DC Fault Protection Strategy Considering DC Network

Partition
Md Habibur Rahman, Lie Xu Liangzhong Yao
Department of Electronic & Electrical Engineering China Electric Power Research Institute
University of Strathclyde, Glasgow, United Kingdom Xiaoying Road, Beijing, 100192, China
[email protected], [email protected] [email protected]

Abstract—This paper investigates DC network partition and isolate the faulty line in a selective manner allowing fast
alternative DC fault protection strategy for Multi-terminal restoration of normal system operation following a DC
HVDC (MTDC) system. Fast acting DC Circuit Breakers fault[3-6].
(DCCBs) or fault blocking DC-DC converters can be configured
at strategic locations to allow the entire MTDC system to be There are several protection methods that have been
operated interconnected but partitioned into islanded DC proposed for MTDC system[3-7]. A protection method of
network zones following faults. In case of any DC fault event, VSC based MTDC system was discussed in [7] in which a
the DCCBs or DC-DC converters at the strategic cable ‘Handshaking’ method using AC circuit breakers and DC
connections that link the different DC network partitions are switchgear were proposed. But the system recovery is slow
opened or blocked such that the faulty DC network zone is which can pose significant operational problems for large
quickly isolated from the remaining of the MTDC system. Thus, scale MTDC systems and connected AC networks due to the
the healthy DC network zone can remain operational or recover loss of the entire network. Fast acting DCCB which is capable
quickly to restore power transmission. Each DC network zone of operating within a few milliseconds can be adopted to
can be protected using AC circuit breakers and DC switches for isolate the faulty cable such that the healthy part of the DC
cost reduction. The validity of the proposed protection strategy network can continue operating or quickly recover. However,
is confirmed using MATLAB/SIMULINK simulations. such DCCBs are likely to have high capital cost, larger
footprint and high on state losses, and thus, their use should be
Index Terms-- DC fault, DC network partition, DC Circuit
Breaker, Multi-terminal HVDC System.
limited. Introducing MMC based DC-DC converter into a
MTDC system in terms of protection can bring significant
I. INTRODUCTION benefit to the entire system. There are many DC-DC converter
configuration have been proposed [8-11] for MTDC system.
The increased global energy demands and the utilization of The main advantages of using DC-DC converters are their
renewable energy generation have raised new requirements for ability to connect DC networks of different voltage levels and
the future electricity grid connection. The development of control the power flow, isolate the fault quickly etc. In case of
renewable energy and increasing the security of supply require any fault event the DC-DC converter rapidly isolates the
large scale offshore and onshore network integration leading faulty section such that the healthy part of the system remains
to transmit large amount of power over a long distance. operational and unaffected.
HVDC becomes a more preferable choice in terms of
transmitting a bulky amount of power over a long distance due In this paper, different MTDC system configurations are
to the transmission losses and smaller cable size for given analysed in terms of DC fault protection considering the
power level when compared to traditional HVAC transmission minimal use of DCCBs or DC-DC converter at strategic
system. locations to allow the entire MTDC system to be operated
interconnected but partitioned into islanded DC network zones
Multi-terminal HVDC system (MTDC) using VSC following faults. The paper is structured as follows: Section II
technology has greater flexibility for large-scale renewable describes the fault behaviour of an MMC based converter. DC
energy integration and transmission network connection due network configuration and different protection options are
to its ability for independent power control both active and outlined in section III and the simulation studies of the
reactive power, AC voltage support, and black–start proposed concept are presented in section IV and section V
capabilities. However, there is a major challenge toward the draws the conclusions.
protection of an MTDC system in the event of a fault at the
DC side of the network including fault protection, fault II. DC FAULT BEHAVIOUR
location and isolation[1-3]. As the rate of rise of DC fault DC faults can cause serious consequences due to the low
current is very high due to the low impedance of the DC impedance of the DC network and the existence of the
network, the protection system has to act fast and an effective freewheeling diodes in half-bridge MMC converters in
protection method needs to detect the fault and its location and HVDC system[12-14]. The simplified equivalent circuit of a

This work is supported in part by China Electric Power Research

Institute (CEPRI)
half bridge MMC during a DC line-to-line fault is shown in III. DC NETWORK PARTITION AND DIFFERENT
Fig. 1. Unlike the two-level VSC, MMC does not have a large PROTECTION OPTIONS
DC link capacitor at converter terminal though cable capacitor
A. DC Network Partition
is still present.
One of the main concerns for large scale MTDC system is
ipa ipb ipc idc Lcable Rcable the potential loose of power transmission across the entire
T1 D1
C0
SM1 SM1 SM1 network in the event of a single fault on part of the DC
VSM
Vc Vpa network. This issue can be partially mitigated by installing fast
D2 SM2 SM2 SM2
T2
acting DCCBs at every DC line connection points though with
SMN
Sub-Module SMN SMN
large increase of system cost. The infrequency of DC fault
Larm Larm Larm events and the inconsistency of power generation from wind
AC System Transformer ia
Vdc farms, the expected cost of losing access to that energy may
ib
ic
not be sufficient to justify huge investment in protection cost.
To justify the cost and reliable protection, the entire MTDC
400kV 1000MVA Larm Larm Larm
SCR=10 400kV/400kV system can be configured in a number of partitions (like
20% SM1 SM1 SM1 different DC network zones within the MTDC system) where
MMC Station Parameter
Vna SM2
a fault anywhere on a particular partition would result in the
SM2 SM2
Arm inductance: 10% partition being isolated by clearance from “slow” DC switches
Number of cells per arm (N): 380
Cell Capacitance: 9.4mF SMN SMN SMN and circuit breakers on the AC side[15] .
ina inb inc Lcable Rcable

Fig. 1. Equivalent circuits of half bridge MMC during DC line-to-line fault

AC System
Option1 or 2
DC Network
DC Cable Zone1
AC System AC System

AC System
AC System
AC System
DCCB
Option 1
DC Network
DC Cable
Zone3
DCCB DCCB
Option 2
AC System DC Cable DC Cable

AC System DC-DC Converter

Fig. 3. Possible DC network partition using least number of DCCBs or DC-

DC converter

In the event of a DC fault it is required from system

Fig. 2. DC current and voltage of MMC during DC line-to-line fault operators’ point of view to keep the permanent ‘loss of infeed’
below the maximum power loss criterion (e.g. 1.32GW, as is
A typical half-bridge MMC based converter has been the present case in GB) to maintain system stability.
considered for analysing the DC fault behaviour. A permanent Therefore, the partitioning of the MTDC system should be
DC line-to-line fault is applied at 1.1s and the MMC based configured in such a way that the power generation export
converter is blocked 1ms after the fault initiation. Fig. 2 shows from offshore is maximised in the event of fault. The power
the system response during the whole fault period. Long (Figs. being supplied from any partition (any DC network zone) to
2(a)-(c)) and short (Figs. 2(d)-(f)) duration time-scale any AC system to which it is connected is less than the limit
waveforms have been presented for ease of analysis. Figs. 2(a) for a permanent ‘loss of infeed’ limit of that AC system[15].
and (d) show the collapse of the converter DC link voltage However, such partitioning reduces the operational flexibility
with oscillations immediately after fault initiation. Due to the of the MTDC system. The aforementioned DC network
arm reactance the DC link voltage can reach to a negative partitioning can be more optimised by placing fast acting
value. Figs. 2(b) and (e) represent the converter DC currents DCCBs or DC-DC converter at strategic location connecting
showing rapid increase after fault initiation. As the MMC different DC network zones to allow the entire MTDC system
converter behaves as an uncontrolled rectifier after IGBT to be operated interconnected pre-fault but partitioned into
blocking, the DC link current cannot decay to zero until AC islanded sections (DC network zones) following faults. A
circuit breaker is opened. In this simulation the tripping time typical DC network configuration is shown in Fig. 3 where
of the AC circuit breaker is set to be 80ms after the converter only limited numbers of fast acting DCCBs or DC-DC
arm current is detected. Until open AC circuit breaker the converters are used to reduce the capital cost of the entire
arms current have to withstand huge over current. Figs. 2(c), MTDC system. In case of any fault event within one DC
and (f) shows large AC currents during DC fault which are network zone at least two of the DC network zones can be
flowing through the anti-parallel diodes.
remain operational or restored quickly after isolating the faulty 1000MVA 1GW
400kV/400kV
800kV DC Link (Option 1)
640kV DC Link (Option 2) 100mH
1GW
1000MVA
400kV/400kV
1GW 100mH
zone depending on the fault location. ACCB 20%
100km
20%
ACCB

B. Protection Option AC System

400kV Station1
100mH
DC Network Zone 1
100mH
Station2
AC System
400kV
SCR=5 MMC Based MMC Based SCR=10
There are a few possible options to clear the DC side faults Option 1 Option 2

Cable
FDCCB1P

FDCCB1N
without causing a large loss of infeed. Different protection

600MW
strategies for MTDC system have been described and analysed

Option 1 or 2

80km
in[3-11, 16-18]. The main purpose of this work, in case of any

Cable
fault event in one DC network zone, the healthy zone can be

600MW
FDCCB3N
FDCCB3P
continued operational or will achieve normal operation after

Cable
isolating faulty zones by means of using the equipped DCCBs
or DC-DC converter. The faulty DC network zone can be 1000MVA
400kV/400kV 1GW 100mH
800kV DC Link
100mH
1GW
1000MVA
400kV/400kV
protected using AC circuit breakers and DC switches. The ACCB 20%
1GW 20%
ACCB
100km
following steps have been taken for the proposed system to AC System AC System
100mH
clear a DC fault.
100mH
400kV Station3 Station4 400kV
SCR=7 MMC Based MMC Based SCR=15

50km
1. Using local current measurement to detect the fault current
flowing through DCCBs and converter arms. FDCCB=Fast DC Circuit Breaker
ACCB=AC Circuit Breaker
DC Cable Parameters
100mH
500MW
500MVA
400kV/400kV
20%
2. If the fault current goes above pre-determined set value the MMC=Modular Multi-level Converter Resistance: 0.009Ω/km
Inductance: 1.4mH/km
ACCB

fast acting DCCBs will be set to open (with a 5ms delay Capacitance: 0.23μF/km
100mH AC System
for this study). In addition, a converter will be blocked if DC Network Zone 2
Station5
MMC Based
400kV
SCR=20

its arm current reaches its threshold protective level.

3. After isolating the faulty zone the system may require to Fig. 4. Block diagram of a proposed five terminal MMC based MTDC system
restart for the normal operation depending on fault
location.
A. Case Study 1 - Option1 (DCCBs)
4. The faulty DC network zone can be protected using
ACCBs. The obtained results representing the system’s behaviour
are presented in Figs. 5(a)-(e), where the responses of the DC
IV. SIMULATION STUDIES voltage magnitudes (as recorded on the DC Link voltage of
Fig. 4 shows the five-terminal MTDC system considered each converter) to the fault are shown.
in the paper consisting of MMC based converter connected to
AC systems. The system contains two DC network zones
which are interconnected by DC cables equipped with fast
acting DCCBs or DC-DC converters. No DCCBs are used
within each DC network zone so as to minimise the cost and
power loss. Within each zone, the DC network zone is
protected using slow AC circuit breakers and DC switches.
The proposed two protection options, i.e. one uses DCCBs and
the other uses DC-DC converters, are applied to the MTDC
system shown in Fig. 4 and simulated using MATLAB-
SIMULINK. Modelling an MMC converter in detail switching
mode requires large computational efforts, thus, average
models with controllable voltage and current sources are used,
and additional semiconductor devices are added to ensure that
the model accurately replicates all possible current paths in a
real converter during DC faults. On this configuration π model
of the cable is used.
The Station 1 and 3 are assigned to transmit 800MW and
600MW power to the DC grid, respectively. Station 5
transmits 300MW power to the AC grid whereas Station 2 and
4 regulate the DC link voltage (800kV for Option 1 and for
Option 2, Zone 1 and Zone 2 DC link Voltages are 640kV and
800kV respectively ) of the entire MTDC system using DC
droop control to ensure effective active power sharing
between the two. For option 2 DC-DC converter’s Station B is
designed to transmit 200MW power from Zone 1 to Zone 2
while Station A is set for controlling the internal AC source.
For simplicity, each converter operates at unity power factor. Fig. 5. System behaviour on the DC side during a DC fault at 2.1s
A DC line-to-line fault is applied at the midpoint of the
transmission line between Station 1 and 2 which is 50 km
away from both stations at 2.1s.
 It is evident from Fig. 5, that the DC fault has severe B. Case Study 2 - Option 2 (DC-DC Converter)
impacts on the entire MTDC system which lead to DC The main concept of this protection arrangement is that, in
voltages collapse and increase in DC link current. The DC case of any fault events in one DC zone, the DC-DC converter
link voltages on each side of the faulty line decrease faster can quickly isolate the faulty zone by blocking its converter
than the other DC link voltages. Figs. 5(f)-(j) demonstrate the such that the healthy zone can remain operational all the times.
DC link current during the fault. It can be seen that the fault
peak current is higher in Station 1 and 2 compare to others The obtained results representing the system’s behaviour
due to the distance of the fault location. are presented in Figs. 7 and Fig. 8. It is evident from Fig. 7,
that the DC fault severely affects the DC voltage in Zone 1
Fig. 6 represents the upper arm currents and converter-side and results in the rapid increase in the DC link current. The
AC currents respectively. Converter blocking is activated DC over current flowing through the DC-DC converter is
when the measured arm currents exceed the pre-determined quickly detected which results in the immediate block of the
set value. In this simulation study Station 1 and 2 are blocked DC-DC converter. This effectively isolate Zone 1 from the
4ms after the fault initiation while converter stations 3-5 healthy Zone 2 and it can be seen that there is little impact on
remain operational. After blocking the converter Station 1 and the DC voltages and currents in Zone 2 due to the fast
2, the AC currents continue increasing (see Figs. 6(f) and (g)) blocking of the DC-DC converter.
through the freewheeling diodes.

Fig. 6. Upper arm and converter-side AC currents during a DC fault at 2.1 s

In this simulation FDCCB1P and FDCCB1N are opened

7ms after fault initiation (2ms detection time plus 5ms
opening delay) whereas for FDCCB3P and FDCCB3N, it is
9ms. After the opening of the fast acting DCCBs, the
complete MTDC system is split into two isolated DC network
zones. For the faulty Zone 1, the AC circuit breakers of
Station 1 and 2 opens around 84ms after the fault initiation,
and the AC current, the arm current and the DC currents
gradually then decay to zero. For DC Zone 2, all the Fig. 7. System behaviour on the DC side during a DC fault at 2.1s
converters remain operational during the fault period though
with some fluctuations in DC voltages, DC link currents, arm Fig. 8 represents the upper and converter-side AC currents
and AC currents. (See Figs. 5 and 6). Despite this fact, all respectively. In this proposed system as soon as a fault occurs,
currents and voltages of the DC Zone 2 remain within only the converters located in Zone 1 and DC-DC converter
acceptable tolerance band. Power transmission continues in stations have been blocked in a selective manner using
DC Zone 2 though power flow was automatically altered due automatic detection and blocking method. In this simulation
to the removing of the interconnection between Zone 1 and 2. study Station 1, 2 and DC-DC converter stations are blocked
their converters at 3ms, 7ms and 6ms respectively after the REFERENCES
fault initiation. After blocking the converter, the AC current [1] E. Kontos, R. T. Pinto, S. Rodrigues, and P. Bauer, "Impact of HVDC
continue increasing (see Fig. 8 (h) and (i)) through the Transmission System Topology on Multiterminal DC Network
freewheeling diodes. Here DC Zone 1 is protected using AC Faults," IEEE Trans. on Power Delivery, vol. PP, pp. 1-1, 2014.
circuit breakers. In this simulation study AC circuit breakers [2] N. Yousefpoor, S. Kim, and S. Bhattacharya, "Control of voltage
equipped in Station 1 and 2 are opened at 83ms and 87ms source converter based multi-terminal DC grid under DC fault
operating condition," in Energy Conversion Congress and Exposition
respectively, after the fault initiation. (ECCE), 2014 IEEE, 2014, pp. 5703-5708.
[3] L. Tang and B.-T. Ooi, "Protection of VSC-multi-terminal HVDC
against DC faults," in Power Electronics Specialists Conference,
2002. pesc 02. 2002 IEEE 33rd Annual, 2002, pp. 719-724 vol.2.
[4] W. Lu and B.-T. Ooi, "DC overvoltage control during loss of
converter in multiterminal voltage-source converter-based HVDC (M-
VSC-HVDC)," IEEE Trans. on Power Delivery,, vol. 18, pp. 915-920,
2003.
[5] J. Yang, J. E. Fletcher, and J. O'Reilly, "Multiterminal DC Wind Farm
Collection Grid Internal Fault Analysis and Protection Design," IEEE
Trans. on Power Delivery, vol. 25, pp. 2308-2318, 2010.
[6] H. Liu, Z. Xu, and Y. Huang, "Study of protection strategy for VSC
based HVDC system," in Transmission and Distribution Conference
and Exposition, 2003 IEEE PES, 2003, pp. 49-54 Vol.1.
[7] L. Tang and B.-T. Ooi, "Locating and Isolating DC Faults in Multi-
Terminal DC Systems," IEEE Trans. on Power Delivery, vol. 22, pp.
1877-1884, 2007.
[8] M. Hajian, D. Jovcic, and W. Bin, "Evaluation of Semiconductor
Based Methods for Fault Isolation on High Voltage DC Grids," IEEE
Trans. on Smart Grid, vol. 4, pp. 1171-1179, 2013.
[9] R. Zeng, L. Xu, and Y. Liangzhong, "DC/DC Converters Based on
Hybrid MMC for HVDC Grid Interconnection," in AC and DC Power
Transmission, 11th IET International Conference on, 2015, pp. 1-6.
[10] I. A. Gowaid, G. P. Adam, A. M. Massoud, S. Ahmed, D. Holliday,
and B. W. Williams, "Quasi Two-Level Operation of Modular
Multilevel Converter for Use in a High-Power DC Transformer With
DC Fault Isolation Capability," IEEE Trans. on Power Electronics,
vol. 30, pp. 108-123, 2015.
[11] D. Jovcic, M. Taherbaneh, J. P. Taisne, and S. Nguefeu, "Developing
regional, radial DC grids and their interconnection into large DC
grids," in PES General Meeting Conference & Exposition, 2014 IEEE,
2014, pp. 1-5.
[12] Y. Gao, M. Bazargan, L. Xu, and W. Liang, "DC fault analysis of
MMC based HVDC system for large offshore wind farm integration,"
in Renewable Power Generation Conference (RPG 2013), 2nd IET,
2013, pp. 1-4.
[13] J. Rafferty, L. Xu, and D. J. Morrow, "DC fault analysis of VSC based
multi-terminal HVDC systems," in AC and DC Power Transmission
(ACDC 2012), 10th IET International Conference on, 2012, pp. 1-6.
[14] J. Rafferty, L. Xu, and J. Morrow, "Analysis of voltage source
converter-based high-voltage direct current under DC line-to-earth
fault," Power Electronics, IET, vol. 8, pp. 428-438, 2015.
[15] K. R. W. Bell, L. Xu, and T. Houghton, "Considerations in design of
an offshore network," in CIGRE Science & Engineering, Feb 2015,
pp. 79-92.
Fig. 8. Arm currents and converter-side AC currents during a DC fault at 2.1s [16] W. Ahmed and P. Manohar, "DC line protection for VSC-HVDC
system," in Power Electronics, Drives and Energy Systems (PEDES),
2012 IEEE International Conference on, 2012, pp. 1-6.
V. CONCLUSIONS [17] C. D. Barker and R. S. Whitehouse, "An alternative approach to
DC network partition with minimum use of fast acting HVDC grid protection," in AC and DC Power Transmission (ACDC
2012), 10th IET International Conference on, 2012, pp. 1-6.
DCCBs or a DC-DC converter at strategic locations of the [18] F. Page, S. Finney, and L. Xu, "An alternative protection strategy for
entire MTDC system are proposed and studied in this paper. multi-terminal HVDC," in 13th Wind Integration Workshop, Berlin,
The main purpose of the protection options is to isolate DC Nov.2014.
faults, without risking the continued operation of the healthy
part of the entire MTDC system. The simulation results
corresponding to DC fault protection have been presented
give a satisfactory result. The proposed concept based on a
five terminal MTDC system have been studied here, can be
extended to different terminal numbers and DC network
partitions which could be cost effective.