is better to maintain the arc in the diffuse mode whenever possible. The
practical methods to keep the arc in the diffuse mode are described in the next section.
2.3.6 Practical design for high current levels
The vacuum arc can be affected by the applied magnetic field. In practical applications,
both radial magnetic field (RMF) and axial magnetic field (AMF) technologies are in
use to prevent the arc from passing into the constricted mode and improve the current
breaking capacity of a vacuum interrupter [101, 104].
A radial magnetic field can cause a rapid rotational movement of the arc, so that the
energy is distributed onto a large surface area of the contact. The radial magnetic field is
produced by the path imposed on the current in the contacts. Spiral contacts or contrate
contacts are widely used to rotate the arc by the self magnetic field of the current. The
root of the arc does not stay in the same location for a long period, which eventually
postpones the appearance of an anode spot. In addition, the vapour is equally spread
throughout the volume of the vacuum interrupter when it starts to recover. Furthermore,
the heat and vapour are uniformly distributed on the shields due to rotation which
ensures the shields do not have hot spots and the vapour can continue to condense on
the shields.
An axial magnetic field delays the formation of an anode spot because of its confining
effect, which helps in maintaining the arc in the diffuse mode. An AMF significantly
reduces the arc voltage and hence the power dissipated in the vacuum interrupter. The
application of an AMF is very important from a practical point of view. Firstly, it can
change the appearance of the arc; the plasma from the cathode spots becomes more
columnar and distinct when an AMF is applied. Secondly, the application of an AMF
reduces the arc voltage sharply, which would delay the formation of the anode spot.
There are several methods used to produce the axial magnetic field using the current
Chapter 2: Literature Review
71
being interrupted such as integrating the coils behind the contacts or using an external
coil to surround the inter contact zone [101].
Both RMF and AMF technologies can offer a high current breaking capacity. These two
techniques have their own advantages and disadvantages. RMF can carry high
continuous current because of its low contact resistance, whilst AMF offers high voltage
rating and contact endurance [104].
2.3.7 Main application fields
Medium voltage levels are the primary application field for vacuum interrupters. The
vacuum interrupter and SF6 interrupter dominates the medium voltage level market. The
advantages of a vacuum interrupter over an SF6 interrupter are summarised in [101]:
• A vacuum interrupter is an enclosed circuit breaker, which adopts maintenance
free design with high electrical endurance.
• The dielectric recovery of a vacuum interrupter is faster than an SF6 interrupter,
which is especially suitable for a severe initial transient recovery voltage.
• A vacuum interrupter requires less operating energy than an SF6 interrupter.
A vacuum interrupter also has several disadvantages compared to an SF6 interrupter:
• A vacuum interrupter uses butt contacts, which need high contact pressure to
minimise contact resistance and avoid separation of the contacts.
• The butt contacts of a vacuum interrupter are usually made of CuCr which has a
higher resistance than the silver-plated multiple contacts of an SF6 interrupter.
• A vacuum interrupter is restricted by overheating.
In low voltage level applications, a vacuum interrupter cannot compete with an air
circuit breaker. The disadvantages which limit the application of a vacuum interrupter in
low voltage systems include:
Chapter 2: Literature Review
72
• A vacuum interrupter has a low arc voltage, which cannot limit the fault current.
An air circuit breaker however has a high arc voltage, which can significantly
reduce the fault current level.
• The continuous current and fault current level in low voltage systems is higher
than medium voltage level applications. The butt contacts with a high contact
pressure and relatively high contact resistance are not suitable.
In conclusion, an air circuit breaker is a simple and economical technology widely used
in low voltage level applications. A vacuum interrupter is employed only in special
situations where enclosed breaking can offer significant advantages such as polluted or
explosive environments [101].
In high voltage level applications, a vacuum interrupter cannot compete with an SF6
interrupter. There are two main limitations of a vacuum interrupter in high voltage level
applications [101]:
• A lighting impulse voltage level in a vacuum interrupter is limited to 123-145kV.
• A vacuum interrupter has the possibility of emitting X-rays when the voltage is
higher than a hundred kV. Therefore, shielded enclosures are necessary to
protect operators. A vacuum interrupter is not a good choice in high field
applications.
For extreme conditions of high di/dt and initial TRV, a hybrid circuit breaker which
consists of a vacuum interrupter and an SF6 interrupter connected in series has been
investigated and tested [105, 106]. The vacuum interrupter withstands the initial TRV
and the SF6 interrupter endures the peak of the TRV. A prototype hybrid circuit breaker
has successfully demonstrated interruption of a short circuit current of 63kA at 145kV
[105].
2.3.8 Summary
Vacuum interrupters are widely used in medium voltage level applications and extreme
situations in low voltage and high voltage level applications. The development and
basic operating principles of the vacuum interrupter have been summarised. The two
Chapter 2: Literature Review
73
main forms of the vacuum arc, the diffuse mode and constricted mode, were described.
Thereafter, the principle of vacuum breakdown and current breaking were outlined.
Finally, the practical design of vacuum interrupters for high current applications was
summarised.
2.4 Vacuum interrupter actuator
There are three main types of vacuum interrupter actuator in use today: the spring
actuator, solenoid actuator and permanent magnetic actuator. The voice-coil type
actuator, which has the advantages of quick response and low moving mass also has the
potential to be used as a vacuum interrupter actuator.
2.4.1 Spring actuator
A conventional spring actuator was commonly used when the vacuum interrupter was
first invented. It is rarely manufactured now because it has several key disadvantages.
Firstly, the spring dissipates its energy and provides lower force during the travel of the
moving contact whereas a vacuum interrupter requires a higher force when approaching
the open or closed position to stop bouncing. Secondly, a spring actuator typically has
approximately 150 mechanical components, many of which are not standardised parts
[107, 108]. During one operation, up to 30 mechanical components are moving. The
spring actuator therefore is extremely complicated with numerous moving parts, which
can cause reliability problems and it also needs regular maintenance.
2.4.2 Solenoid actuator
A solenoid actuator consists of three parts: a coil which is energised by a DC supply; a
magnetic core to provide the path for the magnetic flux and a moving plunger. Both
slow and fast driving speeds are possible through control of the DC supply to control
the current through the coil. As the plunger of the solenoid approaches the open or
closed position, the force on the plunger increases because the plunger decreases the
reluctance of the magnetic flux path as it moves. This would help prevent the vacuum
interrupter from bouncing [108].
The solenoid actuator however suffers from problems of reliability and regular
maintenance requirements. In addition, in order to hold the movable contact in the open
Chapter 2: Literature Review
74
or closed position, the solenoid coil is required to be energised continuously or have
additional mechanical latches installed. The former method dissipates electric energy
continuously. The latter method requires extra energy to release the latches and
therefore the structure is complex, which would further reduce the reliability of the
mechanism. Moreover, the dynamic characteristics of a solenoid actuator are very
difficult to predict due to magnetic flux leakage, eddy currents and saturation, which
makes the system complex and highly non-linear [109].
2.4.3 Permanent magnetic actuator
In 1987, Manchester University Energy Systems Group developed a new type of
permanent magnetic actuator for operating a medium voltage level vacuum interrupter
[110-112]. Thereafter, the permanent magnetic actuator has drawn great interest due to
its simple structure compared with the conventional type of spring actuator [107, 108,
112-115].
The structure of a permanent magnetic actuator is shown in Figure 2.20. It consists of
seven parts: a stationary iron core, a movable armature, two pieces of permanent magnet,
a shaft, a closing and an opening coil [107, 111]. Compared with a spring actuator, the
total number of parts in the permanent magnetic actuator is dramatically reduced. The
reliability therefore is significantly improved and the permanent magnetic actuator only
needs minor maintenance.
Figure 2.20 Typical structure of a permanent magnetic actuator
Chapter 2: Literature Review
75
A vacuum interrupter with a permanent magnetic actuator is commercially available
today. For example, the VM1 vacuum interrupter with a permanent magnetic actuator
designed by ABB offers in excess of 100,000 operations [107, 116].
When the current is supplied to one of the two coils, the resultant magnetic field
produced by the current through the coil and the permanent magnets begins to change
and the force on the armature changes. The permanent magnetic actuator therefore will
open or close the vacuum interrupter. Energy is only required during the opening or
closing operation. After the actuator achieves its open or closed position, the power
supply can be removed. Without any current through the coil and any mechanical latch,
the armature can be held in the open or closed position by the magnetic force produced
by the permanent magnets.
Eddy currents however are induced in the armature of the permanent magnetic actuator
in response to the changes in the current flowing in the coil. The eddy currents generate
the magnetic flux that opposes the magnetic flux produced by the coil current. As a
result, the resultant magnetic flux in the actuator is reduced, which may significantly
degrade the overall performance of the system. Thin laminated sheets of electric steel
are widely used to minimise eddy currents in motor applications [117]. However, the
speed of the permanent magnetic actuator could be very high, so it is not possible to
laminate the armature because it may reduce its structural rigidity.
2.4.4 Voice-coil type actuator
The voice-coil type actuator is an electromagnetic actuator that generally consists of one
or more coils placed in a magnetic field. The Lorentz force is produced on the coil when
current flows through the coil. This force is proportional to the magnetic flux density,
current through the coil and the length of the coil in the magnetic field [118]. This
actuator type has been known for decades and has been used as the source of force in
loudspeakers and the drive mechanism for disk drive read heads, for example [119, 120].
This actuator type can be a fast and powerful means for operating a vacuum interrupter.
It has the following advantages:
Chapter 2: Literature Review
76
• Compact structure with a small number of parts and maintenance free.
• Light weight of moving part compared with the permanent magnetic actuator.
• High acceleration rate and fast response time.
The major disadvantage of this actuator is the same as a solenoid actuator, i.e. the coil
does not have an inherent stable position when the power supply is removed. There is a
need therefore for a latch mechanism. Magnetic latches, which include latch magnets to
provide the latching magnetic field and latch steel pieces fixed on the moving part can
be installed in the actuator directly [118, 121]. The actuator can therefore be latched in
the open or closed position without energy from an external supply. When the vacuum
interrupter is opening or closing, the force on the coil is higher than the latching force,
so that the actuator releases automatically.
2.4.5 Summary
Four different types of vacuum interrupter actuator, i.e. the spring actuator, solenoid
actuator, permanent magnetic actuator and voice-coil type actuator were introduced and
compared. The structure of the spring actuator and solenoid actuator is too complex,
which causes problems in terms of reliability and maintenance. Permanent magnetic
actuators provide reliable operation with a compact structure. However, the speed is
limited by the heavy armature and eddy currents. The voice-coil type actuator is widely
used in loudspeakers and disk drive read heads and shows great potential for use in a
vacuum interrupter. The benefits of the voice-coil type actuator include its simple
structure and maintenance free design.
Chapter 2: Literature Review
77
2.5 Conclusions
This chapter began with a description of the properties of superconductors, which
exhibit zero DC resistivity below a critical magnetic field strength, critical temperature
and critical current density. Superconducting materials such as BSCCO, YBCO and
MgB2 used for FCLs were introduced including their raw materials, manufacturing
processes and properties.
The chapter then presented the necessity for fault current limitation devices because of
the increase in the fault current level in modern power networks. Conventional methods
of fault current limitation were summarised and the topologies and operating principles
of non-superconducting FCLs and SFCLs were highlighted. The advantages and
disadvantages of different SFCL topologies were summarised and compared. In
addition, the related SFCL projects using the above techniques were detailed. A vacuum
interrupter integrated into the resistive SFCL was proposed to solve its overheating and
recovery problem.
The general principles of the vacuum interrupter operation were summarised including
the development and basic structures of the vacuum interrupter; two main forms of the
vacuum arc - the diffuse mode and constricted mode; and the principles of vacuum
breakdown and current breaking.
Actuation mechanism including the spring actuator, solenoid actuator, permanent
magnetic actuator and voice-coil type actuator were also compared. The voice-coil type
actuator shows great potential for a vacuum interrupter and will therefore be
investigated in this thesis.
Chapter 3: SFCL Coil and Experimental Test Rig
78
3 SFCL Coil and Experimental Test Rig
3.1 Introduction
It is necessary to build the MgB2 SFCL coil before any experimental investigation can
be conducted. The design of an early prototype former will be introduced. The
manufacturing process of the SFCL coil and heat treatment process will then be
described. The instrumentation of the SFCL coil will be presented. The experimental
test rig including the test circuits, control system and cryostat system will also be
discussed in this chapter. The prototype former, instrumentation connection circuit and
experimental test rig were designed and built previously by Oliver et al. [15, 17].
3.2 Coil former
A prototype alumina former was designed to support the MgB2 wire. It was
manufactured by Dynamic Ceramic Ltd. using Dynallox 96 (96% Al2O3). The
specification and drawing of the prototype former is shown in Figure 3.1. The prototype
former shown in Figure 3.2 has a blackened appearance due to the earlier heat
treatment.
The selection of the former material was an important aspect of the design. Alumina
was chosen because it has the following advantages:
• Wide temperature range. The heat treatment of the alumina former with the
MgB2 coil is conducted at 700ºC, whilst the operating temperature is around
-243ºC. The temperature band is almost 1000ºC. The maximum operating
temperature of alumina is 1700ºC [122], which is higher than the temperature of
the heat treatment. The minimum operating temperature of alumina is not
available, but previous tests show that the former had no problem at the
cryogenic temperature [17].
Chapter 3: SFCL Coil and Experimental Test Rig
79
• High thermal conductivity. The thermal performance of the former is a critical
factor. The thermal conductivity of alumina is 25W/m·K [122]. Alumina
therefore is able to conduct heat quickly during a fault test due to its high
thermal conductivity. This can help to keep the temperature of the MgB2 wire
below its melting point in a fault test and to recover quickly after the fault is
cleared.
• High dielectric strength and volume resistivity. The dielectric strength of
alumina is 14.6kV/mm [123] and its volume resistivity is higher than 1014Ω·cm
at 25ºC [122]. Alumina therefore can provide excellent electrical insulation for
the coil, which is important for high voltage applications.
Furthermore, the thermal contraction/expansion of alumina over the operating
temperature range needs to match that of the MgB2 wire, to prevent strain being placed
on the wire. The thermal expansion coefficient of alumina is 7.8×10-6K-1 [122]. Hyper
Tech Research who supplied the MgB2 wire indicated that the thermal expansion
coefficient of the MgB2 wire is 10.38×10-6K-1. The difference between the two thermal
expansion coefficients will cause interference between the MgB2 wire and the former of
about 7.224×10-4, which corresponds to about 0.48mm across the circumference when
cooled down to 20K from room temperature. Alumina therefore is a good material
choice for the MgB2 wire.
14.00mm
5.00mm
134.00mm
5.00mm
14.00mm
3.00mm
Figure 3.1 Former specification (side view) [17]
Chapter 3: SFCL Coil and Experimental Test Rig
80
Figure 3.2 Prototype former
3.3 Coil manufacturing process
There are two methods for fabricating the MgB2 SFCL coil: wind-and-react and react-
and-wind [16]. In the wind-and-react method, the MgB2 wire is wound onto the former
and then heat treated whilst in the react-and-wind method, the MgB2 wire is first reacted
and then wound onto the former. The first approach was used because any cracks
resulting from the winding process may be healed when heat treated. A spool of wire
containing 15 metres of 1.28mm MgB2 wire in un-reacted form was delivered by Hyper
Tech.
3.3.1 MgB2 wire current connections and winding process
The SFCL coil was made up of two interleaved solenoid coils. Both coils were wound
from top to bottom but in opposite current directions. This design effectively reduced
the inductance of the coil by cancelling the solenoidal field created by the two coils.
There were four current connections available on the cryostat bottom plate. These
connections were connected to the copper bar connections at the top of the internal
copper bucket in the cryostat through thermally-anchored leads. The copper bar
Chapter 3: SFCL Coil and Experimental Test Rig
81
connections (numbered 1 to 4) and the annular working space within the copper bucket
are shown in Figure 3.3.
Figure 3.3 Cryostat interior
Copper braid was used to connect the MgB2 wire to the copper bars inside the cryostat.
The benefit of copper braid is that it can cope with any expansion/contraction of the
wire during heat treatment and test in nitrogen. It was necessary to hold the MgB2 wire
in place after winding; copper clamps were machined therefore to attach the MgB2 wire
onto the copper braid. The copper clamp would not only hold the wire in place but also
provide the electrical connection. The copper clamp was loosely placed on the copper
braid, as shown in Figure 3.4, and then fitted into the channel at the top of the former
and folded back, to pass out vertically through the grooves in the top lip, as shown in
Figure 3.5.
The two ends of the copper braid were bolted to the copper connections inside the
cryostat. This arrangement was used at the top of the former for each of the two MgB2
wires forming the interleaved coils. At the bottom of the former, the current connections
for the coils were attached to another copper braid by copper clamps. The copper braid
was then joined into a single loop using a clamp.
Chapter 3: SFCL Coil and Experimental Test Rig
82
Figure 3.4 Clamp used to connect wire to copper braid
Figure 3.5 Top copper braid current connection
The MgB2 wire was placed through the copper clamp leaving several centimetres
protruding past the clamp, to allow a soldered joint to be made after heat treatment. The
connection was made by pressing the MgB2 wire and copper braid between the two
clamp plates and screwing the plates together tightly. The former with the copper braid
was then placed on a lathe. The MgB2 wire was wound onto the former by turning the
lathe slowly. The other end of the MgB2 wire was then passed through the copper clamp
at the bottom of the former, again leaving some of the MgB2 wire protruding for a
soldered connection. This process was repeated for the MgB2 wire forming the second
coil. After winding, the copper clamps were fastened against the MgB2 wire and the
protruding portions of the stainless steel screws used in the clamps were removed and
filed flush with the outside of the clamp plate.
A silica rope was wrapped around the copper braid and the MgB2 wire at the top of the
former, and finally a stainless steel jubilee clip was placed over the silica rope and
clamped tightly. This helped to maintain the tension on the MgB2 wire during heat
treatment so that the complete former could be heat treated with the MgB2 wire in the
coil slots. The silica rope was pre-heated to burn out the cellulose.
Chapter 3: SFCL Coil and Experimental Test Rig
83
3.3.2 MgB2 wire heat treatment
Hyper Tech specified that the MgB2 wire should be held at 700ºC for 20 to 40 minutes
to react the elemental magnesium and boron to form MgB2 [13, 16]. This heat treatment
process would also help to repair any cracks that may have occurred during winding.
Dynamic Ceramic advised that a maximum temperature gradient of 150ºC/hour should
be used for the ceramic former. A low temperature gradient is normally specified to
avoid an uneven heating that might cause the alumina to expand more rapidly on one
side than the other, which may create cracks. From previous experience, there was no
problem heating the former with a temperature gradient of 200ºC/hour. The former was
heated therefore in a vacuum oven with a gradient of 200ºC/hour until it reached 700ºC,
and then held constant in the oven at 700ºC for 30 minutes. After 30 minutes, the oven
was cooled down to 600ºC, and then flowing argon gas was used to cool the oven down
to room temperature with a gradient of 200ºC/hour.
Hydrocarbon impurities in the alumina reduced to carbon during the heat treatment
process, resulting in the blackened appearance. The former and coil after heat treatment
is shown in Figure 3.6 (left).
Figure 3.6 Coil after heat treatment (left) and soldered joints (right)
The protruding MgB2 wire ends either side of the copper clamps were then soldered to
the copper braid to provide extra connection length and strength. The soldered joints are
shown in Figure 3.6 (right).
Chapter 3: SFCL Coil and Experimental Test Rig
84
Instrumentation connections were also mounted onto the coil. Voltage taps and BAS16
diode temperature sensors were soldered onto the MgB2 wire in several places. Four
Cryocon S700 commercial temperature probes were mounted on a piece of fibreglass
board before being attached to the former. The coil instrumentation is described in more
detail in section 3.4.
Kapton tape was wrapped round the entire former to provide voltage insulation between
the coil and the cryostat copper bucket. Heat shrink sleeving was also placed over the
exposed copper braid, again to provide voltage insulation between the braid and the
cryostat copper bucket. The manufactured prototype coil and former, ready to be
installed in the cryostat, is shown in Figure 3.7 (left). The former was finally installed
into the cryostat by lowering it into the annular working space and bolting the copper
braid connections onto the cryostat copper bar connections. A picture of the prototype
coil installed in the cryostat, along with the instrumentation wiring, which was brought
out of the top of the cryostat, is shown in Figure 3.7 (right).
Figure 3.7 Coil ready to be installed in the cryostat (left) and installed in the cryostat
(right)
3.4 Instrumentation
3.4.1 Voltage signals
Each of the two interleaved coils has three full turns and one part turn. The voltage of
each turn was measured by soldering voltage taps directly onto the MgB2 wire. The
exact locations of the voltage taps are described in the next chapter.
Chapter 3: SFCL Coil and Experimental Test Rig
85
The connections for the voltage taps were brought out from the cryostat through a
socket on the top plate and connected to a multi-channel high precision linear
differential amplifier built by the electronics workshop. Each voltage channel was
provided with a variable gain setting which could be adjusted by changing a resistor.
High gain was used for the coil when it was superconducting and low gain was used
when the coil was expected to quench. The output from the voltage amplifier was
connected to a LabVIEW interface card. The instantaneous voltage signals were
recorded by LabVIEW during testing.
3.4.2 Current signals
Two Hall Effect current sensors were used to measure the current passing through the
coil. One was rated up to 100A, to provide an accurate current reading for low current;
the other one was rated up to 1000A, to measure high current. A power supply was
manufactured by the electronics workshop at the university to provide ±12V to the
sensors. Appropriate measurement resistors were connected to the outputs to convert the
current signals to voltage signals and then connected to the LabVIEW interface card.
3.4.3 Temperature signals
Twelve Cryocon S700 temperature probes were placed in the cryostat. These probes use
an industrial small bobbin package to provide good thermal contact to the diode
junction. The diameter of each probe is around 8mm. Four of these probes were located
on the four internal current leads connecting the external supply to the current
connections. Instrumentation wires for these four sensors were brought out of the
cryostat through the socket on the bottom plate and connected to a Cryocon 14
temperature monitor. Another four sensors were located on the outside surface of the
copper bucket at the top, middle, bottom and on the cryocooler cold head. These four
probes were connected to a Cryocon 34 temperature controller. The controller and
monitor front panels are shown in Figure 3.8.
The last four probes, which were mounted onto a fibreglass board, are shown in Figure
3.9. The board was then mounted onto the side wall of the former. The probes on the
left and right of the board fitted onto the top and bottom copper braids and MgB2 wire
joints, whilst the other middle two measured the temperature of the former.
Chapter 3: SFCL Coil and Experimental Test Rig
86
The connections to these four probes were brought out of the cryostat through the socket
on the top plate and connected directly to the Cryocon 14 temperature monitor. The
monitor displays the four temperature readings on the front screen and provides one
temperature as an output. This output was connected directly to the LabVIEW interface
card.
Figure 3.8 Temperature monitor (top) and temperature controller (bottom)
Figure 3.9 Cryocon S700 silicon diode temperature probes
Accurate temperature measurement of the MgB2 wire is important in preventing the
wire from overheating. The Cryocon temperature probes were too large to fit into the
slot housing the wire. It was found however that a common diode could be used to
measure temperatures from 20K to 300K [124-126]. The forward voltage drop in the
p-n junction of a diode depends on the temperature at constant current. A constant
current power supply of 10µA was designed for the diodes and built by the electronics
workshop. The BAS16 diode temperature sensor with a SOT23 package was used to
measure the temperature of the MgB2 wire. The BAS16 diode was chosen due to its
Chapter 3: SFCL Coil and Experimental Test Rig
87
low-cost and small size. One spare pin on the diode package was soldered onto the
MgB2 wire to provide good thermal contact. One of the diodes soldered onto the MgB2
wire in the slot is shown in Figure 3.10. BAS16 diodes from the same manufacturer had
similar characteristics with temperature. However, the accuracy can be significantly
improved by individual calibration [124, 125]. All the diodes therefore were calibrated
before carrying out any tests.
The connections to these sensors were brought out of the cryostat again through the
socket on the top plate and connected to an isolated operational amplifier circuit. The
output from the amplifier was connected to the LabVIEW interface card. The
temperature was recorded by LabVIEW during and after testing.
Figure 3.10 BAS16 diode temperature sensor soldered onto the MgB2 wire
3.5 Control
3.5.1 High-current test circuit
A controllable high current supply was built to test the quench behaviour of the SFCL
coil. A schematic of the high-current supply is shown in Figure 3.11. Switch control and
data acquisition was provided by the LabVIEW system. The switch was controlled
directly from LabVIEW to turn on and off. Voltage, current and temperature signals
were also monitored and recorded using the LabVIEW data acquisition system.
A variac was directly connected to the laboratory supply. The output of the variac was
connected to a current step-up transformer with a turns ratio of 4:1 through the point-on-
wave switch. This switch was manufactured by the electronics workshop. The switching
point was at current zero during each test. The switching operation was performed using
a back-to-back thyristor arrangement. A gate signal was sent by LabVIEW to close the
switch. The switch would be turned off at the next current zero-crossing after removal
Chapter 3: SFCL Coil and Experimental Test Rig
88
of the gate signal. Furthermore, the switch outputted a single pulse for each cycle,
which was read into LabVIEW to control the number of current cycles passed through
the coil during testing.
Figure 3.11 High-current test circuit schematic [15, 17]
A variable load resistor was placed in series with the SFCL coil in the secondary side of
the transformer to represent the system impedance during a fault. This resistor could be
varied from 0 to 0.45Ω. The maximum current supplied to the coil, with the variac set at
100% and the load resistance at 0Ω, was estimated to be 1053Apeak. It should be noted
however that this level of current would not be reached in practice because the coil
would quench before this current level was reached.
The test supply was constant voltage rather than constant current. The coil would act
therefore as an SFCL when transitioning from the superconducting state to the normal
resistive state. The impedance of the coil was assumed to be negligible in the
superconducting state. The ratio of the load resistor to the quenched coil resistance
would directly determine the current limitation.
3.5.2 Low-current test circuit
A low-current supply was made by the electronics workshop to provide a constant low
current (up to 3Apeak) signal. A schematic of the low-current supply is shown in Figure
Chapter 3: SFCL Coil and Experimental Test Rig
89
3.12. The low-current supply produced a variable frequency and waveform signal. This
could be used for a frequency sweep to determine the MgB2 coil inductance.
Figure 3.12 Low-current test circuit schematic [15, 17]
The low-current supply was controlled by a gate signal from LabVIEW, which turned
the supply on and off at the next current zero-crossing. Again the unit outputted a single
pulse for each cycle of supply that was read into LabVIEW to control the number of
current cycles passed to the coil.
3.5.3 LabVIEW control programme
All the data acquisition and control was performed using a LabVIEW embedded
system. The graphical user interface is shown in Figure 3.13. The number of current
cycles to pass through the coil could be set by ‘Current cycle to test’ on the left top
corner of the screen. When the ‘Run’ button was pressed, LabVIEW sent an ‘On’ signal
to the switch and the current supply was connected to the coil at the next voltage zero-
crossing.
When the programme was running, the ‘cycle counter’ output from the switch was read
in by LabVIEW and incremented. All the voltage channels, current sensors and diode
temperature sensor signals were also read in at a rate of 2000 samples per second during
a test. These signals were displayed in charts on the screen and saved to a text file on
the PC for post signal processing.
Chapter 3: SFCL Coil and Experimental Test Rig
90
When the number of current cycles was equal to the set value, LabVIEW sent an ‘Off’
signal to the switch. Current flow in the coil would then be turned off at the next current
zero-crossing. All the signals were still read in and recorded at a rate of 10 samples per
second until the ‘Stop post test recording’ button was pressed. The diode temperature
sensor signals from ‘Temperature post test’ would be used as an indication of the
thermal recovery of the MgB2 wire.
Figure 3.13 Screenshot of LabVIEW control programme
3.5.4 Cryostat
The cryostat system was designed and built by Scientific Magnetics [127]. This system
included five major parts: a vacuum insulated cryostat, a cryostat top lid, a cryocooler
compressor, a vacuum pump set, temperature monitors and a controller.
The cryostat consisted of an aluminium vacuum vessel surrounding a cylindrical copper
bucket. As shown in Figure 3.3, a central copper cylinder was mounted inside the
copper bucket at the bottom. The annular space between them was the working space
for the former. A polystyrene cylinder was placed inside the central copper cylinder to
reduce the amount of liquid nitrogen required. A fibreglass neck tube was connected
Chapter 3: SFCL Coil and Experimental Test Rig
91
between the copper bucket and the cryostat top flange. The fibreglass tube is not
designed to be placed in compression and could be damaged therefore if the vacuum is
not established before filling with liquid nitrogen. The vacuum level of the vacuum
vessel should be better than 1×10-3mbar before the liquid nitrogen could be poured into
the bucket [127].
Figure 3.14 Cryostat system: vacuum cryostat vessel with vacuum pump set (left),
vacuum pump (top middle), gauge (bottom middle) and cryocooler compressor (right)
The copper current connections were located around the top lip of the copper bucket, as
shown in Figure 3.3 (numbered 1 to 4), to which the coil current connections were
made. The internal current connections were connected to the external current
connections on the bottom plate by internal leads through the vacuum vessel. These
leads were thermally anchored to the cryocooler cold head to prevent heat transfer from
ambient temperature. The bottom plate of the cryostat contained connections for
instrumentation such as temperature probe connections, heater control, and the four
external current connections. A 50W heater was located at the top of the cryostat copper
bucket. This heater could be controlled to maintain a given temperature at any of the
four temperature sensors connected to the temperature controller. The heater power was
controlled by a PI control loop.
Chapter 3: SFCL Coil and Experimental Test Rig
92
Figure 3.15 Cryostat top plate
The top lid was made up of the top plate and insulation baffles. The top plate, shown in
Figure 3.15, contained four connectors for test instrumentation wiring, pressure relief
valve, nitrogen vapour vent check valve, lifting points and a port through which liquid
nitrogen was poured when the cryostat was closed. These four connectors were used for
the diode temperature sensors, normal voltage taps, detailed voltage taps, and
commercial temperature probe wire connections.
Cryogenic temperature was reached using a Gifford–McMahon AL230 single-stage
cryocooler, as shown in Figure 3.14 (right). Two pipes from the front panel of the
cryocooler carried compressed helium to the cold head and low pressure helium back to
the cryocooler. The cryostat was manufactured so that the copper bucket sat on top of
the cold head, cooling the copper bucket rapidly due to its high thermal conductivity.
The performance of the cryocooler is strongly influenced by the quality of the vacuum
in the vessel. The vacuum pump and gauge are shown in Figure 3.14 (middle). The
vacuum was created using a Leybold PT70 B compact turbo molecular pump with
forepump. The pump line was connected to a valve at the bottom of the cryostat, along
with a pressure gauge.
Chapter 3: SFCL Coil and Experimental Test Rig
93
3.6 Conclusions
Material selection and geometry of the prototype former to support the MgB2 wire was
first introduced. The wind-and-react method was used to manufacture the MgB2 SFCL
coil. Copper clamps were manufactured to attach the MgB2 wire onto the copper braids
for the current lead connections. The MgB2 wire was then wound onto the alumina
former and heat treated in a vacuum oven. The instrumentation consisting of voltage,
current and temperature signals was also described. The high-current and low-current
experimental test circuits and the LabVIEW control system were also presented. Details
of the cryostat system including a vacuum insulated cryostat, a cryostat top lid, a
cryocooler compressor, a vacuum pump set, temperature monitors and a controller were
included.
Chapter 4: Experimental Investigation of SFCL Coils
94
4 Experimental Investigation of SFCL Coils
4.1 Introduction
A single-strand MgB2 SFCL coil with a wire diameter of 1.28mm was initially
investigated. The MgB2 wire sample was delivered on a spool from Hyper Tech. Two
sections of MgB2 wire were used, each over three metres long, for winding the
interleaved design coil on the ceramic former. To react the elemental magnesium and
boron to form MgB2 and anneal possible cracks in the MgB2 wire during the winding
process, the heat treatment process described in the previous chapter was carried out at
Bodycote Heat Treatment Ltd. in Stockport. The instrumentation probes were mounted
onto the coil afterwards. The main purpose for the experimental tests on the single-
strand coil was to determine whether MgB2 in simple round wire form would be suitable
as a resistive SFCL.
A three-strand MgB2 SFCL coil was then tested to determine the current capacity for
multi-strand wire. The full quench current level will be reduced if the quench current
level and impedance of each wire strand is not identical; therefore, ensuring each strand
carries the same current is one of the most important aspects. The diameter of each
individual wire was 0.63mm, and the three wires were hand braided with a pitch of
18.4mm. The smaller diameter wire was used because of the limitation in the slot size of
the former. Two braids were fabricated, hand wound onto the former, heat treated by
Hyper Tech and delivered to Manchester afterwards. The instrumentation probes again
were mounted onto the coil.
The specification of both wires is shown in Table 4.1. The chemical composition of the
two wires was the same; only the percentage of each composition is slightly different.
Chapter 4: Experimental Investigation of SFCL Coils
95
Table 4.1 MgB2 wire specification
MgB2 wire Strand B source Powder Barrier Sheath
1 1617 SMI-97/98 MgB2 Niobium Monel
2 1617-3-B SMI-97/98 MgB2 Niobium Monel
MgB2 wire OD (mm) Mg:B % Powder % Barrier % Sheath
1 1.28 1:2 30.7 17.1 52.2
2 0.63 1:2 30.6 14.8 54.4
4.2 Single-strand SFCL Coil
4.2.1 Instrumentation
Two series of voltage taps were mounted onto the single-strand coil: the normal voltage
taps were connected to each turn; and the detailed voltage taps were equally distributed
on the first turn. There were eight voltage channels on the whole coil, which is shown
schematically in Figure 4.1 (left). Each interleaved coil had three full turns and one part
turn. Voltage channels 1 to 3 and 6 to 8 were full turns, whilst voltage channels 4 and 5
were part turns. In order to monitor the voltage distribution starting from one end of the
coil, eight detailed voltage channels 11 to 18 were connected to the first turn, as shown
in Figure 4.1 (right). The notation A, B and C corresponds to the first two turns in
Figure 4.1 (left).
These two series of voltage channels shared the same voltage amplifier external to the
cryostat; they were used in separate tests for different purposes. The normal voltage
channels were first used to get an overview of the voltage distribution across the coil
and then the detailed voltage channels were used to get the voltage of each section in the
first turn.
Chapter 4: Experimental Investigation of SFCL Coils
96
Figure 4.1 Schematic showing locations of normal voltage taps (left), detailed voltage
taps and BAS16 diode temperature sensors (right)
The distribution of temperature in the cryostat was measured using four Cryocon S700
commercial temperature probes mounted around the former. Six BAS16 diode
temperature sensors were directly mounted onto the wire to measure the temperature of
the wire. The locations of the six temperature diodes, which are denoted as D1 to D6,
are shown in Figure 4.1 (right). Diodes 1 to 4 were mounted on the coil at the middle of
the detailed voltage channels 11, 13, 15 and 17, whilst diodes 5 and 6 were mounted on
the coil in the first and second half of the second turn. The output voltage of each diode
was connected to the LabVIEW data acquisition board.
All the diodes were calibrated before carrying out any tests to improve their accuracy;
the diode voltage could then be easily converted to the corresponding temperature by
interpolation of the calibration data. This recorded diode voltage would indicate the
temperature rise after quench and the recovery process of the coil. It is important to
ensure that the temperature of the coil returns to the normal operating temperature
before the next quench test can be conducted to prevent any possible damage or
degrading of the coil.
Chapter 4: Experimental Investigation of SFCL Coils
97
4.2.2 Calibration of BAS16 diode temperature sensors
Temperature is one of three important parameters for the MgB2 wire to remain in the
superconducting state. As mentioned in section 3.4.3, BAS16 silicon diode temperature
sensors were used to measure the coil temperature. The characteristics of each diode
however might be a little different. Individual calibration of each diode therefore would
significantly improve their accuracy.
After the cryostat was filled with liquid nitrogen, the temperature inside the cryostat was
measured at approximately 77K after several minutes. The cryocooler compressor was
then turned on to bring the temperature inside the cryostat down to 20K. One of the
Cryocon temperature sensors mounted on the fibreglass board was used to provide a
calibrated temperature reading to which the diodes were calibrated. The diode voltages
were recorded using the LabVIEW calibration programme per each change in
temperature, i.e. one Kelvin.
After the entire test was completed, the cryostat was then warmed up to room
temperature by turning off the cryocooler and using the heater inside the cryostat to
warm it up more quickly. The temperature controller was set higher than the measured
temperature in the cryostat, step by step, to ensure the full transition of nitrogen across
the two phase boundaries during warming up. The LabVIEW programme again
recorded the diode voltages per each change in temperature.
The voltage of each diode with respect to temperature when cooling down (1st) and
warming up (2nd) is shown in Figure 4.2. The results show that the voltages of the six
diodes have similar characteristics. It is clear that each diode voltage when warming up
and cooling down is similar at the same temperature. The measured diode voltages in
any subsequent tests therefore could be converted to the equivalent temperatures easily
using simple interpolation. The BAS16 diodes have a closely constant sensitivity of
2.7mV/K between 30K and 300K, but below 30K the sensitivity increases to 3.9mV/K.
Chapter 4: Experimental Investigation of SFCL Coils
98
0.4
0.6
0.8
1
1.2
1.4
1.6
20 90 160 230 300
Temperature (K)
Dio
de
Vo
ltag
e (V
)
Ch.1 1st
Ch.2 1st
Ch.3 1st
Ch.4 1st
Ch.5 1st
Ch.6 1st
Ch.1 2nd
Ch.2 2nd
Ch.3 2nd
Ch.4 2nd
Ch.5 2nd
Ch.6 2nd
Figure 4.2 Temperature diode calibration curves when cooling down (1st) and warming
up (2nd)
4.2.3 Temperature profile
The objective of the temperature profile test was to obtain the impedance of the coil at
different temperatures and to determine if the coil was superconducting below the
critical temperature.
The temperature of the cryostat was slowly controlled from 20K to 40K. The low-
current supply was used to pass two cycles of low level current through the coil at
different temperatures. A gain of 981 was used for the voltage amplifier when the coil
was superconducting, whilst a gain of 51 was used when the coil transitioned to the
normal resistive state. The impedance of each turn was determined using the root mean
square (RMS) voltage divided by the RMS current.
Figure 4.3 shows the impedance per metre of the MgB2 wire based on the average
impedance of each turn. It is obvious that the impedance can be divided into three
sections: below 36.2K, the wire is superconducting and the impedance (inductance) is
constant; between 36.2K and 37.3K, the coil is in the transition zone in which the
impedance increases quickly with temperature; and above 37.3K, the impedance
increases steadily, which is similar to a normal conductor with increasing temperature.
The resistance of the coil was 0.27Ω/m at room temperature (293K), reducing to
Chapter 4: Experimental Investigation of SFCL Coils
99
0.0327Ω/m at 37.3K. The critical temperature is defined as the temperature at which the
resistivity is half of the normal state resistivity. The critical temperature for the coil is
approximately 36.7K.
0
0.01
0.02
0.03
0.04
20 24 28 32 36 40
Temperature (K)
Imp
edan
ce p
er m
etre
(Ω
/m)
Figure 4.3 Temperature profile of the coil
4.2.4 Frequency sweep
A frequency sweep test was carried out to determine the impedance of the coil with
varying frequency in the superconducting state.
The low-current supply was used to pass two cycles of low level current through the
coil with the frequency varying from 10Hz to 100Hz at 25K. The current and voltage
across each turn was recorded by LabVIEW. Again the impedance of each turn was
determined using the RMS voltage divided by the RMS current. Figures 4.4 and 4.5
show the individual impedance of each turn and total impedance of the coil with varying
frequency. It can be seen that the impedance of each turn is similar and they have a
linear relationship to frequency. This indicates that the coil is dominated by its
inductance in the superconducting state.
The interleaved coil design was employed to minimise the inductance; however, there is
still a relatively small and finite coil inductance, which will exceed the internal
Chapter 4: Experimental Investigation of SFCL Coils
100
inductance of the MgB2 wire. A first order equation for the trend line of impedance
gives an estimated coil inductance of 6.26µH.
0
0.0002
0.0004
0.0006
0.0008
0.001
0 20 40 60 80 100Frequency (Hz)
Imp
edan
ce p
er m
etre
(Ω
/m) Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.4 Impedance of the coil with varying frequency at 25K
y = 0.0000088x + 0.0000186
R2 = 0.9996766
0
0.0002
0.0004
0.0006
0.0008
0.001
0 20 40 60 80 100
Frequency (Hz)
Imp
edan
ce p
er m
etre
(Ω
/m)
Average
Linear(Average)
Figure 4.5 Total coil impedance with varying frequency at 25K
4.2.5 Quench tests
Quench tests were carried out for a number of reasons: firstly, to determine whether the
coil would limit a fault current; secondly, if it did, whether it would recover to the
superconducting state afterwards; thirdly, what effect would the operating temperature
Chapter 4: Experimental Investigation of SFCL Coils
101
have on the quench current level; and fourthly, whether the coil would have any
noticeable deterioration after successive quench tests.
The high-current test circuit was used to pass two cycles of high current and the
LabVIEW programme recorded all the instrumentation channels. Quench tests were
first carried out at a temperature of 34K, to reduce the quench current level and hence
reduce the heat dissipated in the coil during quench. The potential peak current level
was gradually increased by adjusting the set point of the variac and the load resistor.
The potential peak current was calculated based on the coil in the superconducting state,
which had negligible impedance. In practice, when the potential peak current was higher
than the quench current level, the calculated peak current could not be achieved because
the coil already quenched before reaching the peak current level.
The potential peak current was increased gradually until the coil started to quench.
Figures 4.6 to 4.8 show the coil response during a short cycle quench test at 34K with a
potential peak current of 311A. It is clear that the coil is quenching (but not
completely). The peak current is limited to 270A in a quarter-cycle. It is interesting that
the whole coil does not quench at the same current level. Figure 4.7 clearly shows that
channels 1, 4, 5, 7 and 8 quench earlier than the other channels.
Figure 4.8 shows the temperature response of the six sections of the coil using the
BAS16 diode temperature sensors. The temperature rise of the first section is about 5K,
whilst the other sections do not show any observable temperature rise. There will be
heat generated if any coil section is quenching, leading to a temperature rise in that coil
section. This demonstrates that the wire is partially quenching. The quench test only
lasted for 50 milliseconds; however, the coil took 50 seconds to recover to the operating
temperature of 34K. The temperature of the coil was checked to make sure the coil was
fully recovered before the next test was carried out, to prevent possible damage to the
coil.
Chapter 4: Experimental Investigation of SFCL Coils
102
-6
-4
-2
0
2
4
6
0 0.01 0.02 0.03 0.04 0.05
Time (s)
Vo
ltag
e (V
)
-300
-200
-100
0
100
200
300
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.6 Coil response at 34K with a potential peak current of 311A
0
1
2
3
0.005 0.007 0.009 0.011 0.013 0.015
Time (s)
Vo
ltag
e (V
)
0
100
200
300
Cu
rren
t (A
)Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.7 Coil response at 34K with a potential peak current of 311A, highlighting the
point of quench
Chapter 4: Experimental Investigation of SFCL Coils
103
32
34
36
38
40
0 10 20 30 40 50 60
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.8 Coil temperature response at 34K with a potential peak current of 311A
The potential peak current was further increased to 372A. In Figure 4.10, channels 1, 4,
5, 7 and 8 quench earlier than the other channels, which is the same as in the previous
quench test. However, channel 3 also quenches in the second test. This indicates that as
the potential peak current increases, more sections of the coil are quenching.
It is clear that the voltage across the first turn rises as the potential peak current
increases. The Joule losses in the first turn have increased because of the high current
and voltage; this explains why the temperature rise of the first section is about 6.5K in
Figure 4.11, which is higher than the previous test. As a result, the recovery time for the
coil is also longer. The BAS16 diode temperature sensors are clearly very important in
indicating the temperatures of the wire sections.
From the previous tests, it can be seen that channels 1, 4, 5 and 8 quench earlier than the
other channels. The particular feature of these channels is that they are the end turns of
the coil. There are three reasons which may explain why the end turns quench first.
Firstly, the end of the coil was connected to the copper braid in the cryostat with the
copper braid connected to the copper bar and then to the external connections. The
temperature of the copper bar would therefore be slightly higher than the MgB2 coil.
Secondly, the copper braid incurs Joule losses when current is flowing through it, which
causes the temperature to rise. This would mean that the coil ends were slightly warmer
Chapter 4: Experimental Investigation of SFCL Coils
104
than the coil centre leading to a reduction in the critical current. Thirdly, the magnetic
field around the end of the coil may be different than at the middle. The magnetic field
can affect the quench of superconductors as discussed in section 2.1.1.2. These explain
why the coil begins to quench initially at the coil ends.
-9
-6
-3
0
3
6
9
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)
-300
-200
-100
0
100
200
300
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.9 Coil response at 34K with a potential peak current of 372A
0
1
2
3
4
5
0.02 0.022 0.024 0.026 0.028 0.03
Time (s)
Vo
ltag
e (V
)
0
100
200
300
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.10 Coil response at 34K with a potential peak current of 372A, highlighting
the point of quench
Chapter 4: Experimental Investigation of SFCL Coils
105
32
34
36
38
40
42
0 10 20 30 40 50 60
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.11 Coil temperature response at 34K with a potential peak current of 372A
Short cycle quench tests were repeated at 32K and 30K, to evaluate the variation in the
quench current with temperature. A sample of 32K and 30K test results are shown in
Figures 4.12 to 4.17. The voltage amplifier had a maximum output voltage of 10.5V and
‘clipped’ the signals that went above this level. As shown in Figure 4.12, the voltage
channel 5 is clipped by the amplifier. It is worth mentioning that in Figure 4.14 the
temperature rises in both the first and third sections of the first turn. This indicates more
sections of the first turn are quenching due to the higher current.
-12
-8
-4
0
4
8
12
0 0.01 0.02 0.03 0.04 0.05
Time (s)
Vo
ltag
e (V
)
-600
-400
-200
0
200
400
600
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.12 Coil response at 32K with a potential peak current of 622A
Chapter 4: Experimental Investigation of SFCL Coils
106
0
2
4
6
8
10
0 0.002 0.004 0.006 0.008 0.01
Time (s)
Vo
ltag
e (V
)
0
100
200
300
400
500
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.13 Coil response at 32K with a potential peak current of 622A, highlighting
the point of quench
30
35
40
45
50
55
0 30 60 90 120
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.14 Coil temperature response at 32K with a potential peak current of 622A
Chapter 4: Experimental Investigation of SFCL Coils
107
-12
-8
-4
0
4
8
12
0 0.01 0.02 0.03 0.04 0.05
Time (s)
Vo
ltag
e (V
)
-700
-500
-300
-100
100
300
500
700
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.15 Coil response at 30K with a potential peak current of 700A
0
2
4
6
8
10
12
0.005 0.007 0.009 0.011 0.013 0.015
Time (s)
Vo
ltag
e (V
)
0
100
200
300
400
500
600
700
Cu
rren
t (A
)Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Current
Figure 4.16 Coil response at 30K with a potential peak current of 700A, highlighting
the point of quench
It is clear from all the tests conducted that the coil successfully recovered to the
superconducting state after quench. The repeated quench tests show that the quench
behaviour was consistent.
Chapter 4: Experimental Investigation of SFCL Coils
108
28
30
32
34
36
38
40
0 10 20 30 40 50 60
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.17 Coil temperature response at 30K with a potential peak current of 700A
From the tests, the quench currents at 34K, 32K and 30K are 245A, 433A and 639A
respectively. As shown in Figure 4.18, the quench current level increases linearly as the
temperature reduces. This demonstrates that a higher quench current level can be
achieved by reducing the operating temperature of the coil. However, the quench
current level along the wire tends to be less uniform when operating at a lower
temperature. Furthermore, reducing the operating temperature would increase the
equipment costs, so there is a compromise between the quench current level and the
normal operating temperature.
200
300
400
500
600
700
30 31 32 33 34
Temperature (K)
Qu
ench
Cu
rren
t (A
)
Figure 4.18 Estimated quench currents versus temperature
Chapter 4: Experimental Investigation of SFCL Coils
109
4.2.6 Long duration quench tests
After the short cycle quench tests were complete, long duration quench tests were
conducted to determine if the coil was capable of limiting the current over a longer time
period because the protection equipment in a power network may take up to several
hundred milliseconds to clear the fault.
The operation of the coil during a ten-cycle quench test with a potential peak current of
372A is illustrated in Figures 4.19 to 4.22. The coil begins to quench at a current of
approximately 276A, and then as the quench progresses rapidly throughout the coil, the
current is limited to 110Apeak after ten cycles. Figure 4.20 shows the voltage of each
turn during a quench test for ten cycles. During this test, all the channels quench except
channel 6.
It was observed that the coil started to quench in the end turns. It was assumed that the
coil started to quench from the end, propagating along the coil. As detailed in section
4.2.1, eight voltage taps were mounted equidistance along the first turn of the coil, but
they shared the same voltage amplifier channels with the normal voltage taps. After it
was confirmed that the coil could survive in a ten-cycle quench test, this quench test
was repeated and the detailed voltages were recorded. Figure 4.21 shows that detailed
voltage channels 11 and 12 start to quench at the first peak current and channels 13 and
14 start to quench in sequence afterwards. This confirms the hypothesis that the coil
starts to quench from the end and propagates along the coil. This also demonstrates that
the coil is partially quenched, and more and more sections are quenching as the
temperature rises. The coil is however still not fully quenched during this test.
The temperature rise of the coil is shown in Figure 4.22. Again, as detailed in section
4.2.1, the first diode temperature sensor was mounted onto the coil between detailed
voltage channel 11 and the second sensor was mounted onto the coil between detailed
voltage channel 13. Figure 4.22 shows that both channels are quenched; the temperature
in these sections therefore increases. It is clear from Figure 4.22 that the temperature
rise of the first and the third sections is 30K and 8K respectively. The voltage on
detailed voltage channel 11 is higher than channel 13; the power dissipated is therefore
higher in the first section. This explains why the temperature rise of the first diode
Chapter 4: Experimental Investigation of SFCL Coils
110
temperature sensor is much higher than the second. Figure 4.22 also shows that the coil
takes more than two minutes to recover.
-300
-200
-100
0
100
200
300
0 0.05 0.1 0.15 0.2 0.25
Time (s)
Cu
rren
t (A
)
Figure 4.19 Coil current response during a ten-cycle quench test with a potential peak
current of 372A
-9
-6
-3
0
3
6
9
0 0.05 0.1 0.15 0.2 0.25
Time (s)
Vo
ltag
e (V
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.20 Coil voltage response during a ten-cycle quench test with a potential peak
current of 372A
Chapter 4: Experimental Investigation of SFCL Coils
111
-2
-1
0
1
2
0 0.05 0.1 0.15 0.2 0.25
Time (s)
Vo
ltag
e (V
)
Voltage Ch.11
Voltage Ch.12
Voltage Ch.13
Voltage Ch.14
Voltage Ch.15
Voltage Ch.16
Voltage Ch.17
Voltage Ch.18
Figure 4.21 Coil first turn detailed voltage response during a ten-cycle quench test with
a potential peak current of 372A
30
40
50
60
70
0 30 60 90 120
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.22 Coil temperature response during a ten-cycle quench test with a potential
peak current of 372A
A one-second quench test with a potential peak current of 372A was conducted to
determine whether the coil can survive for a one-second fault. The operation of the coil
is shown in Figures 4.23 and 4.24. The current is limited to 300Apeak in the first quarter-
cycle then reduces to 63Apeak after one second. This test demonstrates the coil is able to
Chapter 4: Experimental Investigation of SFCL Coils
112
carry the fault current for one second without any damage or degradation. The highest
temperature recorded on the coil is 107.2K during the test, as shown in Figure 4.24.
-300
-200
-100
0
100
200
300
0 0.2 0.4 0.6 0.8 1
Time(s)
Cu
rren
t (A
)
Figure 4.23 Coil current response during a fifty-cycle quench test with a potential peak
current of 372A
30
50
70
90
110
0 30 60 90 120
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.24 Coil temperature response during a fifty-cycle quench test with a potential
peak current of 372A
Chapter 4: Experimental Investigation of SFCL Coils
113
4.2.7 Simulated fault test
A simulated fault test was conducted to determine the coil behaviour when a real fault
occurs. The coil initially carried rated current in the superconducting state when the
current was suddenly increased to simulate a fault. The objective of this test was to
access the current-limiting performance of the coil in a more realistic condition.
Initially, the coil operated with a peak current of 40A. The fault was made by manually
closing a switch which was in parallel with the load resistor, to produce a short circuit
fault. The potential current was 316Apeak without any coil resistance. As shown in
Figure 4.25, in practice the current was limited to 252Apeak in a quarter-cycle, reducing
to 115Apeak over eight cycles. This test demonstrates that the coil can effectively limit a
fault current in a relatively short period of time.
-300
-200
-100
0
100
200
300
0 0.1 0.2 0.3 0.4 0.5
Time (s)
Cu
rren
t (A
)
Figure 4.25 Coil current response to simulated fault at 34K with a potential peak
current of 316A
4.2.8 Temperature rise test
An SFCL would have to operate at or below its rated current most of its life in a
practical application. The purpose of a temperature rise test was to determine whether
the AC losses in the superconducting coil would cause a slow increase in temperature,
leading to an unexpected quench. The normal steady state temperature of the coil was
Chapter 4: Experimental Investigation of SFCL Coils
114
recorded with the coil operating at 30K with a current of 200Apeak for more than one
hour.
Figure 4.26 illustrates the recorded temperature during the test. It is clear that at this
current level there is no obvious temperature rise after one hour of steady operation. A
steady state temperature rise of less than 0.1K was recorded. This demonstrates that an
SFCL manufactured with a similar design to the coil in this test would be able to operate
at rated conditions without experiencing an unwanted quench.
29
29.5
30
30.5
0 1000 2000 3000 4000
Time (s)
Tem
per
atu
re (
K) Temp Ch.1
Temp Ch.2
Temp Ch.3
Temp Ch.4
Temp Ch.5
Temp Ch.6
Figure 4.26 Coil temperature response during continuous 200Apeak current test at 30K
for one hour
4.2.9 Summary
The experimental tests showed that the coil successfully operated as an SFCL with
consistent characteristics. The tests proved that the quench current increased nearly
linearly as the temperature was reduced. Furthermore, the tests confirmed that the coil
was dominated by its inductance in the superconducting state and the resistance
increased quickly and became dominant after the coil started to quench.
The quench process started from the ends of the coil and extended throughout the coil as
the fault current increased. A more uniform quench would be produced when the coil
was operated with a fault current which significantly exceeded the quench current of the
Chapter 4: Experimental Investigation of SFCL Coils
115
coil. The long duration quench test showed that the coil was able to limit the fault
current for one second without any damage.
The coil heated up quickly after quench and took several minutes to recover to the
superconducting state. Some measures therefore need to be taken to protect the coil
from heating up, to reduce the recovery time. One option included placing a fast-acting
vacuum interrupter in series with the SFCL coil.
A steady state temperature rise of less than 0.1K was observed during normal
superconducting operation after the coil was tested for one hour, which demonstrated its
thermal stability in continuous operation.
4.3 Three-strand SFCL Coil
In a practical power system application, parallel MgB2 wires will likely be required in
order to carry current levels in the kA range. Ensuring each strand of wire carries the
same current is a very important issue for multi-strand wires because the critical current
level will decrease if the critical current and impedance of each wire strand is not
closely identical. This was achieved by braiding the strands to equalise the impedance
of each parallel wire path. The three strand wires were braided together, as shown in
Figure 4.27.
Figure 4.27 Picture of the three-strand coil (Courtesy of Hyper Tech)
4.3.1 Instrumentation
It was impossible to measure the current in each strand directly using conventional
current transducers because of the braiding. The voltage taps mounted onto a section of
each strand however would be a good indication of the current distribution. The
locations of the voltage taps are shown in Figure 4.28. Each coil on the former was
made up of three full turns and one part turn. The solid dots in Figure 4.28 indicate the
Chapter 4: Experimental Investigation of SFCL Coils
116
locations of the voltage taps. The wire strands themselves were un-insulated and the
sheaths were in random electrical contact along the length of the coil. Insulation paper
was used to separate and insulate each strand where individual voltage taps were
required on the first turn of the two interleaved coils. The voltage taps on each strand
therefore could be measured and used to monitor each strand current in the braid.
Figure 4.28 Schematic showing locations of the voltage taps
4.3.2 Temperature profile
The impedance of the coil from 20K to 40K using the same method as described in
section 4.2.3 is shown in Figure 4.29. The resistance at 37.3K and at room temperature
are 0.0408Ω/m and 0.38Ω/m for the three-strand coil, whilst they are 0.0327Ω/m and
0.27Ω/m for the single-strand coil.
It is clear that the resistance of the three-strand coil is higher than the single-strand coil.
The chemical composition of the two coils is similar; the resistance therefore is mainly
determined by the cross-sectional area. The cross-sectional area of the single-strand coil
is about 1.37 times the three-strand coil. From the resistance and known cross-section
area, the electrical resistivity of the single-strand coil is 3.47×10-7Ω·m and the three-
strand coil is 3.55×10-7Ω·m at room temperature.
Chapter 4: Experimental Investigation of SFCL Coils
117
0
0.01
0.02
0.03
0.04
0.05
20 24 28 32 36 40
Temperature (K)
Imp
edan
ce p
er m
etre
(Ω
/m)
Figure 4.29 Temperature profile of the coil
4.3.3 Frequency sweep
The impedance of the coil was obtained using the same method as described in section
4.2.4. The low-current supply was used to pass two cycles of low level current through
the coil with the frequency varying from 10Hz to 100Hz at 25K. The voltage channels 1
to 3 and 6 to 8 show similar voltages. The voltages on channels 1 and 6 were used
therefore to calculate the impedance for the first and fifth turns. Figures 4.30 and 4.31
show the individual impedance of each turn and the total impedance of the coil with
varying frequency. Again the results show that the coil is inductance dominated in the
superconducting state. The inductance of the coil is approximately 3.65µH from the
trend line in Figure 4.31. This reduction (when compared to the single-strand coil)
requires further investigation.
0
0.0001
0.0002
0.0003
0.0004
0.0005
0.0006
0 20 40 60 80 100Frequency (Hz)
Imp
edan
ce p
er m
etre
(Ω
/m)
Voltage Ch.1
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.9
Voltage Ch.10
Figure 4.30 Impedance of the coil with varying frequency at 25K
Chapter 4: Experimental Investigation of SFCL Coils
118
y = 0.0000052x + 0.0000110
R2 = 0.9990691
0
0.0001
0.0002
0.0003
0.0004
0.0005
0.0006
0 20 40 60 80 100
Frequency (Hz)
Imp
edan
ce p
er m
etre
(Ω
/m)
Average
Linear(Average)
Figure 4.31 Total coil impedance with varying frequency at 25K
The total voltage drop across each strand is related to its resistance and inductance,
expressed as Equation 4.1. VA, VB and VC are the total voltages across each strand. These
voltages are identical because the coil ends are connected to the copper braid. The
resistances RA, RB and RC consist of the strand resistance and the contact resistance
between each strand and the copper braid. IA, IB and IC are the currents flowing through
each strand. The self-inductance of each strand is denoted as LA, LB and LC. MAB, for
example, represents the mutual inductance between strand A and strand B.
/
/
/
A A A AA AB AC
B B BA B BC B B
CA CB CC C C C
R I di dt VL M M
R I M L M di dt V
M M LR I di dt V
+ =
(4.1)
In the superconducting state, the resistance of each strand can be influenced strongly by
the contact resistance and it is difficult to ensure that the contact resistance of each
strand is identical. It is important in these applications therefore to make sure that the
inductance of each coil strand is sufficient to dominate the individual strand impedance.
The coil inductance is clearly the major factor in helping to ensure each strand carries
the same current. The self-inductance of each strand is determined by the physical size
of the former and the position of each strand on the former. Each individual strand is
closely linked magnetically to the other strands because they have been braided
Chapter 4: Experimental Investigation of SFCL Coils
119
together. Their coefficients of coupling are close to unity therefore and the mutual
inductances will be closely related to the self-inductances. The three strands are braided
therefore to ensure the impedances of each individual strand are closely matched: this is
a common technique used in winding coils in large generators [128], for example.
4.3.4 Current sharing test
In practice, there have been quite a lot of problems in ensuring the uniform distribution
of the current in each strand [129, 130]. The objective of this experimental test was to
determine the current distribution between each of the three strands.
The high-current test circuit was used to pass five cycles of current through the coil at
an initial temperature of 30K. The voltage and current signals were recorded and are
shown in Figure 4.32. This figure show a 90º phase shift between the coil voltage and
current, which again demonstrates that the coil is inductance dominated. It also clearly
shows that the voltages in the first and fifth turns of each strand show no observable
difference, which indicates that the current is the same in each wire strand. In the
superconducting state, the coil inductance clearly plays an important role in current
sharing in a multi-strand wire. This result confirms that with careful coil design, it is
possible to scale-up the current levels using multi-strand wire.
-3
-2
-1
0
1
2
3
0.1 0.12 0.14 0.16 0.18 0.2
Time (s)
Vo
ltag
e (V
)
-90
-60
-30
0
30
60
90
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Voltage Ch.9
Voltage Ch.10
Current
Figure 4.32 Coil response with current flow at 30K
Chapter 4: Experimental Investigation of SFCL Coils
120
4.3.5 Quench tests
Two cycles of high current were passed through the coil to test the coil quench
behaviour. The voltage and current were recorded by LabVIEW.
Figure 4.33 shows the current and voltage waveforms of a quench test at 34K with a
potential peak current of 249A. It is clear from Figure 4.33 that all the turns have
quenched but only partially quenched. The current was limited to 178Apeak within a
quarter-cycle.
Figure 4.34 highlights the voltage of each strand on the first and fifth turns. It can be
seen that the voltages of channels 1 to 3 are higher than channels 6 to 8. This may be
explained by the temperature variation within the cryostat test chamber. Channels 6 to 8
were near the bottom of the copper bucket in the cryostat and close to the cold head.
The temperature therefore would be slightly lower at the bottom of the coil compared to
the connections at the top of the coils. As a result, the quench current would be lower on
the first turn compared to the fifth turn and this is where the quench would be expected
to commence. Figure 4.34 also shows that the voltage of each strand is still exactly the
same after quench, which again demonstrates that the quench current of each strand is
identical. This further confirms that it is possible to use multi-strand MgB2 wire to
scale-up the current levels for power system applications.
-6
-3
0
3
6
0 0.01 0.02 0.03 0.04
Time (s)
Vo
ltag
e (V
)
-200
-100
0
100
200
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Voltage Ch.9
Voltage Ch.10
Current
Figure 4.33 Coil response at 34K with a potential peak current of 249A
Chapter 4: Experimental Investigation of SFCL Coils
121
-6
-3
0
3
6
0 0.01 0.02 0.03 0.04
Time (s)
Vo
ltag
e (V
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.34 Each coil strand voltage response of the first and fifth turns at 34K with a
potential peak current of 249A
Figure 4.35 shows the current and voltage waveforms of a quench test at 32K. Again it
is clear that all the turns are partially quenched. The voltages of channels 1 to 3 still
show no noticeable difference, as are channels 6 to 8. The voltages of channels 1 to 3
are still higher than channels 6 to 8, which is the same as the previous test.
-12
-6
0
6
12
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)
-400
-200
0
200
400
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Voltage Ch.9
Voltage Ch.10
Current
Figure 4.35 Coil response at 32K with a potential peak current of 467A
Chapter 4: Experimental Investigation of SFCL Coils
122
-12
-6
0
6
12
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.36 Each coil strand voltage response of the first and fifth turns at 32K with a
potential peak current of 467A
Figures 4.37 to 4.39 present the results of the quench test carried out at 30K with a
potential peak current of 529A. It can be seen in Figure 4.37 that channels 5 and 10 do
not quench whilst the other channels are partially quenched. The voltages of channels 1
to 3 are still the same, but the voltages of channels 6 to 8 show differences now.
Channels 6 to 8 still quench at the same time but the voltage of channel 6 is lower than
channels 7 and 8 in the first cycle and later the voltage of channel 8 is lower than
channel 7 as well. This is possibly an indication of some minor material variations in the
MgB2 wire. It is also worth pointing out that the voltage difference between channel 6
and channel 7 is smaller in the second cycle compared to the first. This shows that the
parallel strands will try to equalise the currents naturally.
Chapter 4: Experimental Investigation of SFCL Coils
123
-10
-5
0
5
10
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)
-500
-250
0
250
500
Cu
rren
t (A
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Voltage Ch.9
Voltage Ch.10
Current
Figure 4.37 Coil response at 30K with a potential peak current of 529A
-10
-5
0
5
10
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.38 Each coil strand voltage response of the first and fifth turns at 30K with a
potential peak current of 529A
Chapter 4: Experimental Investigation of SFCL Coils
124
-1.5
-1
-0.5
0
0.5
1
1.5
0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Vo
ltag
e (V
)Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.39 Each coil strand voltage response of the fifth turn at 30K with a potential
peak current of 529A
The quench test was repeated and the coil demonstrated repeatable and reliable current-
limiting properties with no detectable degradation of the MgB2 wire performance during
the quench process.
The quench currents at different temperatures, summarised from repeated tests, are
shown in Figure 4.40. The quench current level has a nearly linear relationship to
temperature, which is the same as the single-strand coil. At the same temperature, the
quench current level for the three-strand coil is lower than the single-strand coil. This is
because the equivalent cross-sectional area of MgB2 is smaller than the single-strand
coil. The diameter of the MgB2 wire used in the single-strand coil and the three-strand
coil are 1.28mm and 0.63mm respectively. The portion of MgB2 is 30.7% in the single-
strand coil and 30.6% in the three-strand coil. The equivalent cross-sectional area of
MgB2 is therefore 0.395mm2 and 0.286mm2. It is easy to determine that the quench
current density for the single-strand coil and the three-strand coil are about 620A/mm2
and 531A/mm2 at 34K with a self-field of 50Hz. Hyper Tech suggested that the MgB2
wire, which they manufactured for MRI scanners, had a critical current density variation
between 15% and 20%. Therefore, the quench current density difference of the two
coils was acceptable.
Chapter 4: Experimental Investigation of SFCL Coils
125
0
100
200
300
400
30 31 32 33 34
Temperature (K)
Qu
ench
Cu
rren
t (A
)
Figure 4.40 Estimated quench currents versus temperature
4.3.6 Long duration quench test
Following the short cycle tests, the three-strand coil was subjected to a ten-cycle quench
test to check whether the MgB2 wire could survive a long duration fault typical of a
power system application.
The behaviour of the coil during a ten-cycle quench test at 34K with a potential peak
current of 249A is illustrated in Figures 4.41 to 4.43. The peak current is limited to
177A in a quarter-cycle and further reduces to 69A over ten cycles, as the coil resistance
increases with temperature. It can be seen from Figure 4.42 that all the turns have
quenched but not fully quenched. From Figure 4.43 it is also clear that the voltages
across the parallel strands are closely identical during the ten-cycle test. This test
demonstrates again that the three-strand SFCL coil shares the current equally and can
limit the fault current for 0.2s without any degradation.
Chapter 4: Experimental Investigation of SFCL Coils
126
-200
-100
0
100
200
0 0.05 0.1 0.15 0.2 0.25
Time (s)
Cu
rren
t (A
)
Figure 4.41 Coil current response during a ten-cycle quench test with a potential peak
current of 249A
-8
-4
0
4
8
0 0.05 0.1 0.15 0.2 0.25
Time (s)
Vo
ltag
e (V
)
Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.4
Voltage Ch.5
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Voltage Ch.9
Voltage Ch.10
Figure 4.42 Coil voltage response during a ten-cycle quench test with a potential peak
current of 249A
Chapter 4: Experimental Investigation of SFCL Coils
127
-8
-4
0
4
8
0 0.05 0.1 0.15 0.2
Time (s)
Vo
ltag
e (V
)Voltage Ch.1
Voltage Ch.2
Voltage Ch.3
Voltage Ch.6
Voltage Ch.7
Voltage Ch.8
Figure 4.43 Each coil strand voltage response of the first and fifth turns during a ten-
cycle quench test with a potential peak current of 249A
4.3.7 Summary
The initial investigation of a three-strand SFCL coil has shown that the coil successfully
performed repeatable and reliable operations as an SFCL. Each of the three wire strands
shared the current equally during the current sharing test and demonstrated closely
identical responses during quench current tests at 34K and 32K. However, the voltage
across each strand of the fifth turn was different during quench current test at 30K,
which suggested that the coil tend to be less uniform at lower operating temperature.
It is believed that an SFCL using MgB2 in wire form with multiple strands in parallel
shows considerable potential as a practical method for scaling-up the current levels
required for power system applications.
Chapter 4: Experimental Investigation of SFCL Coils
128
4.4 Conclusions
The experimental investigations conducted on the two coils showed reliable and
repeatable current-limiting properties in short cycle and long time duration quench tests.
Tests on both coils demonstrated that the quench current increased nearly linearly as the
temperature reduced. The quench current density of the single-strand coil was about
16.7% higher than the three-strand coil at 34K with a self-field of 50Hz but in general
this was considered acceptable.
The inductance of the two coils was relatively small but finite, determined by the
physical structure of the former. Tests proved that the coil was dominated by inductance
in the superconducting state; the resistance became dominant after the coil started to
quench.
A significant problem for the resistive SFCL coil is that there are high Joule losses in
the coil after it quenches, so the temperature of the coil rises quickly and takes a long
time to recover to the superconducting state. Solutions are needed therefore to protect
the coil from heating up, to reduce the recovery time. One option is placing a fast-acting
vacuum interrupter in series with the SFCL coil, which will be investigates in the
chapters 6 to 8.
Chapter 5: Modelling of SFCL Coil
129
5 Modelling of SFCL Coil
5.1 Introduction
There are two interdependent aspects of the MgB2 SFCL coil that need to be considered
carefully: the electrical characteristics and the thermal properties. Resistivity of a
superconductor is strongly dependent on temperature, whilst the temperature rise of the
coil is caused by resistance Joule losses. A MATLAB model will be developed to
simulate the behaviour of the Hyper Tech 1.28mm MgB2 wire. There are two main
purposes for this MATLAB model: firstly, the simulation results will be compared with
the experimental results to validate the theory of the SFCL operation; secondly, the
validated MATLAB model will be used to predict and provide guidance for future
SFCL design work.
To acquire more accurate thermal information of the SFCL coil, a finite element (FE)
thermal model will be built using Flux2D [131]. The thermal analysis will be performed
using a transient thermal two-dimensional (2-D) FE solution. Power dissipated in the
coil will be calculated from the experimental results obtained. The FE thermal model
will therefore be used to simulate the thermal response of the coil using the measured
power dissipation. The temperature from the FE thermal model will be first compared
with the experimental results. After validation, the FE thermal model will then be used
to predict the thermal response of the coil for a three-second fault.
The results from the MATLAB model using adiabatic boundary conditions will estimate
the highest temperature; whilst the FE model, assuming perfect thermal contact between
the coil and nitrogen, will estimate the lowest temperature. The practical temperature
rise of the coil will fall between these two boundaries if there are no hot spots in the
coil. This would provide useful design guidance for further development.
Chapter 5: Modelling of SFCL Coil
130
5.2 MATLAB model
The SFCL coil cannot be modelled as a stand-alone device. If an SFCL coil is
connected to a network where the quenched resistance of the coil is dominant, the
current level is limited effectively and then the resistance develops slowly. On the other
hand, if the quenched coil resistance is not dominant, the current-limiting effect is not as
significant and the resistance develops quickly in the coil. The behaviour of the coil
therefore is highly dependent on the network or system impedance and has to be
modelled as part of the network. It was decided that the high-current test circuit should
be used as the network for the MATLAB model of this SFCL coil.
There are three operating regions for a superconductor: the superconducting region, the
transition region and the normal resistive region. In the superconducting region, the
resistance is negligible and in the MATLAB model, the resistance therefore is taken as
zero. In the transition region, the current density rises above the lower critical current
density and the coil starts to develop resistance. The resistance increases due to the
increasing current density. The coil temperature increases as a result due to the Joule
losses. The current density then starts to reduce because of the increasing resistance.
Due to the dissipated losses however, the temperature of the coil either increases above
the critical temperature or the current increases above the reduced upper critical current
density. The coil then transitions to the normal resistive region. In the normal resistive
region, the resistance is mainly dominated by the temperature. The resistance of the coil
is similar to a normal conductor with increasing temperature.
5.2.1 Model description
As described in the previous chapter, the high-current test circuit consisted of a constant
voltage source and a variable load resistor. The load resistance, connecting lead
resistance, coil inductance and transformer impedance were taken into consideration in
the MATLAB model. The fault current was reduced after the coil quenched, which
reduced the Joule losses and temperature rise in the coil. The limited current level
depended on the resistance of the coil following quench compared with the load
resistance.
Chapter 5: Modelling of SFCL Coil
131
In this MATLAB model, the SFCL coil was divided into arbitrary equal sections along
one of the interleaved coils. The MATLAB model time solution used a small time step,
for example, 200µs for a short cycle simulation and 500µs for a long duration
simulation. The calculation was repeated at each time step until the simulation time was
complete. At each time step, the MATLAB model undertook the following procedures:
• Set up the initial conditions such as temperature and resistance of each coil
section.
• Calculate the voltage and current variation in the circuit.
• Calculate the heat generated and temperate rise of each coil section using the
heat equation.
• Calculate the resistance of each coil section with updated current and
temperature.
The details of the calculations at each step are explained in the following section. The
first step is setting up the initial conditions. At the beginning of the simulation, an
operating temperature of 34K, for example, was specified as the initial temperature of
the central coil sections with a linear increase in temperature specified, so that the coil
ends were 0.1K hotter. The initial resistance of all the sections were set to be zero. From
the second time step, the initial conditions were obtained from the simulation result of
the previous time step.
The second step is calculation of the voltage and current variation. The high-current test
circuit consisted of a constant voltage source connected across a load resistor in series
with the SFCL coil through a transformer. The expression for the current variation is
given by:
( )s s c fcls
c
V I R Rdi
dt L
− += (5.1)
where: sdi
dt = Rate of change of current in the circuit (A/s)
Vs = Voltage of supply (V)
Chapter 5: Modelling of SFCL Coil
132
Is = Circuit current (A)
Rc = Circuit resistance, including load, transformer and lead resistance (Ω)
Rfcl = SFCL coil resistance (Ω)
Lc = Circuit inductance, including transformer and SFCL inductance (H)
The third step is calculation of the heat generated and temperature rise of each coil
section. The coil was taken as a solid conductor with adiabatic boundary conditions,
with heat only conducted along the conductor to simulate heat conduction along the coil
[118]. The heat was generated from the resistance Joule losses. The temperature rise
was calculated from the standard one-dimensional (1-D) heat conduction equation
[132]:
2
2
( , ) 1 ( , )d pc
CT x t T x tg
x k k t
ρ∂ ∂+ =∂ ∂
(5.2)
where: T = Temperature (K)
x = Distance along the coil length (m)
t = Time (s)
k = Thermal conductivity (W/m·K)
gc = Power generated in the coil per unit volume (W/m3)
ρd = Density (kg/m3)
Cp = Specific heat capacity (J/kg·K)
where:
2fcl s
c
R Ig
v= (5.3)
where: v = Volume of the SFCL Coil (m3)
It should be noted that the values of the specific heat capacity and thermal conductivity
vary with temperature. It was difficult however to obtain reliable material data at
cryogenic temperatures. The specific heat capacity and thermal conductivity of the
materials used from 20K to 500K therefore were sourced from [133-136], detailed in
Chapter 5: Modelling of SFCL Coil
133
Appendix A. The data was superimposed into single average curves and saved into
tables with respect to temperatures. In the MATLAB model, they were determined
using a lookup table with interpolation between the two nearest temperature points
[118].
The heat conduction equation shows that the rate of thermal energy generation equals
the rate of increase of internal energy, minus the net rate of heat gain by conduction
[132]. A partial differential equation solver was used to solve the above equation. The
MATLAB model however took about ten minutes to simulate one cycle (20ms), which
was too slow to integrate into a power system analysis. A performance test on the
MATLAB model showed that the majority of the processing time was taken by solving
the partial differential equation. Some modifications were therefore necessary to
optimise the MATLAB model. If the heat transfer along the coil was neglected, the heat
equation could be rearranged as:
c
d P
gdT
dt Cρ= (5.4)
If c
d P
g
Cρ is taken as constant during each time step, the equation therefore can be
expressed as:
c
d P
g tT
Cρ∆∆ = (5.5)
The MATLAB model simulation time was significantly reduced to several seconds for
one cycle using this approximation. The results from both MATLAB models will be
compared in the next section.
The fourth step is calculation of resistance. The resistance of the SFCL coil is highly
non-linear. In the MATLAB model it is divided into the following four regions: (1) If
the temperature is greater than 37.3K, the coil is in the normal resistive region and the
resistance of the coil is similar to a normal conductor with increasing temperature. The
Chapter 5: Modelling of SFCL Coil
134
R-T relationship determined from the test results is used to calculate the resistance. (2)
If the temperature is between 36.2K and 37.3K, the coil is developing resistance as a
function of temperature. The resistance was determined therefore from the temperature
profile. (3) If the temperature is lower than 36.2K and the current density exceeds the
critical current density, the coil is developing resistance due to current. An S-curve
approximation was used for the current-resistance profile, as detailed by several papers
[137-139]. (4) If it is none of the above, this means that the coil is in the
superconducting region and the resistance is therefore zero.
5.2.2 Comparison with short cycle quench tests
The MATLAB models including and neglecting heat conduction along the coil were
used to simulate quench behaviour at 34K, 32K and 30K with the same potential peak
current as the experimental tests. Current waveforms from the MATLAB models
compared to the experimental tests are shown in Figures 5.1 to 5.3.
In these graphs, the trace in blue is the experimental test result on the coil; the trace in
red is the MATLAB modelled result including heat conduction along the coil, and the
trace in pink is the MATLAB modelled result without considering heat conduction. It
should be noted that the red and pink traces overlap, so the red trace cannot be seen. The
results demonstrate that the MATLAB model with heat conduction along the coil gives
practically the same results as the one without heat conduction. Heat conduction along
the wire therefore does not affect the SFCL coil behaviour for a short cycle simulation.
The simulation time is significantly reduced by neglecting heat conduction along the
coil.
These results show a good correlation between the MATLAB model and the
experimental results for the first cycle. This confirms that the equations used to model
the quench process of the SFCL coil are reasonable.
In the MATLAB model, the resistance was not allowed to reduce in value during this
simulation, to avoid large oscillations. It is therefore suitable for fault current levels that
are much higher than the quench current level.
Chapter 5: Modelling of SFCL Coil
135
-200
-100
0
100
200
300
0 0.005 0.01 0.015 0.02
Time (s)
Cu
rren
t (A
)
Experimental
Modelled withheat transferalong the wire
Modelledwithout heattransfer
Figure 5.1 Results comparison for a fault at 34K with a potential peak current of 372A
-400
-200
0
200
400
600
0 0.005 0.01 0.015 0.02
Time (s)
Cu
rren
t (A
)
Experimental
Modelled withheat transferalong the wire
Modelledwithout heattransfer
Figure 5.2 Results comparison for a fault at 32K with a potential peak current of 622A
Chapter 5: Modelling of SFCL Coil
136
-400
-200
0
200
400
600
800
0 0.005 0.01 0.015 0.02
Time (s)
Cu
rren
t (A
)
Experimental
Modelled withheat transferalong the wire
Modelledwithout heattransfer
Figure 5.3 Results comparison for a fault at 30K with a potential peak current of 700A
5.2.3 Comparison with long duration quench tests
Figures 5.4 and 5.5 show the comparison between the experimental and modelled
results for ten-cycle and fifty-cycle quench tests. It can be seen that the MATLAB
model shows a good correlation with the experimental results for the first cycle.
However, for subsequent cycles the MATLAB model shows a slightly lower current
than the experimental test. This would indicate that the modelled results give a higher
resistance. A possible reason for this is that the MATLAB model assumes adiabatic
boundary conditions. In practice the coil was wound on the alumina former and
surrounded by the solid nitrogen. There would be heat transfer therefore to the alumina
former and/or the solid nitrogen.
The MATLAB model was then revised to include the heat conduction into the nitrogen.
The coil therefore was modelled as a long straight conductor surrounded by the
nitrogen. The temperature of the coil was used as the boundary temperature between the
coil and the nitrogen. The temperature of the solid nitrogen at a certain distance, i.e. two
and half times the radius from the centre of the coil, for example, was taken as the
ambient operating temperature. This approximation will be checked using an FE
thermal model described in section 5.3.2. The solid nitrogen from the coil boundary to
the distance specified for ambient was divided into twenty annular segments. The heat
conduction equation was used to calculate the temperature variation of each segment by
Chapter 5: Modelling of SFCL Coil
137
setting ‘gc’ to zero, i.e. there is no heat generated in the nitrogen. The temperature rise
of the coil for each step was then calculated as the energy generated by the resistance
minus the heat dissipated into the nitrogen. The heat dissipated into the nitrogen was
calculated as the sum of the increase in the nitrogen internal energy.
-300
-200
-100
0
100
200
300
0 0.05 0.1 0.15 0.2
Time (s)
Cu
rren
t (A
)
Experimental
Modelled
Figure 5.4 Results comparison for a ten-cycle fault at 34K with a potential peak current
of 372A
-300
-200
-100
0
100
200
300
0 0.2 0.4 0.6 0.8 1
Time (s)
Cu
rren
t (A
)
Experimental
Modelled
Figure 5.5 Results comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A
Chapter 5: Modelling of SFCL Coil
138
Figures 5.6 and 5.7 show comparisons of the experimental and modelled results
considering the heat dissipated into the nitrogen for ten-cycle and fifty-cycle quench
tests. It is clear that the model now shows an excellent correlation with the experimental
results during the period measured. However, the simulation time using this MATLAB
model was significant: 24 hours for a fifty-cycle quench.
-300
-200
-100
0
100
200
300
0 0.05 0.1 0.15 0.2
Time (s)
Cu
rren
t (A
) Experimental
Modelled withheat transferto nitrogen
Figure 5.6 Results comparison for a ten-cycle fault at 34K with a potential peak current
of 372A (considering the heat dissipated into the nitrogen)
-300
-200
-100
0
100
200
300
0 0.2 0.4 0.6 0.8 1
Time (s)
Cu
rren
t (A
) Experimental
Modelled withheat transferto nitrogen
Figure 5.7 Results comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A (considering the heat dissipated into the nitrogen)
Chapter 5: Modelling of SFCL Coil
139
The limitation in the MATLAB model is that it can only calculate heat transfer in one
direction with a constant thermal conductivity in each section at each time step. The
MATLAB model therefore only took the nitrogen into consideration. However, the
alumina former has a higher thermal conductivity; it can conduct the heat quicker than
the nitrogen but this MATLAB model did not take it into consideration.
The temperature response from both MATLAB models is shown in Figure 5.8. The
temperature after a one-second quench test from the adiabatic model is 170K, whilst for
the model taking the nitrogen into consideration, it is 135.9K. The measured
temperature of the coil after the quench test is 107.2K. The temperature from the
MATLAB model is clearly higher than the test result. This is caused possibly because
of the thermal resistance between the surface of the coil and the temperature sensor.
0
40
80
120
160
200
0 0.2 0.4 0.6 0.8 1
Time (s)
Tem
per
atu
re (
K)
Modelledadiabatic
Modelledwith heattransfer tonitrogen
Figure 5.8 Temperature comparison for a fifty-cycle fault at 34K with a potential peak
current of 372A with adiabatic and nitrogen boundary conditions
5.2.4 Summary
A MATLAB model of the SFCL coil temperature response has been developed. This
MATLAB model successfully predicted the SFCL response over a range of
temperatures. The MATLAB model can be used to predict the SFCL behaviour in a
power system analysis and also to aid future designs. It is useful to indicate the
Chapter 5: Modelling of SFCL Coil
140
maximum temperature of the SFCL coil with adiabatic boundary conditions because
this is the worst case situation.
The results from the MATLAB model considering heat conduction along the coil were
almost identical to the model ignoring heat conduction for a short cycle faults. Heat
conduction along the coil therefore was neglected to reduce the simulation time.
5.3 Finite element thermal model
After the MATLAB model was validated, a transient thermal FE model was used to
perform a thermal analysis of the coil. Flux2D provides a 2-D numerical approximation
to the following equation [140]:
1d P
T T Trk k g C
r r r z z tρ∂ ∂ ∂ ∂ ∂ + + = ∂ ∂ ∂ ∂ ∂
(5.6)
where: g = Power loss density (W/m3)
r = Distance from the centre along the radial direction (m)
z = Distance from the centre along axis direction (m)
5.3.1 Model description
To build the FE thermal model, it is necessary to define the geometry and the materials
used. A copper spacer was placed in the centre of the cryostat working space. The
middle of the copper spacer was filled with a polystyrene space-filler to save volume
and reduce the required amount of liquid nitrogen. The alumina former with the MgB2
coil wrapped around it was placed in the cryostat annular working space and the coil
was connected to the copper connections using copper braid. Nitrogen took up all the
remaining working space. The geometry was entered as an axisymmetric model into
Flux2D. The FE model geometry and mesh are shown in Figures 5.8 and 5.9. The outer
three surfaces were defined as imposed temperature boundaries, to keep the external
surface of the cylinder at a constant operating temperature of 34K.
Chapter 5: Modelling of SFCL Coil
141
Copper
braid
MgB2
coil
Alumina
former
Polystyrene
space-filler
Nitrogen
Copper
spacer
Figure 5.9 Flux2D FE model of the geometry
Figure 5.10 Mesh of the Flux2D FE model, showing detail of the coil in the slot
The specific heat capacity and thermal conductivity of the materials detailed in
Appendix A also were used in the FE model. The thermal contact resistances were not
taken into consideration in the FE model because of the lack of reliable data. During the
Chapter 5: Modelling of SFCL Coil
142
winding process, gaps between the coil and the former were common. However, the size
of these gaps was difficult to estimate precisely. The gap therefore would be taken as a
variable as part of a sensitivity analysis. The distance of the coil from the former varied
from 0mm to 1.5mm.
The amount of thermal energy dissipated in the coil and copper braid connections were
calculated from the experimental results obtained from the SFCL coil. The FE thermal
model therefore was built to simulate the thermal response of the coil using the
measured power loss. Heat generated in the coil and copper braid was calculated as
follows. The coil was dominated by resistance after quenching. The instantaneous
power loss density in the coil was then determined using the expression:
cm cmcm
V Ig
v=
(5.7)
where: gcm = Power loss density in the coil (W/m3)
Vcm = Measured voltage across the coil (V)
Icm = Measured current passing through the coil (A)
The instantaneous power loss density in the coil during a one-second fault is shown in
Figure 5.11. It is clear that the power loss density in the coil is effectively reduced due
to the limited fault current. This is one of the reasons why the coil was able to survive
during a long duration fault.
Chapter 5: Modelling of SFCL Coil
143
0
1
2
3
4
5
6
7
0 0.2 0.4 0.6 0.8 1
Time (s)
Po
wer
den
sity
(G
W/m
3 )
Figure 5.11 Instantaneous power loss density in the coil during a one-second fault at
34K with a potential peak current of 372A
The cross-sectional area of copper in the copper braid is 16.08mm2, so the resistance of
the copper braid was calculated from:
cbcb
cb
lR
A
ρ= (5.8)
where: Rcb = Resistance of the copper braid (Ω)
ρ = Resistivity of copper (Ω·m)
lcb = Length of the copper braid (m)
Acb = Cross-sectional area of copper in the copper braid (m2)
The circuit was equivalent to two parallel copper braids; each one carried half the
current of the coil. The instantaneous power loss density in the copper braid connections
was determined using the following expression:
2
2cm cb
cbcb
I Rg
v =
(5.9)
where: gcb = Power loss density in the copper braid (W/m3)
Chapter 5: Modelling of SFCL Coil
144
vcb = Volume of the copper braid (m3)
The power loss density in the coil and copper braid was specified using the time varying
values by writing the data into two separate files [141]. In the FE model, the power
density for the coil and copper braid automatically linked to the two files to calculate
the temperature response. For any simulation time between two time points, the power
loss density was calculated using linear interpolation between the two nearest points.
5.3.2 Results and discussion
The temperature response with different gaps using the measured power loss density is
shown in Figure 5.12. The temperature drops slightly near one second because the test
only lasted for 49 cycles (0.98s) and there is no heat generated after this. The FE model
used the space average power dissipated in the coil. It is therefore difficult to say what
the peak temperature in the coil was during testing. However, it is predicted that the
average temperature during a one-second quench test would be in the range from 69.2K
to 122.7K. The temperature rise of the coil wound on the former without a gap is
dramatically lower than the other sizes of gaps. This confirms that the coil would
benefit by minimising the thermal contact resistance between the coil and the former.
The variation in temperature rise in other cases, except perfect contact, is small. The
distance of the coil from the former was therefore set at 1.5mm for the remaining
simulations.
The temperature profile vertically down the former through the centre of the coil with a
gap of 1.5mm from the former is shown in Figure 5.13. The temperature rise focuses in
the coil and drops dramatically around the coil. The temperature profile showing the
detail of the coil in the slot is shown in Figure 5.14. The temperature at the outer
boundary, which is two and half times the radius from the coil centre, is about the same
as the operating temperature of 34K. This confirms that in section 5.2.3, the selection of
this distance as the ambient operating temperature in the MATLAB model is reasonable.
It can be seen from Figure 5.14 that the temperature of the surrounding nitrogen drops
dramatically the further from the coil. In the MATLAB model, the nitrogen was divided
into 20 annular sections between the coil boundary and the nitrogen at 34K. This
sections number was chosen as a compromise between computational time and accuracy.
Chapter 5: Modelling of SFCL Coil
145
0
20
40
60
80
100
120
140
0 0.2 0.4 0.6 0.8 1
Time (s)
Tem
per
atu
re (
K)
0.0mm
0.1mm
0.2mm
0.3mm
0.5mm
0.8mm
1.0mm
1.5mm
Figure 5.12 Time variation of temperature in the centre of the coil for varying distances
from the former
0
20
40
60
80
100
120
140
0 20 40 60 80 100 120
Path length (mm)
Tem
per
atu
re (
K)
Figure 5.13 Temperature profile vertically down the former through the centre of the
coil after a one-second fault, with a gap of 1.5mm from the former
Chapter 5: Modelling of SFCL Coil
146
Figure 5.14 Temperature profile after a one-second fault, with a gap of 1.5mm from the
former
The temperature response from both numerical models with adiabatic boundary
conditions and including the nitrogen is shown in Figure 5.15. The FE model shows a
good correlation with the temperature response predicted by the MATLAB model. The
temperature rise of the numerical models including the nitrogen is about 40K lower than
the simple adiabatic boundary conditions. The temperature rise of the MATLAB model
is slightly higher than the FE model. Under adiabatic boundary conditions in the FE
model, it did not solve correctly if the thermal conductivity of the nitrogen was set to
zero; a low value of 10-3W/m·K therefore was specified. This explains why the FE
model had a slightly lower temperature rise under adiabatic boundary conditions.
The temperature variation following a quench for 0.04, 0.2 and 1 seconds from the
experimental test results, and the two numerical models are summarised in Table 5.1.
The temperature from the test results is consistently lower than those predicted by the
numerical models. The thermal resistance between the surface of the coil and the
temperature sensor may cause lower temperature measurements., Additionally, it is
Chapter 5: Modelling of SFCL Coil
147
possible that the coil might be close to the former, which would produce a lower
temperature rise. Furthermore, both numerical models do not consider phase change of
nitrogen, which absorbs energy without temperature rise.
0
40
80
120
160
200
0 0.2 0.4 0.6 0.8 1
Time (s)
Tem
per
atu
re (
K)
MATLABadiabatic
FE adiabatic
MATLAB withheat transferto nitrogenFE with heattransfer tonitrogenExperimental
Figure 5.15 Temperature response comparison for a one-second fault at 34K with a
potential peak current of 372A from the MATLAB model and the FE model
Table 5.1 Temperature response comparison with fault time variation
Quench
time (s)
Experimental
temperature
(K)
MATLAB
temperature
(adiabatic)
(K)
MATLAB
temperature
(nitrogen)
(K)
FE
temperature
(adiabatic)
(K)
FE
temperature
(nitrogen)
(K)
0.04 41.2 51 48.2 50.4 47
0.2 56.3 89.3 75.6 86.7 71.7
1 107.2 178 135.9 174.6 132
Figure 5.16 shows the temperature profile through the coil centre after a one-second
fault from both numerical models. The temperature in the coil and surrounding nitrogen
is very similar in both models. It shows that the temperature rise is centralised in the
coil, probably due to the relatively poor thermal conductivity of the surrounding
nitrogen. This again indicates the importance of measuring the coil temperature as close
to the coil as possible in order to obtain accurate measurements.
Chapter 5: Modelling of SFCL Coil
148
0
30
60
90
120
150
0 0.8 1.6 2.4 3.2
Length (mm)
Tem
per
atu
re (
K)
MATLAB
FE
Figure 5.16 Temperature profile through the centre of the coil for a one-second fault at
34K with a potential peak current of 372A from the MATLAB model and the FE model
5.3.3 Summary
The FE model showed a reasonable correlation with the experimental test results,
demonstrating that the data for the thermal properties of the materials specified in the
FE model were reasonable. The FE model predicted an average temperature in the range
from 69.2K to 122.7K in the SFCL for a fifty-cycle fault with a gap between the coil
and the former between 0mm to 1.5mm.
Comparing the results from the FE model and the MATLAB model with the
experimental test results, the temperature rise for different time variation demonstrated a
reasonable correlation. The two numerical models are useful therefore for predicting the
quench behaviour and the temperature rise of the coils with a similar design.
5.4 Prediction for three-second fault test
After the MATLAB and the FE models were validated using the experimental test
results, they were used to estimate the quench behaviour for three seconds. Figure 5.17
shows the predicted current response from the MATLAB model. The current level is
limited from 300Apeak to 40Apeak in a three-second fault condition. Figure 5.18 shows
the predict temperature response from both numerical models. The predicted
Chapter 5: Modelling of SFCL Coil
149
temperature from the MATLAB model with adiabatic boundary conditions is 290.3K,
which would be the highest temperature if the coil quenches uniformly. Using the power
loss density from the MATLAB model, the predict temperature from the FE model with
the coil surrounded by nitrogen is 212K. The predicted highest temperature is much
lower than the melting point of the SFCL coil; this coil therefore can withstand a fault
for three seconds.
-300
-200
-100
0
100
200
300
0 0.5 1 1.5 2 2.5 3
Time (s)
Cu
rren
t (A
)
Figure 5.17 Modelled current response for a three-second fault at 34K with a potential
peak current of 372A
0
50
100
150
200
250
300
0 0.5 1 1.5 2 2.5 3
Time (s)
Tem
per
atu
re (
K)
Matlabadiabatic
FE with heattransfer tonitrogen
Figure 5.18 Temperature response comparison for a three-second fault at 34 K with a
potential peak current of 327A from the MATLAB model and the FE model
Chapter 5: Modelling of SFCL Coil
150
5.5 Conclusions
The MATLAB model successfully predicted the SFCL coil response over a range of
temperatures and fault current levels. This validates the theoretical model developed for
the quenching process of the superconducting coil.
However, there are several limitations in the MATLAB model: firstly, it is not as
accurate for faults in which the superconducting material remains in or near the
transition region; secondly, the MATLAB model including heat transfer to the
surrounding nitrogen requires a long simulation time; and thirdly, the MATLAB model
can only consider heat transfer in one direction, for example, along the coil or into the
nitrogen.
The FE thermal model also showed a reasonable correlation with the experimental test
results. The FE thermal model considered the physical structure and materials
surrounding the SFCL coil during test. It provided a detailed temperature profile around
the coil. It was concluded that the coil winding should be manufactured carefully to
minimise the gap between the coil and former, in order to minimise the coil temperature
rise.
The FE thermal model was 2-D axisymmetric using space average power loss density
values in the coil. It was not possible therefore to investigate other thermal issues such
as localised hot spots.
After the two numerical models were validated with experimental test results, the
MATLAB model was initially used to predict the SFCL behaviour for a three-second
fault and the FE thermal model was then used to predict the temperature profile using
the power loss density from the MATLAB model. These two numerical models would
provide reliable predictions for future designs.
Chapter 6: Operating Actuator for Vacuum Interrupter
151
6 Operating Actuator for Vacuum Interrupter
6.1 Introduction
The advantages of integrating a vacuum interrupter into the resistive SFCL system were
discussed in the previous chapters. It is important to design and build a fast-acting
actuation mechanism to operate the DVS10CB vacuum interrupter, as shown in Figure
6.1: the specification is outlined in Table 6.1.
Figure 6.1 DVS10CB vacuum interrupter
Table 6.1 DVS10CB vacuum interrupter specification
Item Rated value
Maximum operating voltage 1.5kV
Rated AC current 320A
Maximum interrupting current 3.2kA
Contact holding force for 3.2kA 2.0kgf
Length 64mm
Diameter 50mm
Contact stroke 2.0mm
Allowed contact wear 1.0mm
Chapter 6: Operating Actuator for Vacuum Interrupter
152
The operating actuator for the vacuum interrupter has features that are clearly different
from that required for other types of circuit breakers such as air, oil and SF6. A vacuum
interrupter requires a comparatively short contact stroke and is required to compress the
butt contacts together with a considerable force, to prevent the contacts from separating
[94]. The requirements of the operating actuator for this application are summarised
below:
• The opening time should be less than 10ms. Once a fault occurs, the contacts
should be opened before the next current zero-crossing to prevent the SFCL coil
from heating up.
• The operating actuator should provide adequate holding force in the closed
position to reduce the contact resistance.
• The operating actuator should provide latching in the open and closed positions.
• The total movement of the actuator should be 5mm, which is 3mm further than
the stroke of the vacuum interrupter. This could provide extra holding force and
eliminate contact wear loss.
• It should be strong enough to withstand the mechanical stress during opening
and closing.
• The overall design should be of reasonable size and weight. A simple structure
and low cost will be a preference.
After comparing the advantages and disadvantages of different actuator types in section
2.4, the voice-coil type actuator was chosen because it has a simple structure and fast
response. An analytical model was developed first to calculate the approximate
magnetic field levels in the airgap. The current through the actuator coil and the force
produced on the coil were then taken into consideration. After the initial calculation
from the analytical model, a 3-D FE model was built using Vector Fields Opera
software [142]. This model was used to simulate the magnetic field distribution in the
airgap from the permanent magnets. Current flowing through the actuator coil in both
directions was then used to produce a vertical electromagnetic force on the actuator coil.
Magnetic latches were employed to hold the actuator in the open or closed position. The
latching force was also calculated using the FE model.
Chapter 6: Operating Actuator for Vacuum Interrupter
153
The complete operating actuator was designed to avoid contact popping, bounce,
rebound and welding. After the size of all the components was finalised, 2-D and 3-D
designs of the operating actuator were drawn using VariCAD. A prototype operating
actuator was then built by the mechanical workshop.
6.2 Analytical model
A preliminary analytical model was built to determine the magnetic field in the airgap
and the electromagnetic force using a classical magnetic circuit model, which was
simplified by ignoring steel reluctance, leakage flux and fringing effects.
6.2.1 Model description
The voice-coil type actuator is based on the simple principle of the loudspeaker. A light,
hollow coil is suspended in a strong magnetic field to allow free movement along the
axial direction [119, 120]. As soon as current flows in the actuator coil, an
electromagnetic force is produced on the coil.
The actuator was made up of steel blocks, permanent magnets and a copper-wound coil.
It was impossible to obtain an annular permanent magnet of suitable size and with a
radially magnetised field. Four pieces of permanent magnet with an arc surface facing
the centre therefore were used to keep the airgap small and equal. Four steel walls, a
bottom steel block and a steel cylinder inside the actuator coil provided a low reluctance
return path for the magnetic flux produced by the magnets. Four magnets were fixed to
the outer steel walls and the actuator coil placed in the airgap between the magnets and
inner steel cylinder. The geometry of the actuator is shown in Figure 6.2. It should be
pointed out that the two steel walls in the front of the figure are hidden to show the
detail at the middle of the actuator. The blocks in grey and black represent the steel and
permanent magnets respectively, whilst the ring in magenta represents the actuator coil.
N48 neodymium-iron-boron (Nd-Fe-B) rare-earth permanent magnets were sourced
from Arnold Magnetic Technologies Corporation. The remanence and coercivity are
1.4T and 975kA/m respectively [143]. The height of the actuator coil was designed to
be 5mm shorter than the permanent magnets, so that full utilisation of the magnetic field
could be achieved. The height of the permanent magnets was 80mm and the height of
Chapter 6: Operating Actuator for Vacuum Interrupter
154
the actuator coil was 75mm. The diameter of the copper wire including polyurethane
insulation was 0.95mm; the actuator coil was made up of 79 turns.
Figure 6.2 Geometry of the actuator (the two steel walls in the front are not shown)
Figure 6.3 Simplified model of the magnetic circuit
Chapter 6: Operating Actuator for Vacuum Interrupter
155
Figure 6.3 shows a simplified actuator model in the cross section through the centre of
one of the magnets. The approximate magnetic flux density in the airgap could be
calculated using Ampere’s Law along the red line, which denotes the mean flux path of
the magnetic circuit.
6.2.2 Magnetic field in the airgap
The magnetic field in the airgap is produced by the magnets when there is no current
flowing through the actuator coil. The permanent magnets operate in the second
quadrant of the hysteresis loop, which is called the demagnetisation curve.
The magnetic characteristic of one of the permanent magnets can be evaluated by
adding the actuator open-circuit load line to the magnet 2nd quadrant B-H characteristics
(drawn from the origin) with a gradient equal to the negative permeance coefficient
[144-146], as shown in Figure 6.4. The intersection of these two curves, point P0, is
called the operating point. Note, for this design, the permanent magnet 2nd quadrant B-H
characteristic is non-linear at 20oC. This material was chosen because of the high
remanence at this temperature.
Figure 6.4 Actuator permanent magnet operating points on the 2nd quadrant B-H
characteristic (demagnetisation curve) at 20oC
Chapter 6: Operating Actuator for Vacuum Interrupter
156
The operating point of the permanent magnet in the model shown in Figure 6.3 can be
further simplified using the following assumptions: the permeability of steel is assumed
to be infinite and hence the reluctance of the steel can be neglected; and there are no
flux leakage and fringing effects in the magnetic circuit.
Because there is no current flowing in the actuator coil, according to Ampere’s Law:
0m m g gH l H l+ = (6.1)
where: Hm = Permanent magnet magnetic field strength (A/m)
lm = Length of the permanent magnet (m)
Hg = Airgap magnetic field strength (A/m)
lg = Length of the airgap (m)
Based on the assumptions that there is no flux leakage and fringing effects, all the flux
from the permanent magnets goes through the airgap, therefore:
g g m mB A B Aφ = = (6.2)
where: Bg = Magnetic flux density in the airgap (T)
Ag = Area of the airgap (m2)
Bm = Permanent magnet magnetic flux density (T)
Am = Permanent magnet arc surface area (m2)
The permeability of free space is µ0, so that:
0g gB Hµ= (6.3)
Substituting Equations 6.2 and 6.3 into Equation 6.1 yields:
0
0g mm m m
g
l AH l B
Aµ
+ =
(6.4)
Chapter 6: Operating Actuator for Vacuum Interrupter
157
Equation 6.4 can be rewritten as:
0gm
m m p mg m
AlB H K H
l Aµ
= − = −
(6.5)
where: Kp = 0gm
g m
Al
l Aµ
, the permeance coefficient
The expression for the second quadrant linear section is given by [144, 145]:
0m m m rB H Bµ µ= + (6.6)
where: µm = Relative permeability of the permanent magnet
Br = Remanence in the permanent magnet (T)
Combining Equations 6.5 and 6.6 yields:
0m
m m rp
BB B
Kµ µ
= − +
(6.7)
Equation 6.7 can be rearranged as:
01 1
r rm
m g mm
pm g
B BB
l AK l A
µ µµ
= = + +
(6.8)
The arc surface area of the magnet is assumed to be the same as the airgap. Equation 6.8
can be simplified therefore as:
1
rm
gm
m
BB
l
lµ
=
+
(6.9)
Chapter 6: Operating Actuator for Vacuum Interrupter
158
From Equation 6.9, it can be seen that it is necessary to keep /g ml l as low as possible to
achieve a large magnetic flux density in the airgap. The actuator coil was placed in the
airgap requiring a tube to support it. The following factors were considered for the
length of the airgap: the distance from the steel cylinder to the tube, the thickness of the
tube, the diameter of the copper wire for the actuator coil, the thickness of the epoxy to
hold the actuator coil onto the tube and the distance from the epoxy to the permanent
magnets. The length of the airgap was initially set to be 5.8mm. The thickness of the
permanent magnets at the centre was 15.2mm. It was therefore determined using
Equation 6.9 that the magnetic flux density between the magnets and the steel cylinder
in the airgap was approximately 1T.
6.2.3 Effect of the actuator coil current on the electromagnetic force
When current flows through the actuator coil, the magnetic field produced by the
magnets may be increased or reduced, depending on the direction of the magnetic field
produced by the actuator coil. The operating point of the permanent magnets in the
second quadrant will move accordingly. For example, in the position where the
magnetic field produced by the actuator coil is in the same direction as the permanent
magnets, the operating point will move up the curve, as shown by P1 in Figure 6.4, and
if the field is in opposite direction, the operating point will move down the curve, as
shown by P2.
The actuator coil of N turns carrying a current I is then considered using the same
approximation. Ampere’s Law can be expressed as:
g gH l NI= (6.10)
where: N = Turns of the actuator coil
I = Current in the actuator coil (A)
Thus the magnetic flux density produced by the actuator coil in the airgap is given by:
0gg
NIB
lµ= (6.11)
Chapter 6: Operating Actuator for Vacuum Interrupter
159
The actuator coil was made up of 79 turns with an estimated current of 50A. The
maximum flux density produced by the actuator coil in the airgap is therefore 0.85T.
The actuator can be regarded as a linear motor which produces a linear force along its
axial direction. When the actuator coil is stationary, it can be simplified to a resistor and
inductor in series. It is also assumed here that the inductance L remains constant. When
the actuator coil starts to move, a back-emf is produced across the actuator coil, whose
polarity is the reverse of the input voltage. Therefore, when the actuator coil is supplied
from a power supply, the instantaneous current flowing through the actuator coil is
given by [118]:
diV RI L E
dt= + + (6.12)
where: V = Power supply voltage (V)
R = Resistance of the actuator coil (Ω)
L = Inductance of the actuator coil (H)
di
dt = Rate of change of current in the actuator coil (A/s)
E = Back-emf (V)
where:
g mf cE B l v= (6.13)
where: lmf = Length of the actuator coil in the magnetic field (m)
vc = Velocity of the copper coil (m/s)
The electromagnetic force is produced on the actuator when it carries a current in the
magnetic field. The Lorentz force acts along the axial direction to open and close the
vacuum interrupter. This force is linearly proportional to the current and the magnetic
flux density in the airgap:
g mfF B Il= (6.14)
Chapter 6: Operating Actuator for Vacuum Interrupter
160
The length of the actuator coil in the magnetic field is 15.8m and the total length of the
actuator coil is 35m. Therefore, if the actuator coil carries a current of 50A, the total
force on the actuator coil would theoretically be 790N. However, due to flux leakage
and fringing effects, the average magnetic flux density would be lower than 1T, and
therefore the force would be lower than 790N.
6.3 Finite element model
The magnetic field distribution in the actuator is non-linear and three-dimensional,
which makes it difficult to evaluate accurately using analytical methods. The actuator
therefore was analysed using a 3-D FE model. The FE model was used to determine the
steady state magnetic field distribution and the static electromagnetic force.
6.3.1 Model description
The 3-D FE model of the actuator was built and the geometry of the model is shown in
Figure 6.5: on the left is the actuator with the two steel walls at the front removed to
show the actuator coil and the permanent magnets in the middle; on the right is a plan
view of the actuator. Different colours are used to represent different materials. Green
represents steel and purple represents the permanent magnets. The actuator coil was
modelled is a red annular conductor.
Figure 6.5 Geometry of the actuator displayed in Vector Fields Opera: full view
without steel walls at the front (left) and plan view (right)
Chapter 6: Operating Actuator for Vacuum Interrupter
161
Due to the symmetric structure with respect to the YZ and ZX planes, the model was
reduced to a quarter, as shown in Figure 6.6. The mesh and computational time
therefore significantly reduced. It should be noted that symmetry cannot be used on the
actuator coil in the FE model.
Figure 6.6 Geometry of the actuator displayed in Vector Fields Opera (only a quarter is
shown using model symmetry)
The characteristics of the magnetic materials were specified by user-defined B-H
curves. The demagnetisation curve of N48 permanent magnet material was specified, as
shown in Figure 6.7 [143]. The magnetisation direction of the magnet was defined as
pointing perpendicular to the steel wall. A non-linear B-H curve, as shown in Figure
6.8, was used for modelling the EN1A steel blocks [147]. EN1A low carbon mild steel
was chosen to manufacture the steel blocks because it is easy to machine and has a
reasonable permeability.
0
0.5
1
1.5
-1000 -800 -600 -400 -200 0
Magnetic field strength (kA/m)
Mag
net
ic f
lux
den
sity
(T
)
Figure 6.7 N48 permanent magnet normal demagnetisation curve at 20ºC [143]
Chapter 6: Operating Actuator for Vacuum Interrupter
162
0
0.5
1
1.5
2
0 5 10 15 20
Magnetic field strength (kA/m)
Mag
net
ic f
lux
den
sity
(T
)
Figure 6.8 EN1A mild steel B-H curve [147]
6.3.2 Magnetic field distribution in the airgap
The current density for the actuator coil was initially set to be zero to simulate the
magnetic field produced by the permanent magnets.
Figure 6.9 shows a 3-D plot of the flux density distribution produced by the permanent
magnets. The flux density at the corner of the steel blocks where the steel walls and the
cylinder were attached to the bottom steel block is large.
Figure 6.10 shows the flux density at the midway point of the airgap between the steel
cylinder and the magnets. It is clear that the flux density is focused at the four sections
where the permanent magnets are located. The maximum flux density in the airgap is
0.965T, which confirms that the maximum flux density calculated from the analytical
model of 1T is reasonable. The average flux density across the whole airgap surface is
0.45T.
Chapter 6: Operating Actuator for Vacuum Interrupter
163
Figure 6.9 3-D plot of the flux density distribution produced by the magnets
Figure 6.10 Flux density distribution in the airgap produced by the magnets
Chapter 6: Operating Actuator for Vacuum Interrupter
164
Figure 6.11 shows the detailed cross-sectional view of the flux density in the actuator.
The flux vectors (cones) represent the direction of the magnetic field, whilst the size
represents the strength of the magnetic flux density. It is clear that most of the magnetic
flux goes through the steel path and airgap; only a small portion near the ends of the
permanent magnet does not. Figure 6.12 shows the radial flux density in the airgap
along the path noted in Figure 6.11. The path is located at the centre of the arc of one of
the magnets from top to bottom and at the midway point in the airgap between the
magnet and the steel cylinder. The flux density, as expected, is relatively constant at the
centre of the magnet but reduces slowly at both ends due to flux leakage and fringing
effects. The flux density at the centre is 0.965T and the average flux density along the
path is approximately 0.9T.
Figure 6.11 Cross-sectional view of the flux density distribution with vectors produced
by the magnets
Chapter 6: Operating Actuator for Vacuum Interrupter
165
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70 80
Length (mm)
Flu
x d
ensi
ty (
T)
Figure 6.12 Flux density distribution in the airgap produced by the magnets (along the
path noted in Figure 6.11)
Figure 6.13 shows the plan view of the flux density distribution. It can be seen that the
airgap flux density around the magnet circumference is concentrated at the magnets.
Figure 6.14 illustrates the radial magnitude of the flux density along the magnet
circumference at the middle of the airgap.
The path is shown in Figure 6.13, and starts from the middle of one of the magnets and
continues to the middle of the adjacent magnet. The maximum flux density occurs at the
mid-point of the magnets, as expect, whilst the minimum flux density occurs at the mid-
point between the magnets again, as expect. At the edge of the magnets, the flux density
is approximately 0.5T, which is about half of the maximum flux density. The magnetic
field at the edge is reduced because of the flux leakage and fringing effects. The average
flux density along this path is 0.48T.
Chapter 6: Operating Actuator for Vacuum Interrupter
166
Figure 6.13 Plan view of the flux density distribution with vectors produced by the
magnets
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70 80 90 100 110
Length (mm)
Flu
x d
ensi
ty (
T)
Figure 6.14 Flux density distribution in the airgap produced by the magnets (along the
path noted in Figure 6.13)
Chapter 6: Operating Actuator for Vacuum Interrupter
167
6.3.3 Effect of the actuator coil current on the electromagnetic force
The actuator coil was modelled as a solid annular conductor with a cross-sectional area
of 67.5mm2 (height of 75mm and thickness of 0.9mm). It was assumed that the
maximum current flowing through the actuator coil was 50A. The current density of the
conductor was therefore specified as 58.5A/mm2 in the FE model.
The direction of current through the actuator coil was initially set to be anti-clockwise.
The magnetic field in the inner steel cylinder is increased because the field from the
actuator coil is in the same direction as that from the magnets. The actuator coil will be
subjected to a downward force along the negative Z-axis and this provides the force to
open the vacuum interrupter. Figure 6.15 shows the resultant radial flux density
distribution produced by the magnets and the actuator coil. It is clear that the magnetic
field in the steel is significantly increased. The corner between the steel cylinder and the
bottom steel block clearly becomes saturated.
Figure 6.16 shows the flux density in the airgap along the vertical path denoted in
Figure 6.11. At the top of the path the magnetic field is considerably increased because
of the field produced by the actuator coil; however, at the bottom of the path the
magnetic field is slightly reduced because part of the steel blocks are saturated by the
increased magnetic field produced by the actuator coil and the magnetic circuit
reluctance increases. The overall average flux density along the vertical direction is
increased. The actuator coil initially receives a downward force of 882N at the top
position. The force produced for a 5mm displacement varies from 882N to 875N. This
confirms that it is an excellent linear actuator providing nearly constant force along its
whole displacement.
The FE model was resimulated with the current direction changed to clockwise. The
flux density distribution with the actuator coil carrying 50A in the clockwise direction is
shown in Figure 6.17. The radial flux density along the vertical path in the airgap from
top to bottom is shown in Figure 6.18. It is clear that the magnetic field in the airgap at
the top of the path is now significantly reduced by the current in the actuator coil and is
slightly increased at the bottom. There is a vertical force of 547N, which tends to move
Chapter 6: Operating Actuator for Vacuum Interrupter
168
the actuator coil upwards. The contact gap in the vacuum interrupter would be closed by
this force.
Figure 6.15 3-D plot of the flux density distribution produced by the magnets and the
actuator coil carrying 50A in the anti-clockwise direction
0
0.2
0.4
0.6
0.8
1
1.2
0 10 20 30 40 50 60 70 80
Length (mm)
Flu
x d
ensi
ty (
T)
Figure 6.16 Flux density distribution in the airgap produced by the magnets and the
actuator coil carrying 50A in the anti-clockwise direction (along the vertical path)
Chapter 6: Operating Actuator for Vacuum Interrupter
169
Figure 6.17 3-D plot of the flux density distribution produced by the magnets and the
actuator coil carrying 50A in the clockwise direction
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70 80
Length (mm)
Flu
x d
ensi
ty (
T)
Figure 6.18 Flux density distribution in the airgap produced by the magnets and the
actuator coil carrying 50A in the clockwise direction
Chapter 6: Operating Actuator for Vacuum Interrupter
170
-1000
-800
-600
-400
-200
0
200
400
600
800
1000
0 10 20 30 40 50
Current (A)
Fo
rce
(N)
Fc - Fr
Closing force, Fc
Reluctance force, Fr
Fo - Fr
Opening force, Fo
Figure 6.19 Actuator force versus current characteristics
The FE model was simulated with the current in the actuator coil varying from 10A to
50A, in steps of 10A, for both the clockwise and anti-clockwise directions. The force on
the actuator coil is illustrated in Figure 6.19, showing that for the same current, the
magnitude of force to open the actuator, Fo, is higher than that to close the actuator, Fc.
The difference between the opening force and closing force is due to asymmetry in the
structure of the actuator that results in a reluctance force component, Fr, which is also
shown in Figure 6.19. The reluctance force shown in Figure 6.19 is calculated by the FE
model by setting the permanent magnet material property to that of air. For the chosen
coordinate definition, the reluctance force is always along the negative Z-axis, is
independent of the current direction and increases closely to the square of the coil
current, as would be expected.
Chapter 6: Operating Actuator for Vacuum Interrupter
171
6.3.4 Magnetic latch
The actuator does not have an inherent stable position when the power supply to the
actuator coil is removed. It is important that if the contacts are in the open or closed
position, the actuator is safely secured in that position. Magnetic latches at both ends
were employed therefore to hold the actuator in position [118, 148], as shown in Figure
6.20. The top magnetic latch was made from a steel plate, a magnet and a latch steel.
The steel plate was fixed at the top of the outer steel wall, forming part of the low
reluctance path for the top latch. The top latch magnet was placed beneath the steel
plate. The latch steel was then placed between the top latch magnet and the main
magnet, as shown in Figure 6.20. The latching force at the top (closed position) was
provided by the top latch magnet, whilst the latching force at the bottom (open position)
was provided by the main magnet.
Figure 6.20 Geometry of the actuator with magnetic latches (the two steel walls and
latches at the front are not shown) (left) and cross-sectional view (right)
The latch magnetic circuit is complex and non-linear; it was almost impractical
therefore to calculate the flux density distribution and force accurately using an
analytical model. The magnetic latches were added into the FE model to simulate the
magnetic flux density distribution in the airgap between the magnet and latch steel and
to estimate the latching force. It was found that the magnetic field in the airgap between
Chapter 6: Operating Actuator for Vacuum Interrupter
172
the magnet and the latch steel is higher when the latch steel is at the top position. The
latch steel therefore was designed to have a bigger area at the bottom than the top, as
shown in Figure 6.20 (right), to increase the latching force at the bottom position.
The force on the latch steel was calculated using Maxwell stress tensor over the surface
surrounding it [118, 148]. The latching force from one latch at the top was 52N. Two
top latches were employed to keep the actuator mechanically balanced. They would
provide a force of 104N in total, which was more than enough to hold the interrupter in
the closed position. Four sets of latches at the bottom position were used: two of these
contained the top latch magnets. The latching force from those with the top latch
magnets was 14N, whilst for the latches without the top latch magnets it was 24N;
therefore, the total latching force in the open position would be 38N.
Coil current is necessary to release the magnetic latches when operating the vacuum
interrupter. This current creates a force on the actuator coil in a direction opposing to
the latching force. The resultant force on the moving part is therefore the force from the
actuator coil minus the magnetic latching force. The force from the actuator coil has to
be larger than the magnetic latching force to operate the vacuum interrupter.
6.4 Design of the full operating actuator
To design a high performance operating actuator for the vacuum interrupter, many
factors need to be considered such as contact popping, bounce, rebound and welding
[94]. The reasons that cause these problems and how to design the operating actuator to
satisfy the requirements will be discussed in the following sections.
6.4.1 Contact popping
Butt contacts are widely used in vacuum interrupters for low interrupting current levels
(<7kA) [94, 149]; the actual contact area is considerably smaller than the apparent area
of the contacts due to macro-roughened surfaces. The magnetic field of the current
passing through the contacts produces an electromagnetic repulsive force tending to
force the contacts apart, called contact popping. This repulsive force is more significant
when high current flows through the closed contacts. When the contacts separate,
popping would lead to arcing in the limited space inside the vacuum interrupter, and this
Chapter 6: Operating Actuator for Vacuum Interrupter
173
could cause severe erosion of the contact surface. In order to prevent contact separation,
a minimum static contact holding force is usually required. The common way to provide
this force is to place a spring along the driving shaft near the vacuum interrupter. The
spring is pre-compressed (commonly called the wipe spring) in order to provide a
considerable force immediately after the contacts are closed. During normal operation,
the contacts would gradually wear and erode. The compression of the spring and
holding force are then reduced, so the pre-compression of the spring helps minimise this
reduction [94].
The required contact holding force for a DVS10CB vacuum interrupter is 2kgf when the
contacts carry 3.2kA. The wipe spring with a stiffness of 2N/mm therefore is placed
next to the movable contact. The spring is pre-compressed by 10mm, so a force of 20N
is placed on the closed contacts. To provide the holding force reliably, the stroke of the
actuator is designed to be 5mm, which is 3mm further than the stroke of the vacuum
interrupter. An extra force of 6N therefore exists on the contacts when the actuator is in
the closed position.
6.4.2 Contact bounce and rebound
Contact bounce on closing is undesirable because, like popping, it can cause arcing
when the contacts separate. It is necessary to analyse the contact dynamics to avoid
contact bouncing. When closing the vacuum interrupter, the movable contact is
travelling towards the stationary contact at a high velocity. The ideal condition would be
that the two contacts move together at the same velocity after collision and slow down
gradually by transmitting the kinetic energy. To control contact bounce, one should
minimise the kinetic energy that must be dissipated and maximise the rate of
dissipation. This kinetic energy could be transmitted into a supporting structure and
eventually dissipated or stored in other components. A flexible supporting spring can be
placed next to the stationary contact of the vacuum interrupter to minimise bounce [94,
150]. The stiffness of the spring has to comply with the following equation [94]:
2 22 1 2 1
21 10
[(1 / ) 1](1 / )M M M M FK
M v
+ − +< (6.15)
Chapter 6: Operating Actuator for Vacuum Interrupter
174
where: K = Spring stiffness (N/mm)
M1 = Mass of the moving part (kg)
M2 = Mass of the stationary part (kg)
F = Resultant force on the vacuum interrupter in the closed position (N)
v10 = Initial velocity of the moving contact before collision (m/s)
A compression spring with a stiffness of 5N/mm was placed therefore between the
interrupter stationary contact and the supporting plate, forming a flexible support
structure to absorb the kinetic energy and avoid contact bouncing.
Contact rebound on opening is a similar problem to contact bounce. If the rebound
brings the contacts close to each other again, the dielectric recovery of the vacuum
interrupter could be compromised. The movable contact therefore must be stopped and
held at the end of the opening stroke. To ensure the movable contact does not rebound,
the kinetic energy has to be absorbed at impact. Building some dissipative element into
the mechanical stopper is feasible. In this prototype, contact opening rebound is
overcome using the magnetic latches, to provide an adequate holding force and a rubber
pad as a damper.
6.4.3 Contact welding
Arcing caused by contact popping and bouncing may melt the surfaces of the contacts,
and this may form a weld when the contacts come together. Although the expected
interrupting current level is much lower than its maximum value, contact welding is still
unavoidable. It is necessary therefore to build a mechanism to break any contact weld
when opening the vacuum interrupter. A snatch bracket is placed between the actuator
and the wipe spring to fracture contact welding. During the process of closing the
vacuum interrupter, the actuator travels further after the contacts are closed, to compress
the wipe spring and provide the holding force to avoid popping. When the vacuum
interrupter opens, the actuator has to travel back the same distance before the contacts
separate. When the contacts start to separate, a significant amount of kinetic energy
therefore has already accumulated in the moving part of the mechanism. This energy is
used to fracture the welding by snatching the moving contact and pulling it away from
the stationary contact.
Chapter 6: Operating Actuator for Vacuum Interrupter
175
6.5 Design of prototype actuator and interrupter
The component properties for the prototype operating actuator such as material and size
are summarised and listed in Table B.1 in Appendix B. After the size of each
component was finalised, the structure of the prototype unit was designed in VariCAD.
The geometry of the 3-D structure of the vacuum interrupter and operating actuator is
shown in Figure 6.21. The complete prototype unit consisted of an actuator, a snatch
bracket, a driving shaft, a wipe spring and a supporting structure.
The actuator includes the steel blocks and cylinder, permanent magnets, an actuator coil
and magnetic latches. The design of the vacuum interrupter is shown in Figures 6.21 to
6.24. A thin fibreglass tube was used to support the actuator coil and prevent it from
deforming. The actuator coil was wound tightly on the fibreglass tube. The steel
cylinder in the middle acts as a guide for the fibreglass tube to move up and down along
its axial direction. A carbon fibre plate was attached on the fibreglass tube to deliver the
force from the actuator coil. Four latch steel blocks were fixed on the carbon fibre plate
to hold the moving part in the open or closed position. Two steel plates were mounted at
the top edge of the steel walls and the latch magnets were placed beneath them. They
provided the magnetic latch in the closed position. Mechanical stoppers are also
designed at the end of both strokes.
The snatch bracket was made from a hollow aluminium cylinder and was fixed onto the
carbon fibre plate by a plastic screw. A brass bolt was used as a shaft to deliver the
force. A wipe spring was placed surrounding the shaft between the snatch bracket and
the movable contact of the interrupter. A supporting spring was placed between the
stationary contact of the interrupter and the supporting plate. The supporting plate was
made of paxolin, a type of synthetic resin bonded lamination (SRBL). The supporting
plate was fixed to the steel walls by four aluminium studs. The supporting plate has two
functions: firstly, it can hold the supporting spring; and secondly, it can provide
electrical insulation terminals for the actuator coil power supply and the vacuum
interrupter.
Chapter 6: Operating Actuator for Vacuum Interrupter
176
Figure 6.21 3-D structure of the operating actuator
Figure 6.22 Cross-sectional view of the vacuum interrupter actuator
(all dimensions in mm)
Chapter 6: Operating Actuator for Vacuum Interrupter
177
Figure 6.23 Plan view of the vacuum interrupter actuator
(all dimensions in mm)
Figure 6.24 Plan view of the stopper and carbon fibre plate with latch steel
(all dimensions in mm)
Chapter 6: Operating Actuator for Vacuum Interrupter
178
6.6 Construction of prototype operating actuator
6.6.1 Actuator stationary part
N48 permanent magnets of the size listed in Table B.1 were manufactured by Arnold
Magnetic Technologies Corporation. Nd-Fe-B is susceptible to corrosion, so nickel
plated magnets were used to protect them from corrosion. A picture of one of the
permanent magnets is shown in Figure 6.25. It can be seen that it has a surface arc to
keep the airgap radial length constant.
Figure 6.25 N48 Nd-Fe-B permanent magnet
The steel blocks and the cylinder were manufactured from EN1A mild steel because it is
easy to machine. The steel cylinder was bolted at the middle of the bottom steel block.
Each magnet was placed on the steel wall with great care and held in place by a tufnol
square frame. The four steel walls with magnets mounted were then bolted onto the
bottom steel block. The steel frame with the permanent magnets is shown in Figure
6.26.
Chapter 6: Operating Actuator for Vacuum Interrupter
179
Figure 6.26 Steel frame with permanent magnets
6.6.2 Actuator moving part
A carbon fibre plate and fibreglass tube was supplied by Tri-cast Composite Tubes Ltd.
The plate was machined round with a diameter of 156mm and then cut to produce four
flat edges, as shown in Figure 6.24. An annular slot with a depth of 2mm was cut into
the plate to place the fibreglass tube. This would provide more contact surface area
between the plate and tube. AWG19 copper wire with a diameter of 0.912mm was used
for the actuator coil. The temperature rise of the actuator coil was not a problem because
the actuator current was only a pulse of a few tens of milliseconds duration with a
maximum current of 50A. The carbon fibre plate and fibreglass tube were placed on the
lathe and 79 turns of copper wire were wound onto the tube. Two small holes were
drilled in the plate to allow both connection ends of the copper wire to pass through.
Four blocks of latch steel were screwed onto the flat edges on the carbon fibre plate.
The top plate and actuator coil on the fibreglass tube is shown in Figure 6.27.
Figure 6.27 Actuator coil on the fibreglass tube
Chapter 6: Operating Actuator for Vacuum Interrupter
180
6.6.3 Complete prototype
The magnetic latches and stoppers in the closed position were assembled at the top edge
of the steel walls. The snatch bracket, shaft, wipe spring, vacuum interrupter, supporting
spring and plate were assembled in turn afterwards. The ends of the actuator coil were
then connected to the terminals on the supporting plate. The assembled prototype
operating actuator with the vacuum interrupter is shown in Figure 6.28.
Figure 6.28 Operating actuator with vacuum interrupter
Chapter 6: Operating Actuator for Vacuum Interrupter
181
6.7 Conclusions
The structure of the vacuum interrupter actuator was analysed and calculated using the
analytical and the FE models. The analytical model was used to calculate the maximum
magnetic field in the airgap. The actuator was taken as an idealised linear model without
considering the reluctance of the steel, flux leakage and fringing effects.
The FE model was built to analyse the flux density distribution and the static force on
the actuator coil with current flow through it. The flux density distribution from the FE
model demonstrated that the analytical model was reasonable. The simulation shows
that the force to open the vacuum interrupter was higher than to close it. This was
because the reluctance force between the actuator coil and the steel frame has the same
direction as the force to open the interrupter. This satisfied the objective that it is more
important to have a higher force to open the actuator than to close it. The force along the
displacement was found to be nearly constant. Magnetic latches in both the open and
closed positions were also added into the FE model. However, the FE model did not
consider actuator coil movement, back-emf, material hysteresis and eddy currents, but it
still provided useful guidance for the design.
A high performance operating actuator were designed and built. The wipe spring,
supporting spring, stopper with rubber pad, magnetic latches and snatch bracket were
included to overcome contact popping, bounce, rebound and welding.
Chapter 7: Design of the Actuator Control Circuit
182
7 Design of the Actuator Control Circuit
7.1 Introduction
The actuator control circuit was designed for two purposes. Firstly, after the SFCL coil
quenches, the voltage across it will increase rapidly due to the increasing coil resistance.
When the voltage is higher than a pre-set threshold voltage level, an ‘open’ signal must
be automatically generated to trigger the vacuum interrupter to open. Secondly, after the
fault is cleared, a ‘close’ signal needs to be generated to close the vacuum interrupter
once the ‘close’ button is pressed. The force required to open and close the vacuum
interrupter is generated from the electromagnetic force on the actuator coil. The
direction of the magnetic field is unchanged; the force direction therefore is changed by
reversing the polarity of the voltage and current.
7.2 Control circuit
7.2.1 Topology selection
The polarity of output voltage and current can be controlled using a full-bridge DC-DC
converter [151]. A single phase full-bridge DC-DC converter, consisting of two legs,
was chosen. Each leg was made up of two switches and their anti-parallel diodes. Figure
7.1 shows a schematic diagram of the actuator control circuit. MOSFETs (G1 to G4)
were selected because they are suitable for low voltage and low current applications.
Their integral body diodes are labelled D1 to D4. The DC bus capacitor C1 was pre-
charged to a voltage between 50V and 100V by a DC power supply, to provide the
energy required to open and close the vacuum interrupter. Resistor R1 was used to limit
the surge current and protect the capacitor [151].
The operating principle for a full-bridge DC-DC converter is explained below. It should
be noted that the positive direction for the coil voltage and current is defined as shown
in Figure 7.1. When MOSFETs G1 and G4 are turned on, G2 and G3 are turned off, a
Chapter 7: Design of the Actuator Control Circuit
183
positive voltage therefore is imposed on the actuator coil, which causes the current in
the actuator coil to flow in the positive direction. The coil current interacts with the
magnetic field produced by the magnets to produce a downward force and open the
vacuum interrupter. When MOSFETs G1 and G4 are turned off, the current cannot
immediately decay to zero due to the coil inductance. The current continues to flow
through diodes D2 and D3 until it is fully discharged. During this period, a negative
voltage is imposed on the actuator coil through the diodes, which forces the current to
decay to zero. When MOSFETs G2 and G3 are turned on, G1 and G4 are turned off, the
direction of the voltage and current reverses. The coil therefore generates an upward
force to close the vacuum interrupter. When MOSFETs G2 and G3 are turned off, the
current flows through diodes D1 and D4 until it decreases to zero. Controlling the
polarity of the output voltage and current of the full-bridge DC-DC converter therefore
controls the opening and closing of the interrupter.
+C1
G1-G G3-G
R1
Actuator coil
G2-G G4-G
G1-S
G2-S
G3-S
G4-S
G4G2
G3G1
+ -i v
D1
D2
D3
D4
GND
50V
-100
V
Figure 7.1 Schematic diagram of the actuator control circuit
7.2.2 Components selection
The components required for the control circuit include MOSFETs, a capacitor and a
current limiting resistor.
The maximum voltage on the capacitor was 100V. It was necessary to work out the
actuator coil resistance to estimate the maximum current level. The resistance of the
actuator coil is given by:
Chapter 7: Design of the Actuator Control Circuit
184
ac
lR
A
ρ= (7.1)
where: l = Total length of the actuator coil (m)
Aac = Cross-sectional area of the actuator coil (m2)
The total length of the actuator coil was 35m, the cross-sectional area was 0.64mm2, and
the resistivity of copper was 1.68×10-8Ω·m at 20ºC. The resistance of the actuator coil
therefore was 0.92Ω. The total resistance of the actuator coil circuit including the
connection wires was about 1Ω. The maximum current through the actuator coil would
be 100A when it stops moving. MOSFETs were chosen therefore for a rating greater
than 120V and 120A; about 1.2 times the maximum operating levels to guarantee a
voltage and current safety margin [152]. IXFH120N20P N-channel enhancement mode
power MOSFETs with a rating of 200V/120A were chosen for the control circuit [153].
The criteria used to select the capacitor were the voltage level and capacitance. The
maximum voltage on the capacitor was 100V, which required the capacitor to be rated
greater than 120V (using a 20% safety margin). When the capacitor was driving the
actuator coil, the capacitor and actuator coil can be regarded as a simple RC circuit. The
voltage across the capacitor, which is time dependent, can be expressed by using
Kirchhoff's current law:
0c c
ac
dV VC
dt R+ = (7.2)
where: C = Capacitance (F)
Vc = Voltage across the capacitor (V)
Solving Equation 7.2 yields:
0( ) ac
t
R CcV t V e
−= (7.3)
where: V0 = Voltage across the capacitor at time t = 0 (V)
Chapter 7: Design of the Actuator Control Circuit
185
The voltage across the capacitor has an exponential decay, so the worst case situation
was considered to guarantee the capacitor can provide enough current to drive the
actuator coil for a certain period of time, i.e. 10ms. The capacitor was pre-charged with
50V; the requirement was that the voltage across the capacitor was no less than 30V
after discharging through the actuator coil for 10ms. From Equation 7.3 the minimum
capacitance was therefore 23mF. Three 8.25mF/450V aluminium electrolytic capacitors
from SIC-SAFCO were connected together in parallel to produce a total capacitance of
24.75mF.
When the capacitors were fully discharged, the maximum voltage difference between
the power supply and capacitor was 100V. A resistor of 20Ω was chosen to limit the
surge current to 5A.
7.3 Trigger signal and MOSFET drive circuit
7.3.1 Trigger signal circuit
The method to determine whether the SFCL coil has quenched or not, was to compare
the coil voltage with the pre-set threshold voltage level. When the coil is in the
superconducting state, the amplitude of the coil voltage is lower than the pre-set
threshold voltage level. Once a fault occurs, the coil starts to quench and the amplitude
of the voltage across the coil increases quickly and will exceed the pre-set threshold
voltage level, which then generates a signal to open the vacuum interrupter. However,
the voltage across the coil is sinusoidal and therefore an absolute value circuit was
necessary to convert it into a positive output signal.
A precision full-wave rectifier circuit was used, as shown in Figure 7.2 [154]. When the
coil voltage is positive, the output of operational amplifier (op-amp) (U3) is negative, so
diode D5 is reverse biased and diode D6 is forward biased. This closes the feedback
loop around op-amp U3 through resistor R21 and forms an inverting amplifier. Op-amp
U4 sums the negative of twice the output of op-amp U3 and the negative of the input
coil voltage, leaving the output of op-amp U4 the same as the coil voltage. When the
coil voltage is negative, diode D5 is forward biased and diode D6 is reverse biased,
which closes the feedback loop around op-amp U3. Op-amp U4 inverts the coil voltage
Chapter 7: Design of the Actuator Control Circuit
186
resulting in a positive output. Thus, the output of op-amp U4 is a positive voltage that
represents the absolute value of the coil voltage, no matter whether it is positive or
negative.
R13
10K C31220pF
Vcoil
C7
0.1uF+15V
GND
GND-15V
R1820K
D5
1N4148
D61N4148GND R19
10K
R2110K
R17
10K
+15VC8
0.1uF-15V
GND
R20
20K
3
26
1 5
74
U3741
3
26
1 5
74
U4
741
-15V
+15V
Abs_Vcoil
Figure 7.2 Schematic diagram of the precision full-wave rectifier [154]
The circuit to generate the ‘open’ signal for the vacuum interrupter is shown in Figure
7.3. The absolute value of the coil voltage is compared with the pre-set threshold
voltage level by the voltage comparator (U5). The pre-set threshold voltage level, which
is connected to the positive input of voltage comparator U5, could be adjusted from 0V
to 7.2V by a potentiometer. When the coil is superconducting, the absolute coil voltage
is lower than the pre-set threshold voltage level and the output of voltage comparator
U5 is held at a high voltage level by a 10kΩ pull-up resistor. After the coil starts to
quench, the output of voltage comparator U5 transitions to a low voltage level until the
coil voltage exceeds the pre-set threshold voltage level. The output of voltage
comparator U5 returns to the high voltage level when the coil voltage drops below the
pre-set threshold voltage level. The output of voltage comparator U5 is connected to the
trigger input of LM555 timer (U2) through a capacitor. Timer U2 is used to build a
monostable circuit [154, 155]. Normally the trigger input of timer U2 is held at the high
voltage level by a 2.4kΩ pull-up resistor. The output of timer U2 stayed at the low
voltage level. The trigger pulse must be of shorter duration than the intended output
pulse to avoid it being triggered again. When the output of voltage comparator U5
transitions to the low voltage level, the trigger pulse changes to a short pulse through
capacitor (C9). When the trigger input falls below a third of the supply voltage, a pulse
is generated for a certain period of time. The duration of the output pulse is determined
by the time constant of an RC network, which consists of a capacitor (C17) and a
Chapter 7: Design of the Actuator Control Circuit
187
resistor (R9). The output pulse terminates when the charge on capacitor C17 equals two
thirds of the supply voltage. The period of output pulse t is given by:
1.1 9 17t R C= × (7.4)
The initial period for the ‘open’ signal is set at 10ms. The resistor R9 and the capacitor
C17 were chosen to be 91kΩ and 0.1µF. The output pulse width could be lengthened or
shortened during testing by adjusting the values of resistor R9 and capacitor C17.
3
21
84
U5A
LM293R1110K
R10
5.1K
C22
0.1uF R1210K
+15V
C230.1uF
GND
GND
TRIG2
Q3
R4
CVolt5 THR 6
DIS 7
VC
C8
GN
D1
U2
555
OPEN
C150.1uF
GND
R9
91K
+15V+15V
C14
0.1uF+15V GND
+15V
GND
R144.7K
C12
0.1uF+15V GND
C9
0.1uF
R162.4K
C24220pF
R15
1M
Abs_Vcoil
Figure 7.3 Schematic diagram of the ‘open’ signal circuit for the vacuum interrupter
The signal to close the vacuum interrupter is produced by another LM555 timer (U1)
monostable circuit, as shown in Figure 7.4. Normally the trigger input of timer U1 is
held at the high voltage level by a 2.4kΩ pull-up resistor and the output of timer U1
stays at the low voltage level. When the ‘close’ button is pressed, the voltage across the
button becomes low. The low pulse changes to a short low pulse by passing through
capacitor (C6), which triggers timer U1 to produce the output pulse. The duration of the
‘close’ signal was initially set at 10ms, which was the same as the ‘open’ signal. A
ceramic capacitor was placed in parallel with the button to eliminate contact bouncing
in the button.
Chapter 7: Design of the Actuator Control Circuit
188
TRIG2
Q 3R4
CVolt5 THR 6
DIS7
VC
C8
GN
D1
U1
555
CLOSE
C100.1uF
GND
C11
0.1uF
R8
91K
+15V
R72.4K
S1
CLOSE
C4
0.1uF
+15V
GND C6
0.1uF
R210K
C50.1uF
Figure 7.4 Schematic diagram of the ‘close’ signal circuit for the vacuum interrupter
7.3.2 MOSFET drive circuit
The gate drive circuit is an important interface between the control signals and the
power MOSFETs. The gate signal must be positive with respect to the source terminal
in order to switch the MOSFET. It is easy to implement a driver for a low-side
MOSFET, i.e. MOSFET G2 or G4, for example, because its source terminal is
connected to ground. However, the source terminal of a high-side MOSFET, i.e.
MOSFET G1 or G3, for example, can be floating between ground and the DC bus
voltage level depending on the states of the MOSFETs in the same leg.
In order to switch the high-side MOSFET, it is necessary to employ a level shifting
circuit and a floating power supply. Level shifting is used to convert a control signal to
a floating signal. There are four potential options to achieve this [151, 156, 157]:
transformer level shifting, opto-couples level shifting, fibre optic link level shifting and
electronic level shifting. The last approach does not provide isolation, but it is a low
cost solution and can be integrated onto a simple chip. International Rectifier
Corporation has developed a series of gate drive ICs utilising this technique [158]. A
floating supply can be achieved using one of the following methods. The simplest way
is to adopt a transformer isolated supply. The second method is ‘charge-pump supply’,
which overlays the voltage of one supply onto another. It is not commonly used for the
driver supply due to its complexity [159]. The last method is a very common and
popular technique called ‘bootstrap supply’, which is made up of a diode and a storage
capacitor. The diode anode is connected to the same power supply as the low-side driver.
Chapter 7: Design of the Actuator Control Circuit
189
The capacitor is charged through the diode when the low-side MOSFET is turned on.
When the low-side MOSFET is turned off, the high-side driver circuit can be powered
from the storage capacitor.
The MOSFETs in the full-bridge DC-DC converter were driven by two IR2113 IC
drivers [160]. These are high voltage and high speed drivers with independent high-side
and low-side referenced output channels. The bootstrap technique is widely used
together with IR2113 IC drivers for high frequency power converters, where switching
devices in the same leg are turned on in sequence. However, it is not suitable for this
application because the capacitor in the bootstrap technique can only be charged when
the low-side MOSFET is turned on. For the actuator control circuit, the MOSFET pairs
G1 and G4 or G2 and G3 only operate occasionally for 10ms, to open or close the
vacuum interrupter. An isolated high frequency DC converter TMA1515S [161] from
Traco Power therefore was used to provide the floating power supply for the high-side
driver. A 20kΩ dummy load resistor was connected to the output of the TMA1515S to
limit any surge voltages. A MOSFET driver for one leg is shown in Figure 7.5. The
structure of the driver for the other leg was the same except the connections for the
‘open’ and ‘close’ signals were swapped. Pull-down resistors were connected to the
high-side and low-side inputs to keep the inputs at the low voltage level when there was
no signal. Four 20Ω resistors, which were recommended in the IXFH120N20P
datasheet, were used as the driving resistors.
LO 1
COM 2
VCC 3
NC 4
VS 5
VB 6
HO 7
VDD9
HIN10
SD11
LIN12
NC8
NC14
VSS13
U6
IR2113
R265.1K
C18
0.1uF
C13 0.1uF
GND
+15V
OPEN
CLOSE
R22 20
R24 20
GND
G2-G
+15V
+C26 10uF
G1-S
G1-G
C19 0.1uF
+C29 10uF
+C25
10uF
R5 10K
R3 10K
GND
Vcc1
GND2 +Vout 4
-Vout 3U8
TMA1515SGND
+15VG1-S
VB1
VB1
R25 20K
Figure 7.5 Schematic diagram of the MOSFET drive circuit
Chapter 7: Design of the Actuator Control Circuit
190
The trigger signal and MOSFET prototype drive circuit were built on a veroboard, as
shown in Figure 7.6. To assemble the MOSFETs onto a heat sink, a thin layer of
silicone paste was placed at the back of the MOSFETs to provide good thermal contact.
The MOSFETs were then screwed onto the heat sink with insulation cloth placed
between the two surfaces. A picture of the vacuum interrupter actuator and its control
circuit is shown in Figure 7.7.
Figure 7.6 Trigger signal and MOSFET drive circuit
Figure 7.7 Vacuum interrupter with its actuator and control circuit
Chapter 7: Design of the Actuator Control Circuit
191
7.4 Conclusions
A full-bridge DC-DC converter was designed to open the vacuum interrupter
automatically once the SFCL coil quenched and to close the vacuum interrupter after a
fault was cleared.
An absolute value circuit was designed to ensure the coil voltage was positive. This was
then compared with a pre-set threshold voltage level using a simple comparator. Once
the coil quenched, the absolute coil voltage rose above the pre-set threshold voltage
level. This caused the comparator to transition to a low logic level, triggering a LM555
timer monostable circuit to generate a pulse for 10ms. After a fault had cleared, a close
button was pressed to trigger another LM555 timer to produce a ‘close’ signal. These
pulses were sent to two IR2113 IC drivers to drive the corresponding MOSFETs. The
control circuit, trigger signal and drive circuit were designed, built and successfully
tested.
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
192
8 Experimental Investigation of an SFCL Coil with Integrated
Vacuum Interrupter
8.1 Introduction
The vacuum interrupter operating actuator and its control circuit was designed and
constructed as described in the previous chapters. It is necessary to test the operating
actuator separately and validate the results with the design models before mounting it
into the SFCL system. The actuator was initially tested at atmospheric pressure. The
ultimate aim was to integrate the whole unit into the vacuum chamber in the cryostat.
The actuator therefore was assembled into a separate vacuum chamber and tested. After
confirming that the actuator and its control circuit worked correctly in the vacuum
chamber, the vacuum interrupter was then placed into the SFCL system for testing.
Quench tests with and without the vacuum interrupter were compared. Simulated fault
tests were carried out to test the SFCL behaviour during a fault with and without the
vacuum interrupter. A fault was simulated experimentally by manually closing a switch
to short-circuit a load resistor.
8.2 Testing of the vacuum interrupter operating actuator
8.2.1 Actuator magnetic field and static force
The magnetic field of the actuator plays an important role in the behaviour of its
operation. The magnetic flux density in the airgap between the steel cylinder and the
permanent magnets was measured using a Gauss meter. The magnetic flux density at the
middle of the arc surface covering the four permanent magnets varied from 0.95T to
1.04T. The maximum magnetic flux density at each side was slightly different because
the airgap in each direction is not absolutely constant. The magnetic flux density
obtained from the analytical model and the FE model was 1T and 0.965T respectively.
The results from both models are within ±5% of the experimental measurements, which
is a good correlation. The experimental measurements also revealed that the magnetic
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
193
flux density at the middle of the permanent magnets was higher than at both edges,
which was the same as obtained from the FE model results.
The static opening force and closing force were measured using an S-type load cell,
Model 615, from Tedea Huntleigh [162]. The load cell was assembled between the top
plate and the carbon fibre plate, which was attached to the actuator coil. The force from
the actuator coil was transmitted to one end of the load cell through the carbon fibre
plate. The other end of the load cell was fixed to the top plate. The reading of the load
cell was obtained using a Maywood 2000 Digital Indicator, which was calibrated by an
Agilent 34420A NanoVolt/MicroOhm meter.
A series of tests were carried out on the actuator, which was supplied by a DC power
supply for ten seconds. The measured force was recorded as the applied current level
was increased from 10A to 50A. A second series of tests were conducted with the same
current level but in the opposite direction. The measured opening and closing forces
together with the predicted forces from the FE model are shown in Figure 8.1 from
which it is clear that the measured opening and closing forces exhibit a good correlation
with the FE model predicted forces. The measured closing force is approximately 10%
lower than the predicted force, which is within the measurement error band and
considerations associated with the mechanical implementation of the force sensing
transducer (i.e. transducer mass and stiffness).
The force measurement demonstrates that the opening force is higher than the closing
force as a consequence of the actuator reluctance. This additional reluctance force
designed for the opening direction of actuation is a desirable feature in terms of the
project objective where opening the vacuum interrupter in as shorter time period as
possible is a primary performance target.
An indirect method was used to measure the latching force in both the open and closed
positions. The current through the actuator coil was increased from 1A, in steps of 1A,
until the actuator coil started to move. The current level can be converted to force, from
the measured force versus current profile. The critical current that caused the actuator
coil to move was 3A in the open position and 8A in the closed position. The latching
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
194
force in the open and closed positions therefore were about 36N and 96N, respectively,
approximately 8% lower than predicted by the FE model. However, it was enough to
hold the vacuum interrupter in both positions.
0
200
400
600
800
1000
0 10 20 30 40 50
Current (A)
|Fo
rce|
(N
)
Opening force
Measured opening force
Closing force
Measured closing force
Figure 8.1 Actuator opening and closing force versus current
8.2.2 Opening the vacuum interrupter at atmospheric pressure
Opening the vacuum interrupter is more critical than closing. If the vacuum interrupter
could open faster when a fault occurs, less energy would be dissipated in the
superconducting coil and the coil could recover in a shorter period of time after the fault
clears. The opening operation of the vacuum interrupter therefore was carefully
evaluated in this section.
During the opening operation, the actuator was activated using its control circuit. The
voltage across the actuator coil was measured using a Lecroy AP300 differential voltage
probe. The current through the actuator coil was measured using a CP150 current probe
and the displacement of the actuator coil was measured by a DFg5 unguided miniature
linear variable differential transformer (LVDT) from Solartron Metrology. The LVDT
is a type of electric transformer, which is made up of three solenoidal coils around a
tube and a ferromagnetic core [163]. The stroke of the DFg5 LVDT is ±5mm. All the
signals were recorded using an oscilloscope and the data were plotted out in figures
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
195
using MATLAB. The response time constant of the DFg5 LVDT was about 3ms [164],
so the displacement is time-shifted by 3ms in the figures.
The opening operation was tested with the voltage of the storage capacitor charged to
50V. The test was repeated by increasing the capacitor voltage every 10V up to 100V.
The actuator was initially tested at atmospheric pressure. The current duration for
opening the vacuum interrupter was initially set at 10ms. Figure 8.2 shows the actuator
coil voltage, current and linear displacement when the vacuum interrupter is opening
with a capacitor voltage of 100V. The opening operation can be divided into the
following phases:
• 0-t1: The current increases quickly but limited by the inductance of the actuator
coil. When the current is less than 8A, the actuator coil is stationary because the
force on the actuator coil is less than the top latching force.
• t1-t2: Once the current exceeds 8A, the actuator coil starts to move. A back-emf,
which is proportional to velocity, appears across the actuator coil. The current
continues to increase but the rate of change of current reduces due to the back-
emf. The velocity of the actuator coil and back-emf are both increasing at the
same time. After 0.0025s, the current starts to drop slightly as the back-emf
increases.
• t2-t3: When the actuator coil achieves a 3mm displacement, the contacts of the
vacuum interrupter start to separate. They may start to open earlier but it is not
easy to determine the exact point they start to open. The current continues to
drop as the velocity increases. When the required force to open the vacuum
interrupter exceeds the force that the actuator coil can provide, the actuator coil
starts to decelerate and the back-emf starts to decrease. The current in the
actuator coil starts to increase steadily as the back-emf drops. When the current
is at its maximum level, the contacts are fully open and the back-emf across the
coil becomes negligibly small.
• t3-t4: The current drops slightly and becomes constant. The moving part of the
actuator moves slightly further compressing the rubber pad on the stopper. A
nearly constant current flows through the actuator coil after it is stopped and
holds the moving part in the open position.
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
196
• t4-t5: The MOSFETs are turned off after 10ms but the actuator coil current
cannot instantly drop to zero due to the coil inductance. The current therefore
freewheels through the MOSFET diodes. A negative voltage is then applied to
the actuator coil, forcing the current to decrease nearly linearly. After the current
decreases to zero, i.e. the energy stored in the inductance is dissipated, the
voltage across the actuator coil returns to zero.
Figure 8.2 Opening operation of the vacuum interrupter with a capacitor voltage of
100V at atmospheric pressure
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
197
During the 10ms operation, the voltage across the actuator coil is reducing because
energy is dissipated in the actuator coil. It should be pointed out that although the stroke
of the actuator was designed to be 5mm, the real stroke increased to 5.5mm because the
rubber pad on the stopper was compressed.
Figure 8.3 shows a comparison of the opening operation with different capacitor charge
voltages. It can be seen that it takes around 9.7ms, 8.5ms, 7.9ms, 7.3ms, 6.5ms and 6ms
respectively, for capacitor voltages from 50V to 100V to fully open the vacuum
interrupter. It is clear that a higher voltage on the capacitor produces a higher current in
the actuator coil and opens the vacuum interrupter in a shorter period of time.
Figure 8.3 Opening operation of the vacuum interrupter with different capacitor
voltages at atmospheric pressure
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
198
8.2.3 Opening the vacuum interrupter in a vacuum
The ultimate objective of this research project was to integrate the vacuum interrupter
and its actuator into the vacuum vessel within the cryostat. It was impossible to place
the unit directly into the existing cryostat because it was not initially designed for this.
To emulate this, the vacuum interrupter and its actuator were mounted into an external
vacuum chamber.
The actuator was held in place using a steel bracket inside the chamber, as shown in
Figure 8.4 (left). The connections to the vacuum interrupter and its actuator coil were
made of solid copper rods passed through a socket on the wall of the vacuum chamber.
Two copper rods with a diameter of 3mm were used for the actuator coil connections,
whilst another two copper rods with a diameter of 6mm were used for the vacuum
interrupter connections. A wooden block with four holes in the correct position for the
copper rods was manufactured. These four solid copper rods were pulled through the
wooden block and then a copper tube which had the same diameter as the socket on the
vacuum chamber, as shown in Figure 8.5. The wooden block and copper tube were held
together using two clamps. Sealing paste was pushed into the gap between the copper
rods and the copper tube to hold the copper rods in position and left until it cured. The
wooden block was then removed; 3M adhesive epoxy was filled into the rest of the
copper tube and left overnight until it cured to provide a good vacuum seal. Four copper
rods were then held in the correct position to ensured good electrical insulation between
each rod and socket. The bare copper rods were covered by heat shrink sleeving to
provide electrical insulation. The external and internal connectors were connected to
these four copper rods. The external connectors are shown in Figure 8.5.
The LVDT power supply and feedback signal were connected through a signal socket.
The vacuum chamber is shown in Figure 8.4 (right). A vacuum pump unit, which was
the same as the one used for the cryostat, was used to pump the air out from the
chamber. The pressure in the chamber was reduced to 0.014mbar during testing.
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
199
Figure 8.4 Operating actuator in the vacuum chamber (left) and external view of the
vacuum chamber (right)
Figure 8.5 External connectors for the vacuum interrupter and its actuator
The opening test in the vacuum chamber with a capacitor voltage from 50V to 100V, in
steps of 10V, is shown in Figure 8.6. There was a serious problem however when
analysing the displacement of the coil. When the initial voltage on the capacitor was
50V, the actuator coil rebounded by approximately 1mm. When the initial voltage on
the capacitor was 60V and 70V, the actuator rebounded and closed again after it opened.
When the initial voltage on the capacitor was 80V to 100V, the actuator operated
correctly.
The reason for this problem was the lack of air resistance in the vacuum chamber and
the instantaneous velocity of the actuator coil was too high to stop when it collided with
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
200
the stopper. That is why the rebound was more serious with a capacitor voltage of 60V
and 70V than 50V. When the capacitor voltage was increased, the opening velocity was
higher and the actuator opened faster. For example, the opening time is 5ms when the
capacitor voltage is 100V, then the actuator is held in the open position for 5ms before
the current is removed and this help to reduce any rebound. This explains why the
actuator works correctly with a capacitor voltage of 80V to 100V.
Figure 8.6 Opening of the vacuum interrupter with different capacitor voltages in the
vacuum chamber
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
201
The duration of the current through the actuator coil was increased to 15ms to solve the
rebound problem. Figure 8.7 shows the opening of the vacuum interrupter with a current
duration of 15ms. It is clear that the actuator works correctly as expected with a
capacitor voltage between 50V to 100V. The contact rebound problem has been
overcome by increasing the current duration. The duration of the current pulse for
opening the vacuum interrupter therefore was set at 15ms.
Figure 8.7 Opening of the vacuum interrupter with different capacitor voltages in the
vacuum chamber with the current duration increased to 15ms
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
202
8.2.4 Comparison of the tests at atmospheric pressure and in the vacuum
The opening of the vacuum interrupter at atmospheric pressure and in the vacuum with
a capacitor voltage of 100V is compared in Figure 8.8. It presents the actuator coil
voltage, current and linear displacement when the vacuum interrupter is opening.
Figure 8.8 Comparison of the opening of the vacuum interrupter at atmospheric
pressure and in the vacuum
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
203
The opening operation can be separated into the following periods:
• 0-t1: The voltage and current are the same in both conditions because the
actuator coil has not moved.
• t1-t2: The actuator coil starts to move and accelerates faster in the vacuum than
at atmospheric pressure because of the lack of air resistance. In the vacuum, a
higher velocity induces a higher back-emf and hence a lower current in the
actuator coil. The displacement of the actuator coil in the vacuum is therefore
higher.
• t2-t3: In the vacuum, the contacts start to separate when the displacement is
3mm. The coil current in the vacuum reduces quicker compared to atmospheric
pressure due to the higher velocity and hence higher back-emf.
• t3-t4: At atmospheric pressure, the contacts start to separate. In order to provide
the force to open the vacuum interrupter, the velocity of the actuator coil
decreases and the coil current increases steadily. The coil current in the vacuum
increases faster than at atmospheric pressure because the back-emf is decreasing
faster. The contacts are fully opened in the vacuum when a displacement of
5mm has been achieved and the current increases to the maximum level.
• t4-t5: At atmospheric pressure, the contacts are fully open when the current
increases to its maximum level.
• t5-: The current drops to a steady state level before the MOSFETs are turned off.
At atmospheric pressure, the current in the actuator coil lasts for 10ms. In the
vacuum, the current duration increases to 15ms to keep it in the open position
for a longer period of time and reduce any rebound.
It is clear from the displacement that the opening in the vacuum is faster than at
atmospheric pressure. The opening time in the vacuum is approximately 5ms, compared
to 6ms at atmospheric pressure.
8.3 Single-strand SFCL coil with and without the vacuum interrupter
After the operating actuator demonstrated successful opening operation, the vacuum
interrupter was placed into the SFCL system. A resistor was connected in parallel but
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
204
external to the cryostat, to act as the fault limiting resistance once the vacuum
interrupter opened. The purpose of the test was to prove whether the operation of the
vacuum interrupter would prevent the SFCL coil heating up and improving recovery
times.
8.3.1 Test rig diagram
The high-current test circuit was modified to carry out the test with the integrated
vacuum interrupter (Figure 8.9). The single-strand SFCL coil was connected in series
with the vacuum interrupter still mounted in the separate vacuum chamber. The bypass
resistor was placed in parallel with them. The load resistor was 0.45Ω and the bypass
resistor was 0.3Ω. A switch was placed in parallel with the load resistor to simulate a
fault by manually closing the switch and short-circuiting the load resistor.
Figure 8.9 Schematic of the high-current test circuit with the vacuum interrupter
8.3.2 Quench test
A quench test was carried out to compare the differences between the SFCL system
with and without the vacuum interrupter. A quench test without the vacuum interrupter
was carried out initially. The actuator control circuit was disabled by discharging the
storage capacitor and turning off the power supply. The amplitude of voltage across the
coil was approximately 0.2V with normal current passing through it. The pre-set
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
205
threshold voltage level therefore was set at 2V during the test. The quench test was
repeated with the vacuum interrupter and the capacitor charged to 100V.
The fault current, coil current, coil voltage and vacuum interrupter voltage during both
quench tests are shown in Figure 8.10. When the vacuum interrupter does not operate, it
is clear that the SFCL coil quenches and limits the fault current to 240Apeak in the first
quarter-cycle and then reduces the current to 160Apeak over the next two cycles. The coil
current is lower than the fault current because after the coil quenches, the bypass
resistor shares some of the fault current.
The voltage across the coil increases after the coil quenches. When the coil voltage is
greater than the pre-set threshold voltage level of 2V, the actuator control circuit
triggers the actuator and the actuator coil starts to move. The vacuum interrupter
contacts start to separate 2.5ms later. After the contacts separate, an arc is drawn
between the contacts inside the vacuum interrupter. The coil voltage suddenly reduces
due to this arc voltage. The coil current and voltage continue to reduce and the arc
extinguishes naturally at the next coil current zero-crossing. The current through the coil
then becomes practically insignificant and the fault current is fully diverted into the
bypass resistor.
The voltage across the vacuum interrupter is then determined by the voltage across the
resistor. The measured arc voltage of the vacuum interrupter was about 12V. This
appeared reasonable because the arc voltage of the vacuum interrupter is made up of
metal vapour and electrons coming from the electrodes. The arc voltage depends on the
contact material and is usually below 80V [94].
Figure 8.11 presents the temperature rise of the coil in a quench test with and without
the vacuum interrupter. When the vacuum interrupter is disabled, 32.3J energy is
dissipated in the coil for two cycles. The temperature rise of the SFCL coil is 13K and
takes approximately 55s to recover to the operating temperature of 34K. When the
vacuum interrupter operates, the energy dissipated in the coil is reduced to 2.7J, which
is approximately 8% of the value without the vacuum interrupter. There is no
discernible temperature rise in the coil. This demonstrates that the energy dissipated in
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
206
the coil is significantly reduced by opening the vacuum interrupter as soon as possible
and as a result, the temperature rise of the coil is practically eliminated.
The SFCL coil has demonstrated successful quench behaviour, and shown that MgB2
wire is a potential candidate material for future SFCL development. With the integrated
vacuum interrupter, the fault current was fully diverted into the bypass resistor at the
next current zero-crossing after the fault was imposed. This significantly reduced the
energy dissipated, temperature rise and recovery time of the coil.
Figure 8.10 Quench test with a potential peak current of 324A with and without the
vacuum interrupter
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
207
30
35
40
45
50
0 20 40 60 80
Time (s)
Tem
per
atu
re (
K)
SFCLwithout VI
SFCL withVI
Figure 8.11 Temperature rise of the coil during a quench test with a potential peak
current of 324A with and without the vacuum interrupter
8.3.3 Simulated fault test
After operation of the SFCL with the vacuum interrupter was demonstrated successfully,
a simulated fault test was carried out. The number of AC cycles supplied was set to be
20 cycles in the LabVIEW control program. Before the test, the switch in parallel with
the load resistor was opened so that normal current flowed in the circuit and the SFCL
coil showed no measurable resistance. The switch was then manually closed to short-
circuit the load resistor and simulate a fault.
The operation of the SFCL coil without the vacuum interrupter was tested initially.
Figure 8.12 shows the fault current, coil current, coil voltage and vacuum interrupter
voltage from the simulated fault test. Normal operating current was 50Apeak and the
switch was manually turned on at approximately 0.25s. The current increases rapidly
and makes the coil quench. In this test it is clear that the SFCL coil quenches and limits
the current to 225Apeak in the first quarter-cycle, then further reduces the current to
75Apeak over the next seven cycles. The voltage across the coil increases after quenching
and 81.7J energy was dissipated in the coil.
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
208
Figure 8.12 Simulated fault test without the vacuum interrupter
Figure 8.13 shows the result from the simulated fault test with the vacuum interrupter
connected. Normal operating current was 50Apeak and the switch was manually closed at
approximately 0.18s to simulate the fault. The coil quenches due to the rising fault
current and the coil voltage increases over the pre-set threshold voltage level of 2V,
which triggers the actuator to open the vacuum interrupter. The vacuum interrupter
starts to open approximately 2.5ms later and an arc is drawn between the contacts. The
arc extinguishes naturally at the next coil current zero-crossing. The bypass resistor then
works as a limiting resistor because the fault current is fully diverted. The energy
dissipated in the coil is reduced to 4.5J, which is approximately 5.5% of the value
without the vacuum interrupter.
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
209
Figure 8.13 Simulated fault test with the vacuum interrupter
Chapter 8: Experimental Investigation of an SFCL Coil with Integrated Vacuum Interrupter
210
8.4 Conclusions
The actuator static force profile was measured using a load cell and was close to that
predicted by the FE model. This demonstrated that the FE model could provide accurate
prediction for the actuator force.
The vacuum interrupter actuator and its control circuit demonstrated successful
operation at atmospheric pressure. The ultimate aim was to integrate the actuator with
the interrupter into the vacuum vessel inside the cryostat. For practical reasons they
were assembled into a separate vacuum chamber for testing. The opening velocity of the
actuator in the vacuum was higher than at atmospheric pressure due to the lack of air
resistance. One operational issue was found during testing in the vacuum: the actuator
rebounded into the closed position due to the high instantaneous velocity. This problem
was solved by increasing the current duration to 15ms, which therefore held the actuator
in the open position longer.
The quench test and simulated fault test were carried out to compare operation with and
without the vacuum interrupter. The SFCL coil limited the fault current to a safe level as
shown in previous tests. The fast-acting actuator opened the vacuum interrupter before
the next current zero-crossing. The fault current was then diverted into the bypass
resistor and the current through the SFCL coil became practically zero. The energy
dissipated in the coil was significantly reduced with the vacuum interrupter. The SFCL
coil with the integrated vacuum interrupter demonstrated successful operation. This
system represents one method of reducing the temperature rise in the resistive SFCL
and hence improves the recovery times.
Chapter 9: Conclusions and Further Work
211
9 Conclusions and Further Research
9.1 Conclusions
This chapter details the contributions from this research project and summarises
conclusions drawn from the work described in the preceding chapters. Further areas of
research are then discussed. All of the aims and objectives outlined in chapter 1 have
been achieved:
• To investigate the behaviour of single-strand and three-strand MgB2 SFCL coils.
• To develop MATLAB and FE models of the single-strand SFCL coil and
compare the simulation results with the experimental test results.
• To design and manufacture a fast-acting operating actuator for a vacuum
interrupter.
• To design and manufacture a vacuum interrupter actuator control circuit.
• To test the vacuum interrupter operating actuator and validate the design process.
• To investigate the behaviour of the SFCL coil with and without an integrated
vacuum interrupter and compare the results.
The concept of the SFCL with an integrated vacuum interrupter is potentially a solution
for a low-cost and reliable SFCL operating in power network applications.
This research project has demonstrated that a resistive SFCL with an integrated vacuum
interrupter using round MgB2 wire shows significant promise for further research and
product development.
The inductances of the single-strand and three-strand SFCL coils were relatively small
but finite due to the interleaved coil design. Additionally, it was confirmed that both
coils were inductance dominated in the superconducting state but the resistance
increased quickly and became dominant after the coil started to quench.
Chapter 9: Conclusions and Further Work
212
Both SFCL coils showed reliable and repeatable current-limiting properties in short
cycle and long duration quench tests. In addition, both SFCL coils showed that the
quench current increased nearly linearly as the temperature was reduced from 34K to
30K. The quench current density of the single-strand coil was about 16.7% higher than
the three-strand coil at 34K with a self-field of 50Hz but in general this was considered
acceptable.
The single-strand coil operated with 200Apeak current for one hour with an observed
temperature rise of less than 0.1K, which indicated that the thermal stability of the
SFCL coil was good and that AC losses could be removed by the cryocooler.
Tests on the three-strand SFCL coil showed that each of the three wire strands shared
the current equally and demonstrated closely identical responses during normal
operation and fault conditions. Multiple strands of MgB2 wire showed potential to be a
practical method to scale-up the current level for SFCL applications.
During a quench test with a potential peak current of 372A for one second, a
temperature of 107.2K was measured on the single-strand SFCL coil, which resulted in
the coil taking more than two minutes to recover to the superconducting state.
A MATLAB model of the single-strand SFCL coil was developed to understand the
MgB2 wire quench process, which could be used to predict the SFCL behaviour in
power system analysis and inform future designs. The SFCL electrical and thermal
response from 34K to 30K was simulated and a good correlation was demonstrated with
the experimental test results.
An FE thermal model was developed to estimate more accurate temperature
distributions for the SFCL coil. The FE model predicted an average temperature of
69.2K to 122.7K in the SFCL coil for a one-second fault and with a gap between the
coil and the former varying between 0mm to 1.5mm. This suggested that the coil
winding should be manufactured carefully to minimise the gap between the coil and
former, in order to minimise the coil temperature rise.
Chapter 9: Conclusions and Further Work
213
The MATLAB and FE thermal models were used to predict the temperature profile of
the single-stand SFCL coil during a three-second fault. A temperature of 212K to 290K
in the coil was predicted, which demonstrated that the SFCL coil could withstand a
three-second fault test.
An analytical model and an FE model were developed to predict the magnetic field in
the airgap and the force on the coil of the voice-coil type actuator. The FE model
suggested that the force along the whole displacement was nearly constant. The actuator
static force measured by a load cell was identical to the force predicted by the FE model.
Both the FE model and measurement showed that the opening force was higher than the
closing force due to the reluctance force (between the actuator coil and the steel frame),
which caused by asymmetry in the structure of the actuator.
The voice-coil type actuator and its control circuit demonstrated successful operation at
atmospheric pressure. They were assembled into a separate vacuum chamber for testing.
The opening velocity in the vacuum was higher than at atmospheric pressure due to the
lack of air resistance.
The SFCL with an integrated vacuum interrupter demonstrated successful operation.
The fast-acting actuator opened the vacuum interrupter before the first half-cycle of the
fault current during the quench and simulated fault tests. The arc in the vacuum
interrupter extinguished at the current zero-crossing and interrupted the current. The
fault current was then diverted into a bypass resistor and the current in the SFCL coil
became negligible. The energy dissipated in the coil and recovery times were
significantly reduced. The SFCL coil with an integrated vacuum interrupter shows good
potential for further research and product development.
Chapter 9: Conclusions and Further Work
214
9.2 Further research
The three-strand SFCL coil demonstrated successful operation with the current shared
equally by each strand. More strands of parallel wires are worthy of being investigated
to determine whether there is a limitation to the number to multiple wire strands that can
be used.
In this research project, the fault detection strategy of an SFCL with an integrated
vacuum interrupter was achieved by comparing the SFCL coil voltage with a pre-set
threshold voltage level. Another method to detect a fault is to monitor the phase shift
between the current through the SFCL coil and the voltage across it. The combination of
these two detection methods would further improve the accuracy and reliability of
detecting a fault condition.
The SFCL with an integrated vacuum interrupter operated automatically when a fault
condition occurred. After the fault was cleared, the vacuum interrupter was closed
manually. A reclosing control scheme integrated into the grid protection systems is a
potential area of further work.
The SFCL with an integrated vacuum interrupter concept has been successfully tested at
low voltage and current levels. A prototype for an 11kV/33kV distribution network
could be designed and manufactured to further validate the concept.
References
215
References
[1] S. Eckroad, "Superconducting Fault Current Limiters," Electric Power Research Institute (EPRI), 1017793, 2009.
[2] M. Noe and M. Steurer, "High-temperature Superconductor Fault Current Limiters: Concepts, Applications, and Development Status," Superconductor Science and Technology, vol. 20, pp. R15-R29, 2007.
[3] S. Eckroad, "Survey of Fault Current Limiter (FCL) Technologies," Electric Power Research Institute (EPRI), 1010760, 2005.
[4] J. Bock, F. Breuer, H. Walter, S. Elschner, M. Kleimaier, R. Kreutz, and M. Noe, "CURL 10: Development and Field-test of a 10kV/10MVA Resistive Current Limiter based on Bulk MCP-BSCCO 2212," IEEE Transactions on Applied Superconductivity, vol. 15, pp. 1955-1960, 2005.
[5] H. J. Boenig and N. M. Los Alamos, "Solid-state Circuit Breaker with Current Limiting Characteristic using a Superconducting Coil," United States Patent: 06/408108, pp. 1-11, 1984.
[6] Y. Xin, W. Gong, X. Niu, Z. Cao, H. Xi, J. Zhang, Y. Wang, B. Tian, and B. Hou, "Development of Superconducting Fault Current Limiters," in International Conference on Power System Technology, 2006, pp. 1-5.
[7] T. Matsumura, T. Uchii, and Y. Yokomizu, "Development of Flux-lock-type Fault Current Limiter with High-Tc Superconducting Element," IEEE Transactions on Applied Superconductivity, vol. 7, pp. 1001-1004, 1997.
[8] S. Elschner, A. Kudymow, S. Fink, W. Goldacker, F. Grilli, C. Schacherer, A. Hobl, J. Bock, and M. Noe, "ENSYSTROB - Resistive Fault Current Limiter based on Coated Conductors for Medium Voltage Application," IEEE Transactions on Applied Superconductivity, vol. 21, pp. 1209-1212, 2011.
[9] A. P. Malozemoff, S. Fleshler, M. Rupich, C. Thieme, X. Li, W. Zhang, A. Otto, J. Maguire, D. Folts, J. Yuan, H. P. Kraemer, W. Schmidt, M. Wohlfart, and H. W. Neumueller, "Progress in High Temperature Superconductor Coated Conductors and their Applications," Superconductor Science and Technology, vol. 21, pp. 1-7, 2008.
[10] L. Martini, "Superconducting Fault Current Limiter Applications," in Symposium on Superconducting Devices for Wind Energy, Spain, 2011, pp. 1-59.
[11] X. Yuan, K. Tekletsadik, L. Kovalsky, J. Bock, F. Breuer, and S. Elschner, "Proof-of-concept Prototype Test Results of a Superconducting Fault Current Limiter for Transmission-level Applications," IEEE Transactions on Applied Superconductivity, vol. 15, pp. 1982-1985, 2005.
References
216
[12] L. Ye, M. Majoros, A. M. Campbell, T. A. Coombs, S. Harrison, P. Sargent, M. Haslett, and M. Husband, "MgB2 Sample Tests for Possible Application of Superconducting Fault Current Limiters," IEEE Transaction on Applied Superconductivity, vol. 17, pp. 2826-2829, 2007.
[13] L. Ye, M. Majoros, A. M. Campbell, T. A. Coombs, D. Astill, S. Harrison, M. Husband, M. Rindfleisch, and M. Tomsic, "Experimental Studies of the Quench Behaviour of MgB2 Superconducting Wires for Fault Current Limiter Applications," Superconductor Science and Technology, vol. 20, pp. 621-628, 2007.
[14] M. Majoros, L. Ye, A. M. Campbell, T. A. Coombs, A. V. Velichko, D. M. Astill, P. Sargent, M. Haslett, M. D. Sumption, E. W. Collings, M. Tomsic, S. Harrison, and M. Husband, "Fault Current Limiting Properties of MgB2 Superconducting Wires," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 1764-1767, 2007.
[15] A. Oliver, A. C. Smith, M. Husband, M. Bailey, and Y. Feng, "Assessment of Small Bend Diameter Magnesium Diboride Wire for a Superconducting Fault Current Limiter Application," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1942-1945, 2009.
[16] M. Tomsic, M. Rindfleisch, J. Yue, K. McFadden, D. Doll, J. Phillips, M. D. Sumption, M. Bhatia, S. Bohnenstiehl, and E. W. Collings, "Development of Magnesium Diboride (MgB2) Wires and Magnets using in situ Strand Fabrication Method," Physica C: Superconductivity, vol. 456, pp. 203-208, 2007.
[17] A. Oliver, "Superconducting Fault Current Limiter using Magnesium Diboride," Doctor of Philosophy Thesis, Faculty of Engineering and Physical Sciences, The University of Manchester, 2008.
[18] M. N. Wilson, Superconducting Magnets. Oxford: Clarendon Press, 1983.
[19] F. Moriconi, F. De La Rosa, F. Darmann, A. Nelson, and L. Masur, "Development and Deployment of Saturated-core Fault Current Limiters in Distribution and Transmission Substations," IEEE Transactions on Applied Superconductivity, vol. 21, pp. 1288-1293, 2011.
[20] H. W. Neumueller, W. Schmidt, H. P. Kraemer, A. Otto, J. Maguire, Y. Jie, D. Folts, W. Romanosky, B. Gamble, D. Madura, A. P. Malozemoff, N. Lallouet, S. P. Ashworth, J. O. Willis, and S. Ahmed, "Development of Resistive Fault Current Limiters based on YBCO Coated Conductors," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1950-1955, 2009.
[21] M. Endo, T. Hori, K. Koyama, I. Yamaguchi, K. Arai, K. Kaiho, and S. Yanabu, "Operating Characteristics of Superconducting Fault Current Limiter using 24kV Vacuum Interrupter Driven by Electromagnetic Repulsion Switch," Journal of Physics: Conference Series, vol. 97, pp. 1-5, 2008.
[22] I. Shimizu, Y. Naito, I. Yamaguchi, K. Kaiho, H. Mizoguchi, and S. Yanabu, "Operation of Superconducting Fault Current Limiter Using Vacuum Interrupter Driven by Electromagnetic Repulsion Force for Commutating Switch," Electrical Engineering in Japan, vol. 164, pp. 52-61, 2008.
References
217
[23] A. C. Rose-Innes and E. H. Rhoderick, Introduction to Superconductivity. Oxford; New York: Pergamon Press, 1978.
[24] B. Seeber, Handbook of Applied Superconductivity, vol. 1. Bristol; Philadelphia: Institute of Physics Publishing, 1998.
[25] J. D. Doss, Engineer's Guide to High-temperature Superconductivity. New York; Chichester: John Wiley and Sons, 1989.
[26] H. Maeda, Y. Tanaka, M. Fukutomi, and T. Asano, "A New High-Tc Oxide Superconductor without a Rare Earth Element," Japanese Journal of Applied Physics, vol. 27, pp. L209-L210, 1988.
[27] Superconductor Technologies, "HTS Wire Overview," http://www.suptech.com/hts_wire_overview.php, Accessed on: 19/09/2011.
[28] M. Li, B. Ma, R. E. Koritala, B. L. Fisher, K. Venkataraman, and U. Balachandran, "Pulsed Laser Deposition of YBCO Thin Films on IBAD-YSZ Substrates," Superconductor Science and Technology, vol. 16, pp. 105-109, 2003.
[29] M. S. Hatzistergos, H. Efstathiadis, J. L. Reeves, V. Selvamanickam, L. P. Allen, E. Lifshin, and P. Haldar, "Microstructural and Compositional Analysis of YBa2Cu3O7-δ Films Grown by MOCVD before and after GCIB Smoothing," Physica C: Superconductivity, vol. 405, pp. 179-186, 2004.
[30] SuperPower, "2G HTS Wire," http://www.superpower-inc.com/content/2g-hts-wire, Accessed on: 19/09/11.
[31] X. Xiong, K. P. Lenseth, J. L. Reeves, A. Rar, Y. Qiao, R. M. Schmidt, Y. Chen, Y. Li, Y. Y. Xie, and V. Selvamanickam, "High Throughput Processing of Long-length IBAD MgO and Epi-buffer Templates at SuperPower," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 3375-3378, 2007.
[32] M. W. Rupich, U. Schoop, D. T. Verebelyi, C. L. H. Thieme, D. Buczek, X. Li, W. Zhang, T. Kodenkandath, Y. Huang, E. Siegal, W. Carter, N. Nguyen, J. Schreiber, M. Prasova, J. Lynch, D. Tucker, R. Harnois, C. King, and D. Aized, "The Development of Second Generation HTS Wire at American Superconductor," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 3379-3382, 2007.
[33] X. Li, M. W. Rupich, C. L. H. Thieme, M. Teplitsky, S. Sathyamurthy, E. Thompson, D. Buczek, J. Schreiber, K. DeMoranville, J. Lynch, J. Inch, D. Tucker, R. Savoy, and S. Fleshler, "The Development of Second Generation HTS Wire at American Superconductor," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 3231-3235, 2009.
[34] American Superconductor, "Amperium™ Wire," http://www.amsc.com/products/amperiumwire/, Accessed on: 20/09/2011.
[35] J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani, and J. Akimitsu, "Superconductivity at 39K in Magnesium Diboride," Nature, vol. 410, pp. 63-64, 2001.
[36] E. W. Collings, M. D. Sumption, M. Bhatia, M. A. Susner, and S. D. Bohnenstiehl, "Prospects for Improving the Intrinsic and Extrinsic Properties of
References
218
Magnesium Diboride Superconducting Strands," Superconductor Science and Technology, vol. 21, pp. 1-14, 2008.
[37] M. Tomsic, M. Rindfleisch, J. Yue, K. McFadden, and J. Phillips, "Overview of MgB2 Superconductor Applications," International Journal of Applied Ceramic Technology, vol. 4, pp. 250-259, 2007.
[38] M. S. A. Hossain, C. Senatore, R. Flukiger, M. A. Rindfleisch, M. J. Tomsic, J. H. Kim, and S. X. Dou, "The Enhanced Jc and Birr of in situ MgB2 Wires and Tapes Alloyed with C4H6O5 (Malic Acid) after Cold High Pressure Densification," Superconductor Science and Technology, vol. 22, pp. 1-8, 2009.
[39] L. Ye, M. Majoros, A. M. Campbell, T. A. Coombs, S. Harrison, P. Sargent, M. Haslett, and M. Husband, "Investigations of Current Limiting Properties of the MgB2 Wires Subjected to Pulse Overcurrents in the Benchtop Tester," Superconductor Science and Technology, vol. 20, pp. 320-326, 2007.
[40] M. Steurer and F. Frohlich, "Current Limiters - State of the Art," in Fourth Workshop and Conference on EHV Technology, Bangalore, 1998, pp. 1-7.
[41] K. Kunde, M. Kleimaier, and L. Klingbeil, "Integration of Fast Arcing Electronic Fault Current Limiters (EFCL) in Medium-voltage System," in 17th International Conference on Electricity Distribution, Barcelona, 2003.
[42] P. Duggan, "Technical and Economic Evaluation of a Solid State Current Limiter," Electric Power Research Institute (EPRI), 1001816, 2002.
[43] M. Steurer, K. Frohlich, W. Holaus, and K. Kaltenegger, "A Novel Hybrid Current-limiting Circuit Breaker for Medium Voltage: Principle and Test Results," IEEE Transactions on Power Delivery, vol. 18, pp. 460-467, 2003.
[44] C. S. Chang and P. C. Loh, "Designs Synthesis of Resonant Fault Current Limiter for Voltage Sag Mitigation and Current Limitation," in IEEE Power Engineering Society Winter Meeting, 2000, pp. 2482-2487.
[45] G. G. Karady, "Principles of Fault Current Limitation by a Resonant LC Circuit," IEE Proceedings C: Generation, Transmission and Distribution, vol. 139, pp. 1-6, 1992.
[46] H. Wang, H. Xi, and G. Tang, "Research of Main Circuit on the Series Resonance Fault Current Limiter," in International Conference on Power System Technology, 2006, pp. 1-5.
[47] J. Skindhoj, J. Glatz-Reichenbach, and R. Strumpler, "Repetitive Current Limiter based on Polymer PTC Resistor," IEEE Transactions on Power Delivery, vol. 13, pp. 489-494, 1998.
[48] R. Strumpler, J. Skindhoj, J. Glatz-Reichenbach, J. H. W. Kuhlefelt, and F. Perdoncin, "Novel Medium Voltage Fault Current Limiter based on Polymer PTC Resistors," IEEE Transactions on Power Delivery, vol. 14, pp. 425-430, 1999.
[49] J. L. Rasolonjanahary, J. P. Sturgess, E. F. H. Chong, A. E. Baker, and C. L. Sasse, "Design and Construction of a Magnetic Fault Current Limiter," in 3rd IET International Conference on Power Electronics, Machines and Drives, 2006, pp. 681-685.
References
219
[50] S. Elschner, F. Breuer, M. Noe, T. Rettelbach, H. Walter, and J. Bock, "Manufacturing and Testing of MCP 2212 Bifilar Coils for a 10MVA Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 13, pp. 1980-1983, 2003.
[51] J. Bock, F. Breuer, H. Walter, M. Noe, R. Kreutz, M. Kleimaier, K. H. Weck, and S. Elschner, "Development and Successful Testing of MCP BSCCO-2212 Components for a 10MVA Resistive Superconducting Fault Current Limiter," Superconductor Science and Technology, vol. 17, pp. S122-S126, 2004.
[52] M. Noe, A. Kudymow, S. Fink, S. Elschner, F. Breuer, J. Bock, H. Walter, M. Kleimaier, K. H. Weck, C. Neumann, F. Merschel, B. Heyder, U. Schwing, C. Frohne, K. Schippl, and M. Stemmle, "Conceptual Design of a 110kV Resistive Superconducting Fault Current Limiter using MCP-BSCCO 2212 Bulk Material," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 1784-1787, 2007.
[53] M. Noe, J. Bock, A. Hobl, and J. Schramm, "Superconducting Fault Current Limiters - Latest Developments at Nexans SuperConductors," in 10th EPRI Superconductivity Conference, 2011.
[54] S. Eckroad, "Superconducting Power Equipment," Electric Power Research Institute (EPRI), 1019995, 2010.
[55] S. Eckroad, "Superconducting Power Equipment," Electric Power Research Institute (EPRI), 1021890, 2011.
[56] L. Kovalsky, Y. Xing, K. Tekletsadik, A. Keri, J. Bock, and F. Breuer, "Applications of Superconducting Fault Current Limiters in Electric Power Transmission Systems," IEEE Transactions on Applied Superconductivity, vol. 15, pp. 2130-2133, 2005.
[57] Y. Xie, K. Tekletsadik, D. Hazelton, and V. Selvamanickam, "Second Generation High-temperature Superconducting Wires for Fault Current Limiter Applications," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 1981-1985, 2007.
[58] J. C. H. Llambes, "Recent Development of SFCL in the USA," in 23rd International Superconductivity Symposium, Japan, 2010.
[59] R. B. Dalessandro, M. Bocchi, V. Rossi, and L. F. Martini, "Test Results on 500kVA-class MgB2-based Fault Current Limiter Prototypes," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 1776-1779, 2007.
[60] L. Martini, M. Bocchi, R. Brambilla, R. Dalessandro, and C. Ravetta, "Design and Development of 15MVA Class Fault Current Limiter for Distribution Systems," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1855-1858, 2009.
[61] B. Gromoll, G. Ries, W. Schmidt, H. P. Kraemer, B. Seebacher, B. Utz, R. Nies, H. W. Neumueller, E. Baltzer, S. Fischer, and B. Heismann, "Resistive Fault Current Limiters with YBCO Films 100kVA Functional Model," IEEE Transactions on Applied Superconductivity, vol. 9, pp. 656-659, 1999.
[62] W. Schimdt, H. P. Kraemer, H. W. Neumueller, U. Schoop, D. Verebelyi, and A. P. Malozemoff, "Investigation of YBCO Coated Conductors for Fault Current
References
220
Limiter Applications," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 3471-3474, 2007.
[63] I. Shimizu, Y. Naito, I. Yamaguchi, K. Kaiho, and S. Yanabu, "Application Study of a High-temperature Superconducting Fault Current Limiter for Electric Power System," Electrical Engineering in Japan, vol. 155, pp. 20-29, 2006.
[64] T. Hori, A. Otani, K. Kaiho, I. Yamaguchi, M. Morita, and S. Yanabu, "Study of Superconducting Fault Current Limiter using Vacuum Interrupter Driven by Electromagnetic Repulsion Force for Commutating Switch," IEEE Transactions on Applied Superconductivity, vol. 16, pp. 1999-2004, 2006.
[65] T. Koyama and S. Yanabu, "Study and Development of Superconducting Fault Current Limiter with High Speed Reclosing," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1868-1871, 2009.
[66] T. Koyama and S. Yanabu, "Adaptation of Superconducting Fault Current Limiter to High-speed Reclosing," Physica C: Superconductivity, vol. 469, pp. 1745-1748, 2009.
[67] H. Kang, C. Lee, K. Nam, Y. S. Yoon, H. M. Chang, T. K. Ko, and B. Y. Seok, "Development of a 13.2kV/630A (8.3MVA) High Temperature Superconducting Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 18, pp. 628-631, 2008.
[68] B. W. Lee, K. B. Park, J. Sim, I. S. Oh, H. G. Lee, H. R. Kim, and O. B. Hyun, "Design and Experiments of Novel Hybrid Type Superconducting Fault Current Limiters," IEEE Transactions on Applied Superconductivity, vol. 18, pp. 624-627, 2008.
[69] G. H. Lee, K. B. Park, J. Sim, Y. G. Kim, I. S. Oh, O. B. Hyun, and B. W. Lee, "Hybrid Superconducting Fault Current Limiter of the First Half Cycle Non-limiting Type," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1888-1891, 2009.
[70] S. W. Yim, S. D. Yu, H. R. Kim, M. J. Kim, C. R. Park, S. E. Yang, W. S. Kim, O. B. Hyun, J. Sim, K. B. Park, and I. S. Oh, "Construction of Testing Facilities and Verifying Tests of a 22.9kV/630A Class Superconducting Fault Current Limiter," Physica C: Superconductivity, vol. 470, pp. 1611-1614, 2010.
[71] O. B. Hyun, S. W. Yim, S. D. Yu, S. E. Yang, W. S. Kim, H. R. Kim, G. H. Lee, J. Sim, and K. B. Park, "Long-term Operation and Fault Tests of a 22.9kV Hybrid SFCL in the KEPCO Test Grid," IEEE Transactions on Applied Superconductivity, vol. 21, pp. 2131-2134, 2011.
[72] W. Kim, O. Hyun, C. Park, S. Yim, S. Yoo, S. Yang, and H. Kim, "Dynamic Characteristics of a 22.9kV Hybrid SFCL for Short-circuit Test Considering a Simple Coordination of Protection System in Distribution Networks," IEEE Transactions on Applied Superconductivity, pp. 1-4, 2011.
[73] O. B. Hyun, K. B. Park, J. Sim, H. R. Kim, S. W. Yim, and I. S. Oh, "Introduction of a Hybrid SFCL in KEPCO Grid and Local Points at Issue," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1946-1949, 2009.
[74] H. Boenig and D. Paice, "Fault Current Limiter using a Superconducting Coil," IEEE Transactions on Magnetics, vol. 19, pp. 1051-1053, 1983.
References
221
[75] E. M. Leung, I. Rodriguez, G. W. Albert, B. Burley, M. Dew, P. Gurrola, D. Madura, G. Miyata, K. Muehleman, L. Nguyen, S. Pidcoe, S. Ahmed, G. Dishaw, C. Nieto, I. Kersenbaum, B. Gamble, C. Russo, H. Boenig, D. Peterson, L. Motowildo, and P. Haldar, "High Temperature Superconducting Fault Current Limiter Development," IEEE Transactions on Applied Superconductivity, vol. 7, pp. 985-988, 1997.
[76] E. Leung, B. Burley, N. Chitwood, H. Gurol, G. Miyata, D. Morris, L. Ngyuen, B. O'Hea, D. Paganini, S. Pidcoe, P. Haldar, M. Gardner, D. Peterson, H. Beonig, J. Cooley, Y. Coulter, W. Hults, C. Mielke, E. Roth, J. Smith, S. Ahmed, A. Rodriguez, A. Langhorn, M. Gruszczynski, and J. Hoehn, "Design and Development of a 15kV, 20kA HTS Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 10, pp. 832-835, 2000.
[77] Z. Wang, J. Zhang, D. Zhang, H. Li, Y. Guan, Q. Bao, L. Lin, and L. Xiao, "Design and Test of High-Tc Superconducting Coils for a Three-phase 10.5kV/1.5kA Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 16, pp. 658-661, 2006.
[78] D. Hui, Z. Wang, J. Zhang, D. Zhang, S. T. Dai, C. Zhao, Z. Zhu, H. D. Li, Z. Zhang, Y. Guan, L. Xiao, L. Lin, L. Li, L. Gong, X. Xu, J. Lu, Z. Fang, H. Zhang, J. Zeng, G. Li, and S. Zhou, "Development and Test of 10.5kV/1.5kA HTS Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 16, pp. 687-690, 2006.
[79] L. Li, L. Gong, X. Xu, J. Lu, Z. Fang, and H. Zhang, "Field Test and Demonstrated Operation of 10.5kV/1.5kA HTS Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 2055-2058, 2007.
[80] F. Moriconi, F. De La Rosa, A. Singh, B. Chen, M. Levitskaya, and A. Nelson, "An Innovative Compact Saturable-core HTS Fault Current Limiter - Development, Testing and Application to Transmission Class Networks," in IEEE Power and Energy Society General Meeting, 2010, pp. 1-8.
[81] Y. Xin, W. Gong, X. Niu, Z. Cao, J. Zhang, B. Tian, H. Xi, Y. Wang, H. Hong, Y. Zhang, B. Hou, and X. Yang, "Development of Saturated Iron Core HTS Fault Current Limiters," IEEE Transactions on Applied Superconductivity, vol. 17, pp. 1760-1763, 2007.
[82] X. Yin, W. Gong, X. Niu, Y. Gao, Q. Guo, L. Xiao, Z. Cao, H. Hong, A. Wu, Z. Li, X. Hu, B. Tian, J. Zhang, Y. He, Y. Wang, J. Cui, S. Ding, J. Wang, A. Ren, and F. Ye, "Manufacturing and Test of a 35kV/90MVA Saturated Iron-core Type Superconductive Fault Current Limiter for Live-grid Operation," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1934-1937, 2009.
[83] Y. Xin, J. Zhang, and W. Gong, "Voltage Surge Protection Circuit for Superconducting Bias Coil," IEEE Transactions on Applied Superconductivity, vol. 20, pp. 1118-1121, 2010.
[84] H. Hong, Z. Cao, J. Zhang, X. Hu, J. Wang, X. Niu, B. Tian, Y. Wang, W. Gong, and Y. Xin, "DC Magnetization System for a 35kV/90MVA Superconducting Saturated Iron-core Fault Current Limiter," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1851-1854, 2009.
References
222
[85] Y. Xin, H. Hui, W. Gong, F. Ye, J. Wang, B. Tian, A. Ren, and M. Zi, "Superconducting Cable and Superconducting Fault Current Limiter at Puji Substation," in International Conference on Applied Superconductivity and Electromagnetic Devices, 2009, pp. 392-397.
[86] Y. Sun, W. Gong, J. Wang, H. Hong, B. Tian, and Y. Xin, "DC Bias System of a 35kV/90MVA Saturated Iron Core SFCL," Cryogenics, vol. 51, pp. 257-260, 2011.
[87] Y. Xin, J. Wang, H. Hong, W. Gong, J. Zhang, X. Niu, A. Ren, D. Si, M. Zi, Z. Xiong, and F. Ye, "Field Tests on a 35kV/90MVA Superconducting Fault Current Limiter," in International Conference on Power System Technology, 2010, pp. 1-5.
[88] Y. Xin, H. Hong, J. Wang, W. Gong, J. Zhang, A. Ren, M. Zi, Z. Xiong, D. Si, and F. Ye, "Performance of the 35kV/90MVA SFCL in Live-grid Fault Current Limiting Tests," IEEE Transactions on Applied Superconductivity, vol. 21, pp. 1294-1297, 2011.
[89] T. Matsumura, H. Shimizu, and Y. Yokomizu, "Design Guideline of Flux-lock Type HTS Fault Current Limiter for Power System Application," IEEE Transactions on Applied Superconductivity, vol. 11, pp. 1956-1959, 2001.
[90] T. Matsumura, A. Kimura, H. Shimizu, Y. Yokomizu, and M. Goto, "Fundamental Performance of Flux-lock Type Fault Current Limiter with Two Air-core Coils," IEEE Transactions on Applied Superconductivity, vol. 13, pp. 2024-2027, 2003.
[91] S. H. Lim, "Analysis on Current Limiting Characteristics of a Flux-lock Type SFCL with Two Triggering Current Levels," Physica C: Superconductivity, vol. 471, pp. 1354-1357, 2011.
[92] W. Gong, J. Zhang, Z. Cao, H. Hong, B. Tian, Y. Wang, J. Wang, X. Niu, J. Qiu, S. Wang, and Y. Xin, "HTS DC Bias Coil for 35kV/90MVA Saturated Iron-core Fault Current Limiter," Physica C: Superconductivity, vol. 468, pp. 2050-2053, 2008.
[93] T. Yazawa, K. Koyanagi, M. Takahashi, M. Ono, M. Sakai, K. Toba, H. Takigami, M. Urata, Y. Iijima, T. Saitoh, N. Amemiya, and Y. Shiohara, "Design and Experimental Results of Three-phase Superconducting Fault Current Limiter using Highly-resistive YBCO Tapes," IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1956-1959, 2009.
[94] A. Greenwood, Vacuum Switchgear. Stevenage: Institution of Electrical Engineers, 1994.
[95] ABB, "VD4 - Vacuum Circuit-breaker," http://www05.abb.com/global/scot/scot235.nsf/veritydisplay/124a7a7dc93496cac1257720003df1de/$file/2457_VD4_6Seiter_GB.pdf, Accessed on: 30/09/2011.
[96] Schneider Electric, "Powersub™ Vacuum Substation Circuit Breaker, Type FVR," http://products.schneider-electric.us/linkservid/9D29C4F9-F271-4135-AA5660B3DECCB8AC/showMeta/0/, Accessed on: 30/09/2011.
References
223
[97] Y. Zhang, Z. Liu, Y. Geng, and J. Yao, "A New Axial Magnetic Field Contact for 126kV Single Break Vacuum Interrupters," in 23rd International Symposium on Discharges and Electrical Insulation in Vacuum, 2008, pp. 280-283.
[98] Y. Zhang, Y. Geng, L. Yu, J. Yan, Z. Liu, C. Hu, and J. Yao, "Axial Magnetic Field Strength Needed for a 126kV Single Break Vacuum Interrupter," in 24th International Symposium on Discharges and Electrical Insulation in Vacuum, 2010, pp. 320-323.
[99] J. Yan, Z. Li, Y. Geng, Z. Liu, C. Hu, and J. Yao, "Experimental Characterization of Current Chop in 126kV Vacuum Interrupters," in 24th International Symposium on Discharges and Electrical Insulation in Vacuum, 2010, pp. 237-240.
[100] M. P. Reece, "A Review of the Development of the Vacuum Interrupter," Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol. 275, pp. 121-129, 1973.
[101] P. Picot, "Vacuum Switching," Schneider Electric, 198, 2000.
[102] J. M. Lafferty, Vacuum Arcs Theory and Application. New York; Chichester; Brisbane; Toronto: John Wiley and Sons, 1980.
[103] E. Husain and N. RS, "Analysis of Paschen Curves for air, N2 and SF6 Using the Townsend Breakdown Equation " IEEE Transactions on Electrical Insulation, vol. 17, pp. 350-353, 1982.
[104] S. Theoleyre, "MV Breaking Techniques," Schneider Electric, 193, 1999.
[105] R. P. P. Smeets, V. Kertesz, D. Dufournet, D. Penache, and M. Schlaug, "Interaction of a Vacuum Arc with an SF6 Arc in a Hybrid Circuit Breaker During High-current Interruption," IEEE Transactions on Plasma Science, vol. 35, pp. 933-938, 2007.
[106] M. Liao, X. Cheng, X. Duan, and J. Zou, "Study on Dynamic Arc Model for High Voltage Hybrid Circuit Breaker using Vacuum Interrupter and SF6 Interrupter in Series," in 24th International Symposium on Discharges and Electrical Insulation in Vacuum, 2010, pp. 174-178.
[107] E. Dullni, "A Vacuum Circuit-breaker with Permanent Magnetic Actuator for Frequent Operations," in 18th International Symposium on Discharges and Electrical Insulation in Vacuum, 1998, pp. 688-691.
[108] B. A. R. McKean, "Magnets and Vacuum - the Perfect Match [MV Distribution Switchgear]," in 5th International Conference on Trends in Distribution Switchgear: 400V-145kV for Utilities and Private Networks, 1998, pp. 73-79.
[109] B. Lequesne, "Dynamic Model of Solenoids under Impact Excitation, Including Motion and Body Currents II," in IEEE Industry Applications Society Annual Meeting, 1988, pp. 149-157.
[110] IPEN ENG, "Background - PM Actuators in MV Switchgear," http://www.ipeceng.com/products/magact/background.html, Accessed on: 2/10/2011.
[111] X. Lin, H. Gao, and Z. Cai, "Magnetic Field Calculation and Dynamic Behavior Analyses of the Permanent Magnetic Actuator," in 19th International
References
224
Symposium on Discharges and Electrical Insulation in Vacuum, 2000, pp. 532-535.
[112] H. Jiang, R. Shuttleworth, B. A. T. Al Zahawi, X. Tian, and A. Power, "Fast Response GTO Assisted Novel Tap Changer," IEEE Transactions on Power Delivery, vol. 16, pp. 111-115, 2001.
[113] R. Shuttleworth, "Electrical Changeover Switching," United States Patent: 238210, pp. 1-12, 1997.
[114] A. J. W. Lammers, P. P. Leeufkens, and G. C. Schoonenberg, "MV Vacuum Switchgear based on Magnetic Actuators," in 5th International Conference on Trends in Distribution Switchgear: 400V-145kV for Utilities and Private Networks, 1998, pp. 86-90.
[115] J. H. Kang, C. Y. Bae, and H. K. Jung, "Dynamic Behavior Analysis of Permanent Magnetic Actuator in Vacuum Circuit Breaker," in 16th International Conference on Electrical Machines and Systems, 2003, pp. 100-103.
[116] E. Dullni, H. Fink, and C. Reuber, "A Vacuum Circuit-breaker with Permanent Magnetic Actuator and Electronic Control," in Conference and Exhibition on Electricity Distribution, Nice, France, 1999.
[117] E. S. Hamdi, Design of Small Electrical Machines. Chichester; New York; Brisbane; Toronto; Singapore: John Wiley and Sons, 1994.
[118] E. P. Furlani, Permanent Magnet and Electromechanical Devices: Materials, Analysis, and Applications. San Diego; London: Academic, 2001.
[119] J. Borwick, Loudspeaker and Headphone Handbook. London: Butterworths, 1988.
[120] M. Colloms, High Performance Loudspeakers, 5th Ed. New York: John Wiley and Sons, 1997.
[121] B. P. Szeifert and M. P. Dunk, "Magnetic Latch for a Voice Coil Actuator," pp. 1-18, 2008.
[122] Dynamic Ceramic, "Dynallox Alumina Ceramics Data Sheet," http://www.dynacer.com/PDF/dynalloxproperties.pdf, Accessed on: 01/11/2008.
[123] Accuratus, "Aluminum Oxide, Al2O3," http://accuratus.com/alumox.html, Accessed on: 01/11/2008.
[124] A. P. Rijpma and H. J. M. t. Brake, "Cryogenic Thermometry with a Common Diode: Type BAS16," Cryogenics, vol. 46, pp. 68-69, 2006.
[125] C. J. Yeager and S. S. Courts, "A Review of Cryogenic Thermometry and Common Temperature Sensors," IEEE Sensors Journal, vol. 1, pp. 352-360, 2001.
[126] J. Veprek and P. Strnad, "Stability of Silicon Diodes as Temperature Sensors in the Range 4.2 - 273 K," Cryogenics, vol. 24, pp. 245-248, 1984.
[127] S. Harrison, "SuFCLEMP Demonstrator Cryostat," Scientific Magnetics, Abingdon, 2006.
[128] M. G. Say, Alternating Current Machines. Bath: Pitman Press, 1976.
References
225
[129] K. Shimohata, S. Yokoyama, T. Inaguchi, S. Nakamura, and Y. Ozawa, "Current Distribution Measurement in YBCO Thin Film for a Superconducting Fault Current Limiter," Cryogenics, vol. 43, pp. 111-116, 2003.
[130] O. B. Hyun, H. R. Kim, J. Sim, Y. H. Jung, K. B. Park, J. S. Kang, B. W. Lee, and I. S. Oh, "6.6kV Resistive Superconducting Fault Current Limiter based on YBCO Films," IEEE Transactions on Applied Superconductivity, vol. 15, pp. 2027-2030, 2005.
[131] Cedrat Group, http://www.cedrat.com/en/software-solutions/flux.html, Accessed on: 07/06/2009.
[132] M. N. Ozisik, Heat Transfer: a Basic Approach. New York: McGraw-Hill, 1985.
[133] Y. S. Touloukian, Thermophysical Properties of Matter the TPRC Data Series, vol. 1-5. New York: IFI/Plenum, 1970.
[134] D. E. Gray and B. H. Billings, American Institute of Physics Handbook, 3rd Ed. New York: McGraw-Hill, 1972.
[135] R. C. Weast, CRC Handbook of Chemistry and Physics: a Ready-reference Book of Chemical and Physical Data, 55 Ed. Cleveland, 1974.
[136] G. W. C. Kaye and T. H. Laby, Tables of Physical & Chemical Constants, 16th Ed. Harlow: Longman, 1995.
[137] L. Ye and K. P. Juengst, "Modeling and Simulation of High Temperature Resistive Superconducting Fault Current Limiters," IEEE Transactions on Applied Superconductivity, vol. 14, pp. 839-842, 2004.
[138] L. Ye and A. M. Campbell, "Case Study of HTS Resistive Superconducting Fault Current Limiter in Electrical Distribution Systems," Electric Power Systems Research, vol. 77, pp. 534-539, 2007.
[139] H. van Weeren, N. C. van den Eijnden, W. A. J. Wessel, P. Lezza, S. I. Schlachter, W. Goldacker, M. Dhalle, A. den Ouden, B. Haken, and H. H. J. Kate, "Adiabatic Normal Zone Development in MgB2 Superconductors," IEEE Transactions on Applied Superconductivity, vol. 15, pp. 1667-1670, 2005.
[140] "FLUX Version 10.3 User's Guide," Cedrat Group, 2009.
[141] "How to Define Properties With Tables in Files?," Cedrat Group, 2009.
[142] Cobham, "Opera-3d - Electromagnetic Design in Three Dimensions," http://www.cobham.com/about-cobham/aerospace-and-security/about-us/antenna-systems/kidlington/products/opera-3d.aspx, Accessed on: 16/02/2009.
[143] Arnold Magnetic Technology Corporation, "Sintered Neodymium-iron-boron Magnets N48," http://www.arnoldmagnetics.com/Neodymium_Literature.aspx, Accessed on: 01/10/2009.
[144] D. Hadfield, Permanent Magnets and Magnetism: Theory, Materials, Design, Manufacture and Applications. London: John Wiley and Son, 1962.
[145] H. C. Roters, Electromagnetic Devices, 1st Ed. New York: John Wiley and Sons, 1941.
References
226
[146] R. L. Ferrari, An Introduction to Electromagnetic Fields. New York; London: Van Nostrand Reinhold, 1975.
[147] J. Woolman and R. A. Mottram, The Mechanical and Physical Properties of British Standard Steels (B.S.970-1955), vol. 1. Oxford: Pergamon, 1964.
[148] M. McCaig and A. Clegg, Permanent Magnets in Theory and Practice. London: Pentech Press, 1977.
[149] P. Slade, "The Vacuum Interrupter Contact," IEEE Transactions on Components, Hybrids, and Manufacturing Technology, vol. 7, pp. 25-32, 1984.
[150] P. Barkan, "A Study of the Contact Bounce Phenomenon," IEEE Transactions on Power Apparatus and Systems, vol. PAS-86, pp. 231-240, 1967.
[151] N. Mohan, T. M. Undeland, and W. P. Robbins, Power Electronics: Converters, Applications and Design. Hoboken, N.J.: John Wiley and Sons, 2003.
[152] G. Ortiz, D. Bortis, J. Biela, and J. W. Kolar, "Optimal Design of a 3.5kV/11kW DC-DC Converter for Charging Capacitor Banks of Power Modulators," in IEEE Pulsed Power Conference, 2009, pp. 1406-1411.
[153] IXYS Corporation, "IXFH120N20P," http://ixapps.ixys.com/DataSheet/DS99223(IXFH-FK120N20P).pdf, Accessed on: 08/06/2010.
[154] A. S. Sedra and K. C. Smith, Microelectronic Circuits, 5th Ed. Oxford University Press, 2004.
[155] Philips Semiconductors, "Application Note AN-170: NE555 and NE556 Applications," http://www.doctronics.co.uk/pdf_files/555an.pdf, Accessed on: 10/01/2012.
[156] M. H. Rashid, Power Electronics Handbook: Devices, Circuits, and Applications. Burlington, Massachusetts: Academic Press, 2006.
[157] IXYS Corporation, "Application Note AN-0002: MOSFET/IGBT Drivers: Theory and Application," http://www.ixysrf.com/pdf/switch_mode/appnotes/5mosfet_driver_theory_and_applications.pdf, Accessed on: 10/01/2012.
[158] International Rectifier, "Application Note AN-978: HV Floating MOS-gate Driver ICs," http://www.irf.com/technical-info/appnotes/an-978.pdf, Accessed on: 10/09/2010.
[159] "Application Note AN-978: HV Floating MOS-gate Driver ICs," International Rectifier.
[160] International Rectifier, "IR2113 High and Low Side Driver," http://www.irf.com/product-info/datasheets/data/ir2110.pdf, Accessed on: 15/09/2010.
[161] Traco Power, "DC/DC Converters TMA Series, 1Watt," http://www.tracopower.com/products/tma.pdf, Accessed on: 20/01/2012.
[162] Tedea-Huntleigh, "S - Type Load Cell Model 615," http://docs-europe.electrocomponents.com/webdocs/028f/0900766b8028f75c.pdf, Accessed on: 09/06/2010.
References
227
[163] Solartron Metrology, "Inductive Technology Principle of Operation," http://docs-europe.electrocomponents.com/webdocs/0ca9/0900766b80ca9df9.pdf, Accessed on: 08/06/2010.
[164] Solartron Metrology, "DC Miniature Displacement Transducer," http://www.solartronmetrology.com/products/LVDT-displacement-transducers.cfm, Accessed on: 08/06/2010.
[165] R. A. Matula, "Electrical Resistivity of Copper, Gold, Palladium, and Silver," Journal of Physical and Chemical Reference Data, vol. 8, pp. 1147-1298, 1979.
[166] E. Bauer, C. Paul, S. Berger, S. Majumdar, H. Michor, M. Giovannini, A. Saccone, and A. Bianconi, "Thermal Conductivity of Superconducting MgB2," Journal of Physics: Condensed Matter, vol. 13, pp. L487-L493, 2001.
[167] T. Muranaka, J. Akimitsu, and M. Sera, "Thermal Transport Properties of MgB2," Physical Review B, vol. 64, pp. 1-3, 2001.
[168] C. Liu, G. Yan, S. Du, W. Xi, Y. Feng, P. Zhang, X. Wu, and L. Zhou, "Effect of Heat-treatment Temperatures on Density and Porosity in MgB2 Superconductor," Physica C: Superconductivity, vol. 386, pp. 603-606, 2003.
[169] "Niobium Technical Data Sheet," http://www.wahchang.com/pages/products/data/pdf/Niobium.pdf, Accessed on: 15/01/2009.
[170] Praxair Technology, "Praxair Material Safety Data Sheet: Nitrogen, Refrigerated Liquid," http://www.praxair.com/praxair.nsf/0/1a2e41d3c1fb1c238525658a00775df8/$FILE/p4630j.pdf, Accessed on: 16/01/2009.
[171] Food and Agriculture Organization of the United Nations, "The Use of Ice on Small Fishing Vessels," ftp://ftp.fao.org/docrep/fao/006/y5013e/Y5013E03.pdf, Accessed on: 15/12/2009.
Appendix A – Materials Data
228
Appendix A – Materials Data
A1 Alumina
Table A.1 Thermal conductivity of alumina
Temperature (K) Thermal conductivity (W/m ·K) Reference 20 23 77 150 194 48 273 35 373 26
[134]
20 24 30 47 35 63 60 150 85 160 95 140 200 50
[133]
84 125.9 126 48.1 249 35.1 100 133 150 77 200 55 250 43.4 273.2 39.7 300 36 350 30.7 400 26.4 500 20.2
[133]
Appendix A – Materials Data
229
Table A.2 Specific heat capacity of alumina
Temperature (K) Specific heat capacity (J/kg·K) Reference 30 2.58 35 4.31 40 6.78 45 10.20 50 14.63 55 20.30 60 27.25 65 35.51 70 44.94 75 55.56 80 67.61 85 80.88 90 95.07 95 110.04 100 125.94 105 142.59 110 160.08 115 178.11 120 196.73 125 215.69 130 235.02 135 254.51 140 274.22 145 293.93 150 313.63 155 333.38 160 353.05 165 372.41 170 391.70 175 410.74 180 429.28 185 448.10 190 466.09 195 483.67 200 501.66 205 518.82 210 535.55 215 551.87 220 568.19 225 583.67
[133]
Appendix A – Materials Data
230
Temperature (K) Specific heat capacity (J/kg·K) Reference 230 599.15 235 614.21 240 628.86 245 643.49 250 657.31 255 670.69 260 684.08 265 697.05 270 709.60 275 722.16 280 733.87 285 745.59 290 756.89 295 768.18 298.16 774.88 300 778.64 305 789.10 310 799.56 315 809.19 320 818.81 325 828.43 330 837.64 335 846.00 340 855.21 345 863.58 350 871.53 360 887.84 370 902.90 380 916.71 390 930.52 400 943.07 410 955.20 420 966.50 430 977.38 440 987.42 450 997.47 460 1006.67 470 1015.46 480 1023.82 490 1032.19 500 1040.14
[133]
Appendix A – Materials Data
231
Table A.3 Density of alumina
Density (kg/m3) Reference 3670 [122]
A2 Copper
Table A.4 Thermal conductivity of copper
Temperature (K) Thermal conductivity (W/m ·K) Reference 30 4300 35 2900 40 2050 45 1530 50 1220 60 850 70 670 80 570 90 514 100 483 150 428 200 413 250 404 273.2 401 300 398 350 394 400 392 500 388
[133]
Appendix A – Materials Data
232
Table A.5 Specific heat capacity of copper
Temperature (K) Specific heat capacity (J/kg·K) Reference 30 26.65 35 41.84 30 26.65 35 41.84 40 58.99 45 77.69 50 96.94 55 116.57 60 135.73 65 153.80 70 171.00 75 187.61 80 202.88 85 216.69 90 230.33 95 241.54 100 251.79 110 270.58 120 286.98 130 301.04 140 312.38 150 322.38 160 331.29 170 338.78 180 345.18 190 351.04 200 356.18 210 360.54 220 364.38 230 367.89 240 371.12 250 374.34 260 377.36 270 380.16 273.15 380.87 280 382.38 300 386.06 373 393.29 473 414.22 573 422.58
[133]
Appendix A – Materials Data
233
Table A.6 Density of copper
Density (kg/m3) Reference 8933 8950
[134]
Table A.7 Resistivity of copper
Temperature (K) Resistivity (Ω·m×10-8) Reference 30 0.00828 35 0.0147 40 0.0239 45 0.0358 50 0.0518 55 0.0727 60 0.0971 70 0.154 80 0.215 90 0.281 100 0.348 125 0.522 150 0.699 175 0.874 200 1.046 225 1.217 250 1.387 273.15 1.543 293 1.678 300 1.725 350 20.63 400 2.402 500 3.09
600 3.79
[135, 165]
Appendix A – Materials Data
234
A3 Magnesium diboride
Table A.8 Thermal conductivity of magnesium diboride
Temperature (K) Thermal conductivity (W/m ·K) Reference 25 9.04 39.8 15 50 17.3 65.7 20 100 22.5 150 22.88 200 23.27 250 24.33 300 26.06
[166, 167]
Table A.9 Specific heat capacity of magnesium diboride
Temperature (K) Specific heat capacity (J/kg·K) Reference 21.12 0.911
23.05 2.00
25.06 3.097
27.19 4.19
29.66 4.83
32.41 6.74
34.92 9.38
37.75 12.20
41.37 14.85
45.43 22.59
54.12 40.99
57.42 50.46
60.99 60.21
65.42 77.32
70.26 100.92
75.6 123.43
81.29 153.39
85.52 170.41
91.29 199.28
97.04 226.35
102.83 253.76
108.58 287.39
114.13 311.64
119.2 343.13
[133]
Appendix A – Materials Data
235
Temperature (K) Specific heat capacity (J/kg·K) Reference 124.4 368.28
129.79 393.42
136.05 439.32
140.35 461.91
154.92 538.48
173.53 625.09
183.37 675.29
194.73 726.34
208.21 782.83
219.27 828.43
236.77 897.89
238.69 899.98
246.14 920.89
248.39 928.85
254.87 948.09
259.61 974.04
266.88 990.35
274.04 1014.20
279.19 1024.66
286.69 1039.72
287.69 1035.54
298.81 1046.84
[133]
26.44 5.28
128.41 400.62
145.03 488.6
160.29 570.69
183.37 675.29
208.21 782.83
248.39 928.85
279.19 1024.66
300.14 1039.31
[133]
273.71 913.79 [133]
238.69 899.98
254.87 948.09
287.69 1035.54
304.22 1041.39
[133]
Table A.10 Density of magnesium diboride
Density (kg/m3) Reference
2600-2630 [168]
Appendix A – Materials Data
236
A4 Monel (Nickel alloy)
Table A.11 Thermal conductivity of monel
Temperature (K) Thermal conductivity (W/m ·K) Reference 25.43 9.08 40.44 12.4 55.54 14.3 70.52 15.2 84.36 16 98.90 16.5 113.26 17.1 127.12 17.5 142.56 18.1 156.84 18.6 171.86 19 198.90 19.8 213.90 20.2 229.03 20.6 245.02 21.1 263.14 21.7 282.81 22.3 295.2 22.6
[133]
25.98 5.65 40.92 9.5 55.32 11.8 70.02 13.3 86.51 14.4 101.15 15.4 115.14 16.1 130.22 17 143.22 17.7 160.39 18.3 175.46 18.8 204.02 19.7 220.29 20.1 239.98 20.8 258.13 21.3 288.14 21.8 361.2 28 392.2 30.5 416.2 30.1 432.7 31.4 465.2 32.2 467.7 31 484.7 33.1 500.2 33.5
[133]
Appendix A – Materials Data
237
Table A.12 Specific heat capacity of monel
Temperature (K) Specific heat capacity (J/kg·K) Reference 73 259.41
123 305.43
173 347.27
223 380.74
273 405.85
373 447.69
473 259.41
[133]
116 297.06 144 322.17 200 364.01 293 418.4 366 447.69 478 476.98 589 489.53
[133]
Table A.13 Density of monel
Density (kg/m3) Reference
8840 [134]
A5 Niobium
Table A.14 Thermal conductivity of niobium
Temperature (K) Thermal conductivity (W/m ·K) Reference 20 229
25 187
30 145
35 116
40 97
45 84
50 76
60 66
70 61 80 58
90 56.3
100 55.2
150 53
200 52.6
[133]
Appendix A – Materials Data
238
Temperature (K) Thermal conductivity (W/m ·K) Reference 250 53
273.2 53.3
300 53.7
350 54.4
400 55.2
500 56.7
[133]
Table A.15 Specific heat capacity of niobium
Temperature (K) Specific heat capacity (J/kg·K) Reference 21.6 12.59 24.78 18.07 28.89 26.57 33.44 37.40 38.21 50.46 42.98 66.65 47.77 78.49 52.52 92.13 57.46 106.39 59.01 110.08 62.38 118.53 65.37 125.85 67.55 130.67 69.73 135.65 72.6 140.46 74.2 145.31 79.23 155.23 84.62 164.68 89.99 173.64 95.58 181.84 101.43 189.66 107.34 196.23 112.93 202.67 118 206.89 123.15 210.92 128.57 215.09 133.91 220.08 139.54 223.47 145.2 227.19 150.61 230.12 156.29 232.59 161.87 235.86 167.6 238.28 173.2 239.79 178.87 241.29
[133]
Appendix A – Materials Data
239
Temperature (K) Specific heat capacity (J/kg·K) Reference 184.5 243.43 189.92 246.06 195.2 247.11 198.91 247.73 200.81 249.03 201.16 248.74 204.13 249.74 206.51 250.04 209.84 250.87 211.68 251.83 212.25 251.58 215.09 252.76 216.5 253.09 217.49 253.47 220.8 254.81 221.21 254.14 222.91 254.76 225.98 255.64 228.03 255.98 231.91 256.69 233.76 257.48 237.37 257.90 238.98 258.15 243.73 258.32 244.35 259.32 248.53 259.74 249.2 259.99 253.6 260.87 254.94 260.79 259.03 261.22 260.72 261.21 265.3 262.09 265.8 262.21 270.51 261.5 270.64 263.01 273.15 269.03 373 272.38 473 275.89 573 279.62
[133]
Table A.16 Density of niobium
Density (kg/m3) Reference 8570 [169]
Appendix A – Materials Data
240
A6 Nitrogen
Table A.17 Thermal conductivity of nitrogen
Temperature (K) Thermal conductivity (W/m ·K) Reference 20 0.4 25 0.32 30 0.27 35 0.23 40 0.2 45 0.18 50 0.16 60 0.169 65 0.16 70 0.151 75 0.1413 80 0.1322 85 0.1231 90 0.1142 95 0.1053 100 0.0966 125 0.052 150 0.01385 160 0.01474 170 0.01562 180 0.01651 190 0.01739 200 0.01826 220 0.01989 240 0.02145 260 0.02298 280 0.02449 300 0.02598 400 0.03252 500 0.03864
[133]
Table A.18 Specific heat capacity of nitrogen
Temperature (K) Specific heat capacity (J/kg·K) Reference 25 970.73 30 1233.57 50 1481.48
[134]
Appendix A – Materials Data
241
Temperature (K) Specific heat capacity (J/kg·K) Reference 70 2001.20 100 14874.58 150 1039.42 200 1039.42 250 1039.42 298.15 1039.42 400 1043.91 500 1055.85
[134]
Table A.19 Density of nitrogen
Temperature (K) Density (kg/m3) Reference
0.15 1140 [134]
77.4 808.5 294.23 1.16
[170]
A7 Polystyrene
Table A.20 Thermal conductivity of polystyrene
Thermal conductivity (W/m ·K) Reference 0.057 0.045 0.037 0.034 0.033
[171]
Table A.21 Specific heat capacity of polystyrene
Temperature (K) Specific heat capacity (J/kg·K) Reference 77 380 173 710 273 1110 473 2150
[136]
Table A.22 Density of polystyrene
Density (kg/m3) Reference
10-33 [171]
Appendix B – Components for Operating Actuator
242
Appendix B – Components for Operating Actuator
Table B.1 Summary of components for prototype operating actuator
Item Material Number Size (mm)
Length 198
Width 30 Steel block for wall EN1A steel 4
Height 120
Length 228
Width 228 Steel block for bottom EN1A steel 1
Height 30
Radius 63 Steel cylinder EN1A steel 1
Height 88
Length 50
Width 14 Steel block for latch EN1A steel 4
Height 14
Length 50
Width 20 Steel block for latch EN1A steel 2
Height 6
Length 80
Width 80 Steel block for stopper EN1A steel 4
Height 6
Length 20
Width 20 Steel block for stopper EN1A steel 4
Height 6
Length 50
Thickness 15.3-20 Permanent magnet N48 Nd-Fe-B
magnet 4
Height 80
Length 50 Permanent magnet for
latch
N48 Nd-Fe-B
magnet
2
Width 10
Appendix B – Components for Operating Actuator
243
Height 6
Length 200
Width 200 Moving plate Carbon fibre 1
Height 6
Inner diameter 63.5
Outer diameter 65.5 Thin tube for actuator
coil Fibreglass 1
Height 98
Turn 79
Radius 66 Actuator coil AWG19 copper
wire 1
Height 75
Natural length 35 Wipe spring Stainless steel 1
Diameter 12
Natural length 30 Supporting spring Stainless steel 1
Diameter 12
Radius 63 Non-magnetic cylinder Aluminium 1
Height 11
Inner diameter 18
Outer diameter 25 Snatch bracket Aluminium 1
Height 30
Diameter 8 Fixing pole Aluminium 4
Length 110
Diameter 8 Shaft Brass 2
Length 30
Length 180
Width 180 Supporting plate Paxolin 1
Height 10
Length 20
Width 20 Stopper damper Rubber 4
Height 2
Diameter 63 Stopper damper Rubber 1
Height 2