CI RED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0369

FAULT CURRENT SOURCE TO ENSURE THE FAULT LEVEL

IN INVERTER-DOMINATED NETWORKS
Frank VAN OVERBEEKE
EMforce RV. - the Netherlands
vanoverbeeke @emforce.nl

ABSTRACT The prevailing standard is to disconnect such generators

from the electricity network in case of a fault in the
network, in order to avoid the formation of electrical
This paper addresses the issue ofensuring thefault level in islands. In many countries network operators demand
inverter-dominated islanded networks. This issue is sophisticated anti-islanding detectors for every small
particularly gaining significance with the advent of generator.
microgrids. It is proposed to add one component to the
network which supplies thefault current when required and Over recent years this standard has been abandoned for
which keeps the network alive long enough for all inverters large wind farms in regions where a significant portion of
to resume operation, if they have switched off during the the regional load is supplied by local distributed generation.
transient. This component is named a Fault Current Source It has been recognized that keeping the wind farms
(FCS). connected during and immediately after a fault in the
network, is essential to maintain the stability ofthe network.
If the network is in normal operation, the FCS charges its There are good reasons to presume that this approach will
internal energy buffer and then remains dormant. As soon also penetrate into the domain of medium and low voltage
as a phenomenon is detected which is representative ofa networks in the near future.
fault, the FCS attempts to restore the system voltage which
was present just before the fault, thereby injecting as much Many low-voltage generators like PV systems, small wind
current as necessary. Once the fault is cleared, the FCS turbines and domestic CHP, are owned by individual home
provides the reference needed by all inverters to restart owners. These individuals seek ways to optimize the
operation. Subsequently the FCS switches offagain. "profit" (which is not perceived as a financial benefit only)
from their generator :

A prototype FCS has been built and tested Preliminary test • by sharing electricity with their close neighbours
results are presented in this paper, full results are to be without having to pay the distribution network
presented during the conference. operator disproportionally for delivering their
power next door;
• by using their private generator as a back-up in
case of a fault in the electricity distribution
INTRODUCTION INTO MICROGRIDS
network;
One of the emerging network concepts in the field of • by offering some kind of control over their
distributed generation is the microgrid. A microgrid is a generator to the DNO or to an energy service
network, usually at low voltage level, in which the operation company (ESCO) , to be used for example as
of network, loads, local generation and local storage is balancing power. In such cases the DNO or
coordinated to deliver particular services to its customers . ESCO may offer a financial compensation in
Such services could be optimal fuel efficiency, optimal cost return.
of electricity, or increased reliability by enabling islanded
operation. The Microgrid philosophy has been developed as a set of
concepts and tools with which these services can be
The main driver behind this concept is the increasing implemented . In the ED 5th and 6th Framework
amount of decentralised electricity generation, e.g. from Programmes, two projects have been executed covering all
wind turbines, solar PV systems and CHP systems. These aspects ofthese concepts . The aspects addressed range from
generators have been installed as a consequence of fundamental electrical issues like dynamic stability in the
government incentives aiming to increase the use of millisecond domain and black start sequences up to
electricity from renewable sources and to reduce the CO2 regulatory issues and market models. All publications from
contribution of electricity generated from fossil fuels. these projects are available from www.microgrids .eu.

CIRED2009 Session 4 Paper No 0369

CI RED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0369

PROTECTION IN A MICROGRID specific country and which would be purchased by the

customer at the local DIY store if he were to expand his
own switchboard. The fault level provided by the microgrid
One of the main benefits of a microgrid is the option to run must then support the operating currents and clearing times
a local or regional network in islanded mode. Nevertheless, needed by such devices .
most of the time the microgrid will be operated in grid-
connected mode. Moreover islanding may involve some Commercially available PV inverters do not comply with
kind of degraded mode, such as a loss of power quality or this recommendation. These inverters switch off
temporary disconnection of non-essential loads. immediately whenever the output current exceeds the rated
current by a relatively small percentage, for example 20% .
However, even in degraded mode the safety ofthe network The same argument applies to inverters used in for example
must not be compromised. In particular this means that microturbines and in small wind turbines. Therefore a
short-circuit faults must detected and isolated within the microgrid needs preferably at least one generator delivering
prescribed time windows, in order to prevent overheating of fault ride-through on any short-circuit behind a customer's
conductors or unsafe levels of touch voltages . On the other switchboard.
hand it is desirable that the microgrid does not collapse as a
consequence of a simple short-circuit fault within the In the sample network analyzed in the MICROGRIDS
premises ofa single customer . That is what would happen in project [1], this role is assigned to the flywheel storage
a microgrid which is dominantly supplied from inverters system which is used to absorb the transient after a sudden
due to their ultra-fast overcurrent protection. disconnection of the microgrid from the external
distribution network. However from the analysis made, it
Analysis of these requirements for a sample low-voltage follows that the flywheel's inverter must be able to sustain a
microgrid [1, 2] has led to the following conclusions: factor 3 overload for several seconds. This may pose a
problem when trying to purchase such a system in the
1. Generally the size of an LV microgrid is such that there market. Moreover, not all microgrids may be equipped with
is no added value in sectionalizing the network when a fault a storage system.
occurs in the network itself. Only in a radial network with
tapered feeders , this approach may be necessary to protect FAULT CURRENT SOURCE
the lower-diameter conductors at the far end ofthe feeders.
As a consequence, it is accepted that the microgrid loses A Fault Current Source (FCS) is a network component
power after a fault in the microgrid's network. This specifically intended to provide additional fault level to a
approach also avoids the issue of misoperation of network. The component was developed with inverter-
protections as a consequence of reverse current flows. dominated networks in mind, but is also applicable in
networks which lack fault level because of long supply
2. The fault level of the microgrid must be sufficient to lines.
operate protection devices behind any customer's meter, so
that a fault within the customer's premises does not Operation
propagate into the network.
The operation of an FCS is illustrated in figure 1. Its power
It is strongly recommended to use standard protection circuit remains dormant during normal operation of the
devices within the customer's installation. By "standard" we network (1). Whenever a fault occurs (2), it is activated (3)
refer to protection devices which are normally used in that and attempts to restore the original voltage, thereby

Figure 1: Voltage in the network in normal operation (1), when the faults occurs (2) and after triggering of the FCS (3).

CIRED2009 Session 4 Paper No 0369

CI RED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0369

injecting a fault current into the network. This current will Triggering circuit
cause a fuse or circuit breaker to clear the fault. The key component of an FCS is a circuit which senses the
Subsequently the FCS maintains the original voltage and state of the network and starts the inverter whenever it
frequency, to enable inverters which had turned off, to detects a fault. In order to be able to operate an FCS as a
resynchronize and reconnect to the network. After some plug-and-play device, this circuit should act only on locally
time, typically 5 seconds, the FCS turns off. measured information. Technical details of the detection
algorithm will be published later because of a pending
Location patent application.

Once triggered, the FCS acts like a low-impedance voltage Charging module
source. The most logical location of an FCS in the network The energy storage device is charged with electricity from
is therefore the main distribution busbar, which would be the network. This must be done in a controlled manner,
right next to the distribution transformer if the network is because high inrush or charging currents are generally
normally supplied from an MV network. unwanted in distribution networks with a low fault level.
On the other hand, after having operated the FCS must be
Implementation recharged as soon as possible to be ready for the next fault.
Depending on the topology of the system, the power
An FCS consists of four subsystems: electronic converter can be used to recharge the storage
• An electricity storage device device.
• A power electronic converter
• A triggering circuit Modularity
• A charging module The FCS is connected in parallel to the network and acts as
a current-limited voltage source when triggered. Therefore
Electricity storage device it is relatively simple to use multiple devices in parallel. For
An FCS is applied ifthe network by itself cannot provide a example, a typical unit would have a rating of 500 A peak
fault current during a fault. As an LV network is dominantly current and an available stored energy of 100 kJ. If a
resistive, a fault current is associated with a significant network requires a higher current, or if for reasons of
amount of real power which has to be injected into the reliability the network operator prefers to have redundant
network for the time needed to clear the fault. The energy units, two or more devices can be connected in parallel.
required to deliver power during a certain time must be
stored inside the FCS. The energy could be stored in a COMPARSION WITH SIMILAR DEVICES
flywheel, a battery, an ultracapacitor or any other
technology which is suitable for fast discharges. It is helpful to defme the role and design of an FCS
compared with other Power Quality related devices.
Power electronic converter
When the FCS is triggered, the available energy must be Flicker compensators
released into the network in a controlled manner. The best
way to do this is by trying to maintain the voltage and The FCS is connected in parallel to the distribution network.
frequency of the network as they were prior to the fault. In This is very similar to the topology applied for a Power
this way, the original voltage is automatically restored as Quality Optimizer (PQO) which was introduced by
soon as the fault is cleared by the appropriate protection EMforce in 2007 [3]. The main difference is that a PQO is
device. Therefore a power electronic converter is used as an in operation continuously and is rated to supply relatively
interface between the energy storage device and the small currents, whereas an FCS operates only when
network. In many cases the electrical energy is available as triggered and can supply fault currents. The devices have in
DC. The distribution network is AC, so that the interface is common that the thermal design has been optimised for the
a DC to AC inverter. delivery of short peak currents. A PQO needs some energy
In LV distribution networks most faults behind the customer storage, but much less than required for an FCS.
meter are single phase to earth/neutral faults. Therefore the
inverter must provide a neutral connection and preferably Static VAR Compensators
act independently for each phase. This approach minimizes
the annoyance in the other two phases during a single-phase A Static VAR Compensator (SVC) is an inverter, connected
fault. in parallel to the network to inject or absorb reactive power.
An SVC is rated for continuous operation and has a thermal
design which is similar to a traditional inverter. It is
generally not designed to deliver fault currents. It does not

CIRED2009 Session 4 Paper No 0369

CI RED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0369

supply active power and therefore its stored energy is very

small. SVCs are rarely used in LV networks.

Dynamic Voltage Restorers

A Dynamic Voltage Restorer (DVR) is a device connected

in series with a voltage-sensitive feeder or load. It provides
correction of the amplitude and sometimes even the
waveform of the voltage supplied from a network. A DVR
is equipped with some electricity storage to be able
reconstruct the voltage if the network does not supply
enough power during a transient. A DVR operates only on
transients and remains dormant during normal operation of
the network. However because of its series connection, the
semiconductors of a DVR have to carry the load current
continuously and the device has to be thermally rated for
continuous load. Generally a DVR does not significantly Figure 3: Prototype FCS.
increase the fault level of a network.
Figure 3 shows the prototype in its initial assembly. From
PROTOTYPE left to right the ultracapacitor stacks, the electrolyte buffer
capacitors , and the inverter can be seen. The charging
In order to demonstrate the concept of an FCS and test its circuit and control electronics are mounted on the back side
properties in a practical configuration, a prototype was built of the assembly and are not visible from this side.
using a 500 A single-phase inverter. A simplified circuit
diagram is shown in figure 2. An ultracapacitor stack which CONCLUSIONS
can deliver approx . 100 kJ is used as an energy buffer. A
topology has been chosen with a single half-bridge and the A Fault Current Limiter is a promising device to provide
network neutral connected to the centre tap ofthe inverter's fault level in inverter-dominated distribution networks. The
buffer capacitors. The pre-charging circuit consists of a concept is plug-and-play and can be used in a modular
transformer supplying 2 kW at a relatively low voltage. manner.
Once the buffer has been charged to approx. 100 V, the
inverter is used as a boost converter to further charge the Preliminary testing of the prototype has proved the
buffer. As soon as the buffer voltage has exceeded 600 V, capabilities of the FCS and the quality of the detection
the transformer is disconnected and the main contactor is algorithm. The correct parametrization of the algorithm is
closed, connecting the inverter phase directly to the underway. Full test results will be presented during the
network. The inverter then continues charging the buffer to conference.
a maximum value of 860 V.
Acknowledgement
The buffer consists of 320 ultracapacitor cells, each rated
140 F / 2.7 V. Proper voltage sharing between the cells is This work was supported by the European Commission
ensured by parallel resistors. As these have been chosen (EU) within the framework of EU Project MORE
rather conservatively, the buffer has a relatively high self- MICROGRIDS , Contract No. 019864 (SES6).
discharge rate; the time constant is approx. 9 hours. In this
application such a value is not prohibitive. REFERENCES
[1] X. Wu, N. Jayawarna, N. Jenkins, "Electrical
Protection Schemes for the MicroGrid",
L,---- , -----../ - - - , - phase L1 MICROGRIDS project report, 20 July 2004.
Ultracapacitor
[2] N. Jayawarna, M. Lorentzou, S. Papathanassiou, N.
Hatziargyriou, "Safety Guidelines for a Microgrid",
f------cf------'--------'------'- Neutral MICROGRIDS project deliverable DEI , November
2004.
[3] F. van Overbeeke, B. Tanovic, T. Meeks, "Mitigation
of flicker in a rural LV network", Proceedings
ClRED2007, paper 0057.
Figure 2: Circuit diagram of the prototype FCS.

CIRED2009 Session 4 Paper No 0369