Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Modelling DC
ERC Starting Grant, 2011-2015
May 2018
Dragan Jovcic
School of Engineering
University of Aberdeen
1. DC transmission networks are required for large-scale offshore energy evacuation,
2. DC grids will be fundamentally different from AC transmission systems,
3. Technical challenges include:• DC voltage stepping
• DC fault isolation
• DC grid control and dynamics
4. DC/DC and DC hubs will play key role in DC grids. They do not exist.
5. Modeling challenges • Numerous converters
• High-frequency dynamics
• Coupling between DC and AC dynamics,
6. Control/stability challenges • DC grid dynamics are two orders of magnitude faster than AC grids
• Distributed control is preferred
1. Background
2
2. DC/DC converters
3Figure 1. Connecting two existing HVDC using a DC/DC converter.
Using DC/DC converter to connect two HVDC lines
• Power trading between two DC lines of different voltage levels,
• Improved operating flexibility,
• Two protection zones, DC faults are not transferred across DC/DC,
• Two HVDC of different manufacturers. DC/DC resolves multivendor issues,
• DC/DC becomes:
• transformer (DC voltage stepping) ,
• power flow regulator,
• DC Circuit Breaker.
DC/DC
+/-400kV
+/-300kV
DC cable (200km)
DC cable
(20-100km)
500MW
DC cable (500km)
+/-300kV
DC cable (300km) DC cable (200km)
2. DC/DC converters
4
Fig.2. 10-terminal DC grid with 3 local radial DC systems and 3 DC/DC
=
Ba-A0
Ba-B0
= AC/DC Converter Station
SA0
SB0
Ba-A1
Ba-B1
Ba-B3
Bm-A1
Bb-B2
Bm-B2 Bm-B3
Bm-F1
Bm-E1
Bb-B1-1
Bb-B4
Bm-C1
Bb-C2
Bb1-D1-1
Cm-A1
Cb1-A1-1
Cb-B2Cm-B2
Cm-B3
Cm-F1
Cm-E1
Cb1-D1-1
Cb-C2
Cm-C1
Bo-C2
Bo-D1
Bo-E1
Bo-F1
800MVA
800MVA
800MVA
800MVA
200MVA
800MVA1200MVA
800MVA
1200MVA
50km
200km
200km
200k
m
200k
m
200km
200km
200km 50km
800M
W
210km
800MW
DC/DC Converter
300k
m80
0MW
280k
m200M
W
200km
800MW
200km
1400MW
300k
m12
00M
W
300k
m12
00M
W
200km1200MW/200km
1200
MW
447k
m
M
DC Cable
DC Overhead line
=DC
DC=
1200MW
450km
1200
MW
730k
m
790
175
324
0
89
86
392
971
324
500
188
103
1200
1200
450
100
4970
750
100
1300
900
618107
0
800MW
500
309
309
1000500
1000
500497
100
500
600
392
800
±400kV±400kV
±400kV±400kV
±400kV
±400kV
±200kV
±200kV
±200kV
±200kV
±400kV
971
447k
m12
00M
W
=
=
=
=DC D
C=
±400kV
=DCDC=
=
=
=
=
1200MVA
1200MVA
790
=
=
800MVA=
500
=DC DC
=
300k
m80
0MW
500
=
=
=
1200MW/200km
750
750
7501200MVA
=
971
1200MVA
1200MVA
971
±400kV
448
783
783
Bb-B2-1 Bb-B2-2
Bb-B1-2
Bb1-D1-2
Cb1-D1-2
Cb1-A1-2
Bb1-A1-1 Bb1-A1-2
Hybrid DC CB
Objectives of Multi-port DC hub in the DC Grid
• Acts like an electronic substation, connecting numerous DC lines,
• Enable any port to trade power with any other port in the DC hub,
• Ability to connect DC lines of different dc voltage ratings, and different manufacturers,
•
• Ability to control power in each DC line,
• Connect/disconnect any DC line “on the fly” without affecting operation of the grid,
• Isolate the faulted DC cable and provide undisturbed operation of the remaining grid,
• DC Circuit breaker may not be required.
3. DC Hubs
5
AC1
DC
HUB
AC2
AC3
1
2
3
~
~
±60kV
±80kV
±120kV
±320kV
±400kV
VSC1
VSC2
VSC3
VSC4
VSC5
AC4
AC5
~
~
~
DC Cable 1
DC Cable 2
DC Cable 3
DC Cable 4
DC Cable 5
Fig.3. 5-terminal DC grid with a 5-port DC hub
DC fault tolerance and high security/redundancy of DC hubs
• Any port is readily disconnected for DC faults,
• Any phase is readily disconnected in case of an internal fault,
(graceful degradation),
• Use redundant phase to meet N-1 criterion (substitute faulted
phase with a spare),
6
V1dc
Port 1
C1
L1 CB1
Port 3
C3
L3CB3
Bus_A
Bus_B
Bus_C
Bus_D
Bus_G
vcA
Port 2
C2
L2 CB2
Port 4ignd
vcB
vcC
vcD
Inner LCL
circuit
V1dc
V4dc
V4dc
V3dc
V3dcC4
L4CB4V2dc
V2dc
Fig.4. 4-port, 4-phase DC hub
3. DC Hubs
-4
-3
-2
-1
0
1
2
3
4
0.9 0.95 1 1.05 1.1 1.15 1.2
Pid
c(p
u)
Time(s)
P1dc
P2dc
P3dc
P4dc
-6
-4
-2
0
2
4
6
0.45 0.5 0.55 0.6 0.65
Pid
c(p
u)
Time(s)
P1dc
P2dc
P3dc
P4dc
7
Advantages:
•DC fault is only a local disturbance
•DC hub inherently reduces fault current,
•DC fault is readily isolated
•No need for fast protection, reliability is high
•Each DC line can have different voltage
level
•Each DC line can have different HVDC
technologies
3. DC Hubs
Fig.5. North Sea DC grid with 4 DC hubs
8
4. DC Grid demonstrators
Fig.6. 5-terminal, 900V, DC Grid demonstrator at Aberdeen HVDC research centre
9
4. DC Grid demonstrators
Fig.7. Testing 30kW, 200V/900V LCL DC/DC converter
a. DC/DC step up operation b. DC/DC step down operation
c. HV side DC fault d. LV side DC fault
10
5. DC Grid modelling
DC Grid modeling challenges:
• Model of N-terminal DC grid is substantially mode complex than N-node AC grid,
• DC grids will include numerous MMC AC/DC, DC/DC converters,
• Simulation based on average modeling (PSCAD,EMTP) is still very slow,
• CIGRE 10 terminal DC grid model, 20s of real time takes 4 hours simulation using average model (20µs).
• Standard simulation platforms (PSCAD,EMTP) support only trial and error in time domain,
• Eigenvalue studies or frequency domain studies are required but not possible,
• Medium frequency (300Hz-1000Hz) will be used in the dc/dc converter to reduce size,
• Simulation step needs to be small to comply with the fastest sampling in the dc/dc converter,
• Multiple DQ frames are required at different frequencies,
• Non-linear elements can not be directly transferred to DQ frame,
11
5. DC Grid modelling
Fig. 8. MMC 10th order DQ analytical model validation.
12
5. DC Grid modelling
Original system
(Kp_PLL=30, Ki_PLL=500)
System with increased PLL gains
(Kp_PLL=300, Ki_PLL=5000)
-14.56 ± j313.2
-17.82± j129.5
-6.98 ± j317
-33.74± j101.3
System 1
(KP_CCSC=0.5, KI_ CCSC =50)
System 2
(KP_CCSC=10, KI_CCSC=50)
-20.1 ± j122.2
-155.0± j637.8
-6.2 ± j129.7
-165.2± j681.6
DQ frame average linearized model enables eigenvalue studies
• High PLL gains cause subsynchronous frequency instability at 45Hz.
• High gains of circulating current suppression controller deteriorate stability at 20Hz
• Status of converter average modelling
Converter Analytical, linearised model
MMC AC-DC 10th order model Developed
MMC AC-DC (blocked state) Challenging
MMC isolated DC-DC (dual active bridge) Complex model
MMC DC-DC Challenging
Line commutated AC-DC converter Developed
13
6. DC Grid control
DC grid control challenges:
•DC grid is more difficult to control than traditional AC systems,
•There is no common frequency, which indicates power unbalance,
•DC voltage indicates global power unbalance but it also changes with local power flow,
•DC grid dynamics are 2 orders of magnitude faster than AC grid dynamics,
•DC grid components have low (overcurrent/undervolatge) tolerances (trip at 0.85pu),
•There are no passive loads with stabilising feedback (lower voltage still draws same current),
•All components are controllable. Numerous control loops,
•No inertia. GW powers should be balanced within 1-2ms.
14
6. DC Grid control
Fig. 10. DC Grid Dispatcher Controller
DC Grid controller demands:
•Grid must be stable without communication with despatcher. Distributed primary response,
•Converters should contribute grid power balancing for a disturbance. Secondary response ensures power
balance (automatic),
•Optimisation and re-dispatching by dispatcher. Tertiary response (human intervention),
Primary/secondary response is critical but may require trade-offs:
•Droop based control enables power balancing but dynamic stability may not be good,
•Fast DC voltage control is required at each terminal, but may not be optimal,
Fig. 9. 3-level controller for DC grid terminals
15
6. DC Grid control
CIGRE DC Grid:
• 5 Offshore VSC terminals,
• 6 onshore VSC terminals,
• 2 DC/DC converters,
• 2 separate DC systems
• One DC system is bipolar,
• Meshed DC lines,
• Onshore AC systems,
Control system requirements:
• automatic power balance,
• optimal operating point,
• stable recovery for large
disturbances,
Fig. 11. CIGRE 11-terminal DC grid benchmark
16
6. DC Grid control
Fig. 12. DC voltages after terminal Cb-A1 outage (1GW loss),
Simulation of an outage of a 1GW VSC terminal:
•Primary response maintains stability. All variables are
within operating limits. No VSC tripping.
•Secondary response balances power. A new operating
point is established within 0.5s.
•Tertiary response re-establishes 1pu average DC
voltage within 2s.
Further research work is required:
1. DC/DC converters as multifunctional DC grid components,• Role in DC grids (voltage stepping, fault isolation, power flow control)
• Topologies, optimisation of losses, reliability, ...
2. DC Hubs (electronics DC substations), • Role in DC grids (voltage stepping, fault isolation, power flow control)
• Topologies, optimisation of losses, reliability, ...
• Control, modelling,
3. Modelling DC grids,• Simulation of grids with many converters,
• Analytical modeling (for eigenvalue studies),
4. Control/stability challenges, • Grid control layers,
• Distributed and centralized control,
• Robustness, flexibility,
5. Hardware (low power) demonstrations,
7. Conclusion – further research
17