11th EPRI Superconductivity Conference
Development and Optimization of Large
Direct Drive Superconducting Generators
for Off-Shore Wind Farms
Philippe J. Masson, Ph.D
Assistant Professor
University of Houston
Department of Mechanical Engineering
Texas Center for Superconductivity
pjmasson@uh.edu
Oct. 29th, 2013
Houston, TX
http://eetweb.com/wind/wind-turbines-go-supersized-20091001/
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
1
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
2
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
3
Global Energy – A “Hungry” Market
•
Existing and expanding global economies have a large appetite for Energy…
…with no signs of letting up!
250
Projections
History
(1012 KWh)
200
+45%
World Primary Energy Consumption
 USA-2007
150
100
50
0
1980
1985
1990
1995
2000
2005 2010
2015
2020
2025
2030
Sources: History: EIA, International Energy Annual 2005 (June-October 2007). Projections: International Energy Agency World Energy Projections Plus (2008)
“In order to meet the 45% increase in projected demand, an investment of over $26 trillion will be required …”
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
4
Global Offshore Wind Market
> 80 GWatts of new installations (>8,000 ea 10 MWatt Turbines)
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
5
Price Range of Renewable Electricity (2008)
Solar
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
6
Attaining Cost Of Energy Goals for Wind
•
10 GW @ 10c/kWh by 2020, 54 GW @ 7c/kWh by 2030…today at 27c/kWh
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
7
Wind Generators – Requirements
• Large power output
(less generators)
• High reliability; high
availability
• Low capital cost
(CAPEX)
• Ease of manufacture
and assembly
• Ease of maintenance
• High efficiency
http://www.windpowerengineering.com/design/mechanical/understandi
ng-costs-for-large-wind-turbine-drivetrains/
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
8
HTS Generators – Value Proposition
• Generator mass reduction
– Some impact on LCOE
• High reliability; high availability
– No gear box
– No thermal cycling
– Sealed system
• Long maintenance time if problem in cryostat
• No failure allowed at cryogenic temperature
• Low capital cost (CAPEX)
– Driven by cost of conductor and cooling system
– Expected to be competitive for large systems
• Ease of manufacture and assembly
• Ease of maintenance
– Less reliable components can be located outside of the cryostat
– Modular approach can be used (individual cryostats…)
http://www.ngpowereu.com/news/
• High efficiency
– Low losses, high efficiency at fractional power output
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
europes-push-on-offshorerenewables/
9
HTS Machine Possible Configurations
Back Iron
•
•
•
•
•
•
•
Salient Iron
Ironless All-Cryo
*
*
*
*
*
*
*
*
*
*
*
*
Low mass
Large number of poles
Low xd
High short circuit
current and torque
High peak field
Large quantity of HTS
conductor
Low cooling
requirements
Courtesy of C. Oberly, AFRL
*
•
•
•
•
•
•
•
•
Low mass
Low number of poles
Low xd
Fault current
limitation
High peak field
Very large quantity of
HTS conductor
AC losses
High cooling
requirements
Field Windings
* Armature Windings
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
•
•
•
•
•
•
Iron
Eddy Current Shield
Heavy
Reg. xd
Low peak field (if not
saturated)
Low quantity of HTS
conductor
Low cooling
requirements
Modular cryostat
possible
Cyrogenic region
Output Terminals
10
Design Requirements
• Economic
–
–
–
–
• Thermal
Low cost conductors
Low cost cryocoolers
Superconductor availability
Cost effective manufacturing
• Mechanical
– Torque transmission/torque tube
• Composite - steel
– Large Lorentz forces (peak field >4 T)
– Torque and forces applied on
conductors
– Heat leaks need to be minimized
• Conduction through shaft
• Current leads
• Splices
– AC losses
•
Multifilament conductors
• Stability
– Quench detection/protection
– Fault current/torque
MgB2 conductor
2G conductor
Carbon fiber composite thermal conductivity
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
11
Partially and Fully Superconducting Machines
Thermal Shield
Armature Wdg.
Rotor
Stator
Fully Superconducting (FSG)
Field Wdg.
Shaft
Apparent Power output of an electrical generator:
Backiron
S  B K s  r La
Rotor
Shaft
0
r
2
0

p
Field Wdg.
Thermal Shield
EM Shield
Stator
Rotor contribution
Active volume
Limited by conductor
Larger radius
Backiron
performance.
needed for PSG
Partially Superconducting (PSG)
More conductor
because of the
needed in PSG
limited electrical
because of the large
loading
air gap
S = apparent power (VA)
Stator contribution
Rotation speed
Br0 = no-load excitation field (T)
Much higher values
Frequency needs to
Ks = electrical loading (A/m)
obtained in FSG
be kept low in FSG
R0 = average radius of armature winding (m)
La = active length (m)
because of high current
to limit AC losses
 = angular frequency (rd/s)
density in
p = number of pole pairs
superconductor
Armature Wdg.
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
12
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
13
Choice of Conductor
•
The conductor defines the operating temperature of the system
•
Key conductor parameters :
– Engineering critical current density @ operating field
– Filament size (AC applications)
– Ratio superconductor/ non superconductor
– Minimum quench energy
– Normal zone propagation velocity
– Minimum bending radius
– Cost
MgB2 conductors
BiSrCaCuO conductors
YBCO conductors
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
14
High Level Conductors Comparison
1G (BSCCO)

2G (YBCO)

MgB2

 
 




• The choice of conductor is done at the system level considering the total cost of system
conductor-cooling system
• MgB2 is very promising:
• Price point of MgB2 moving towards $10/kAm @ 2T, 20 K
• Development of high filament count conductors (~10 mm)
• 2G (YBCO) is improving fast:
• Current price point of YBCO at $400/kAm @ 2T, 60 K
• Active development towards cost reduction and multi-filaments
• 4X performance increase
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
15
Current Cost of Conductors at 2 T (2012)
Cost of conductor
1000
100
$/kAm
4x improvement expected (UH – ARPA-e)
YBCO-$/kAm
10
MgB2-$/kAm
2x improvement expected (mass
production, higher Jc) (DOE/NASA)
1
0
20
40
60
80
T (K)
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
16
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
17
Large Wind Turbine Generators
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
18
Fully Superconducting Generator (MgB2)
Turbine hub support
and bearings
Fully Superconducting
Generator
Approximate Specifications:
Diameter:
3.5 m [140 in]
Length:
7.6 m [300 in]
Mass:
140 t [319,000 lbs]
Power Converters
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
19
Fully Superconducting Rotating Machine
•
10 MW, Direct Drive, Fully Superconducting Generator
– Detailed Conceptual Design ✓
– Cost of Energy Analysis ✓
– De-risking Program Development: 2012-2014
•
•
•
•
•
•
Mechanical components
Manufacturing
Detailed LCOE
MgB2 Conductor development
Fault conditions
AC losses (modeling/experimental)
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
20
Fully Superconducting Generator (FSG) Block Diagram
System
Controls
Back Iron
Stator Cryostat (@ 15-20K)
Stator (Superconducting)
Power In 
Main Shaft
(10 RPM)
Rotor (Superconducting)
Heat
Exchange
Rotor
Excitation
(DC Power)
Heat
Exchange
Cryocooling
System
(gas Helium)
Rotor
Cryostat
(@15-20K)
Current
Leads
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
Power
Conversion
(AC-DC-AC)
Power Out 
21
FSG Design - High Fidelity Sizing Models
Philippe Masson, AML
Proprietary Sizing Model is applied to obtain Superconducting Generator Design
1. Primary inputs  Power requirement and RPM
1. High fidelity analytical sizing is performed from first principles based on a simplified geometry
1. Primary outputs  Generator parameters
• Electromagnetic
• Thermal
• Structural
Wdg.
Parameters
Avg. ro
2. Optimization is performed for minimum
mass with constraints on the AC losses
Number of
Layers
3. Final Outputs:
•
Active length of coils
•
Overall principal dimensions
•
Rotor and stator field requirements
•
Key machine parameters
Air gap
Length: d
Wdg.
Parameters
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
22
Philippe Masson, AML
AC Losses and Machine Mass Trade-Off
• AC losses can be reduced at the expense of additional weight
• Cryocooler represents a small fraction of the total weight
350
3.50%
300
3.00%
cryocooler
weight
generator
weight
250
2.50%
200
2.00%
150
1.50%
100
1.00%
50
0.50%
0
0.00%
0
100
200
300
400
500
600
cryocooler weight (% of total mass)
Generator Weight (metric tons)
Machine weight vs. AC losses in stator
700
AC losses (W @ 20 K)
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
23
3D FEA Models
Structural
Electromagnetic
•
FEA models for generator design and
optimization
Electro-thermal
Thermal
heat pulse
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
24
Philippe Masson, AML
Stator Current Limitation
•
The superconducting stator acts as fault current limiter
superconductor
resistive matrix
 Provides lower short circuit torque
 Provides lower short circuit power without impacting dynamics
• Fault Sequence:
1. Short circuit occurs and phase current rises
2. Current level exceeds Ic (MgB2 critical current)
3. Stator current diffuses from the superconductor to the matrix
material throughout entire coil
4. Coil impedance changes from inductive to resistive
5. Single phase impedance rises from ~ 4 to >600 
6. Fault current drops from >1,000 A to ~ 6 A (dissipating ~22 kW)
7. Phase imbalance is detected
8. Temperature rise in windings has to be limited to less then 60K
Multi-filament MgB2 conductor
short circuit
Phase current vs. time
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
Temperature distribution after
1 second of short circuit
25
FSG Summary
• Accomplishments
– Design and detailed CAD drawings
– Manufacturing plan
– Detailed LCOE model
GHe Transfer
Line
• Ongoing work
Cooling
System
– De-risking tasks (structural material,
multi-filamentary conductors, cooling, quench…)
– AC losses in superconducting stator
• Modeling
• Experimental validation
– Higher fidelity LCOE model considering dynamic
analysis
Sample
Cryostat
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
26
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
27
ARPA-E funded program for High Performance
Superconducting Wires and Coils
• ARPA-E funded program
• UH-led program with SuperPower, TECOWestinghouse, Tai-Yang Research and NREL.
Technology Impact
Present-day superconducting wire constitutes more than
60% of the cost of a 10 MW superconducting wind
generator. By quadrupling the superconducting wire
performance at the generator operation temperature, the
amount of wire needed would be reduced by four which will
greatly enhance commercial viability and spur a
tremendous growth in wind energy production in the U.S.
High-power, Efficient
Wind Turbines
Engineered nanoscale defects
4x improved
wire
manufacturing
Lower cost
Generator coils
Quadrupling Superconductor Wire Performance for Commercialization of 10 MW
Wind Generators
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
28
2 MW HTS Generator Conceptual Design
• Warm iron rotor core
• Individual cryostats for
HTS coils
• Copper air-gap stator
• GHe cooling system
• Design of 2MW generator
used as starting point for
the 10 MW generator
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
29
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
30
Machine Configurations
The rotor magnetic configuration is controlled using binary parameters.
The gray parts are magnetic
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
31
Geometry Parameterization
• Geometry is
entirely
parameterized
Stat_slot_hei
ght
Stat_winding_wi
dth
Stat_core_hei
ght
• Poles are
represented by
nondimensional
parameters
airgap
Maxwell Tensor
Surface
Rot_crown_
width
Rot_winding_w
idth
Rotor_radius
Stat_winding_
height
Rot_crown_heig
Rot_pole_he
ht
ight
R
spacing_winding
_pole
Rotor bore
Rot_winding_
height
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
32
Model Input and Output Parameters



11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
Fixed Parameter
Value between 0 and 1
Binary value
33
Main Algorithm is Implemented in Python
Machine geometries and sources are created in Python using classes for
each component.
Motor
The superconductor operating point is found
using a dichotomy in Python.
PDEParameters
Stator
Python generates the FlexPDE scripts.
Rotor
FlexPDE is used as a solver and creates
output text files imported in the Python
environment.
Supply
WindingRotor
WindingStator
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
34
4X Conductor Characteristic
• For this study, only the
field in the c-axis
direction is considered.
• Engineering current
density Jc(B) was
estimated for tapes
based on:
– 12mm width
– 0.15 mm thickness
– 2760 A @ 2.75 T
@ 35 K
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
35
Design Requirements
• Minimum length of 4X HTS conductor
– Ideally less than 10 km
– 30 K operation
– 30% margin on Jc
• Cost should be more penalized than mass in optimization
• Investigate the impact of
– Rotor-stator gap
– Number of pole pairs
– Aspect ratio
• on
– Generator active mass
– Length of 4X HTS wires
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
36
Example of Data Generated
• Monte-Carlo
exploration of design
space
• Very time consuming
• Need to fix
parameters and
perform optimization
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
37
Examples of Designs Generated
model A
Iron poles
15 km of HTS
model B
Iron poles
Lower rotor field
Thicker stator
7 km of HTS
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
model C
No iron poles
High excitation field
92 km of HTS
38
38
Comparison Between 3 Selected Designs
Model
Number of poles
Air gap (m)
Radius rotor (m)
J Stator (A/mm2)
J Rotor (A/mm2)
Torque (N.m)
Filling factor of the Superconductor
Angle of maximal Torque (rad)
Length of the machine (m)
Active Mass (metric tons)
Mass Winding Stator (metric tons)
Mass Winding Rotor (metric tons)
Mass Core (metric tons)
HTS 4X wire (km)
Max Field on the Superconductor (T)
Max Field in air gap (T)
A
28
0.011
2.6875
1.80E+06
1.16E+08
1.20E+07
10
0.0703
3.299023
217
31.61
1.03
184.75
15.12
1.597
2.051
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
B
28
0.04
2.60
1.80E+06
1.20E+08
1.20E+07
10
0.0809
2.886807
169
48.29
0.53
120.5
7.74
1.80
1.42
C
12
0.04
1.425
1.80E+06
1.50E+08
1.20E+07
4
0.1528
2.0956
135
38.84
2.54
93.52
92.82
4.936
3.635
39
Impact of the Rotor-Stator Gap
• Mass and HTS length increase with air gap length
• Air gap length should be kept as small as possible
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
40
Impact of the Aspect Ratio
• Minimum HTS length for Lactive/Rrotor of 0.5
• Mass increases with aspect ratio
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
41
Impact of the Number of Poles
• Minimum mass for 36 poles
• HTS length decreases with number of poles
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
42
Mass vs. Number of Poles and Aspect Ratio
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
43
Number of Poles for Minimum Mass
32-36 poles
• Optimum number of poles between 32 and 36
• Optimum number of poles fairly independent from L/R
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
44
Conductor Length vs. p and L/R
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
45
Next steps
• Perform parametric sweep/multi-objective optimization
based on cost and mass
– Ongoing (OpenMDAO)
• Complete detailed mechanical and thermal design
• Test of a full scale HTS coil
using 2X conductors
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
46
HTS Coil Operating Point
• Coil performance based on estimated 4X wire
• Ic around 1500 A
• Transverse Peak field of about 1.8 T
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
47
Outline
•
•
•
•
•
•
Introduction
Superconducting generators
HTS conductor choice
Fully Superconducting Generator - DOE
Partially Superconducting Generator - ARPA-e
Conclusion
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
48
Summary
• Availability of low cost HTS conductor is paramount
to the economical viability of HTS wind generators
– Conductor development for wind generators is driven by 2
projects (ARPA-E and DOE) in the US
• 4X improvement in transport current at 3T, 30K in 2G tapes
• Fine filament high current density MgB2 conductors
• Preliminary LCOE calculations show very promising
results for both YBCO and MgB2
• De-risking and modeling activities will take the
technology to the level of full-scale prototype
development
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
49
Aknowledgements
• Advanced Magnet Lab
– Dr. Rainer Meinke
– Mr. Vernon Prince
• University of Houston
–
–
–
–
–
Prof. Selvamanikam
Prof. Denis Netter
Dr. Clement Lorin
Dr. Yaw Nyanteh
Mr. Nicolas Escamez
11th EPRI Superconductivity Conference – Houston, TX - 10/29/2013
50