Bremen Workshop 2009

8/13/2019 Bremen Workshop 2009

1/8

AbstractIncreasing levels of wind penetration in

autonomous power systems has set intensively high standards

with respect to wind turbine technology during the last years.

The special features of non-interconnected power systems make

security issues rather critical and operation of wind farms as

conventional power plants is becoming a necessity as wind

turbines replace conventional units in the production side. This

paper includes the study case of Rhodes island, in Greece, where

rapidly increasing wind penetration has started to impose serious

security issues for the immediate future. The scenarios studied

here correspond to reference year of study 2012 and include wind

farms with three different wind turbine technologies namelyDoubly Fed Induction Generator (DFIG), Permanent Magnet

Synchronous Generator (PMSG) and Active Stall Induction

Generator (ASIG) based wind turbines. Aggregated models of

the wind farms are being used and results for different

simulation cases are being analyzed and discussed. The ability of

wind farms to assist in some of the power system control services

currently carried out by conventional synchronous generation is

being investigated and discussed. The power grid of the island,

including speed governors and automatic voltage regulators, was

simulated in the dedicated power system simulation program

Power Factory from DIgSILENT.

Index Terms--wind power penetration, autonomous systems,

uninterrupted operation, frequency control.

I. INTRODUCTIONTechnical and regulatory issues regarding the interaction

between large wind farms and power system are underconstant discussion, as wind power penetration increases in

modern power systems. The requirements, that wind farms

have to fulfil, are being continuously updated. Among these,

voltage and frequency control play an important role. Voltage

control and reactive power control of variable speed wind

turbines has been mainly under focus, while frequency control

has started to appear as emerging need under increasing wind

power penetration conditions and due to the extended

replacement of conventional generators by large wind farms in

power supply. The impact of wind farms in frequency

phenomena is even more vital in non-interconnected powersystems, where the power system inertia is limited. Short

circuits at the grid and faults, i.e. sudden loss of the largest

conventional unit in the system, are under study in this paper

1 Authors are with National Technical University of Athens, School of

Electrical and Computer Eng. Electric Energy Systems Lab2Authors are with Ris National Laboratory, Wind Energy Department, P.O.

Box 49, DK-4000 Roskilde, Denmark3Author is with Public Power Corporation S.A. Athens, Greece

The auxiliary services provided by wind turbines, like fault

ride through capability (FRT) and frequency control, are

investigated in this article through detailed modelling for all

different components of the system. Dynamic security of

power systems has to be carefully examined, before windpower penetration limits are expanded. The response of

conventional units, the load dependency on frequency and

voltage and the wind turbines response during events that

affect system frequency are some of the key aspects that have

to be modelled in detail for this kind of investigations.

For these investigations, the study case of the Rhodespower system is used. Three different types of conventional

generators gas, diesel and steam units and three different

types of wind turbines Active Stall Induction generator

(ASIG), Doubly Fed Induction generator (DFIG) and

Permanent Magnet Synchronous generator (PMSG) windturbines are included in the simulation platform, which was

developed. This variety of components gives the chance for a

wide range investigation of key issues for modern power

systems.

In Section II of the article, the Rhodes power system and

the basic modelling principles applied in this survey are also

given. In Section III, the basic aspects of dynamic security in

isolated power systems are analyzed and definitions regardingfault ride through capability and frequency control in power

systems are given. Section IV analyses the results for voltage

drops in the system due to short circuits and frequency

deviations due to sudden loss of generation. The effect ofancillary services of modern wind farms on the power system

operation is demonstrated.

II. POWER SYSTEM MODELA. Rhodes power system

The system under study is the small size island system ofRhodes. Rhodes power system production for the reference

year 2012 includes two power plants and five wind farms. In

2012, the total installed wind power capacity and the

maximum annual power demand are assumed to be about 48

MW and 233 MW, respectively (see Table I).

The basic characteristics of Rhodes power system in 2012

are summarized in Table I:

Investigating power control in autonomous

power systems with increasing wind power

penetrationI. D. Margaris

1, A. D. Hansen

2, P. Sorensen

2, N. D. Hatziargyriou

1,3


2/8

TABLE IBASIC CHARACTERISTICS OF RHODES POWER SYSTEM IN 2012

Rhodes power system

Max Power Demand (MW) 233.1

Rated Thermal Power (MW) 322.9

Rated Wind Power

Generation (MW)48.8

Power system simulation studies for 2012, were based on

modified operational data and additional generating units andwind farms, which are expected to be online by the year of

study, 2012, [1]. The protection system, mainly under/over

frequency and voltage protection relays, is also included in the

dynamic power system model.

The load scenario, which was chosen to use in the

simulations, is the Maximum Wind Power Penetration

scenario (in percentage of the load demand). The penetration

reaches 30%, which is used as a rule of thumb for secure wind

power penetration in autonomous systems. This occurs during

the winter season, when wind power is significant and load is

not as large as in the summer season. The system inertia underthese conditions is restricted, as not so many conventional

units participate in the production. This scenario is consideredto be the worst case from the dynamic security perspective for

the power system.

B. Thermal power plant models

The conventional generating capacity includes diesel, gas

and steam plants with different ratings and control attributes.

Each thermal plant contains several control blocks, which are

essential for power system dynamic simulations, i.e. voltage

controller, primary controller (governor), primary mover unit

and the synchronous generator, [2].

The following three different models, already existing as

built-in standard models in Power Factory [3], are used for the

governors of the Rhodes power system: GAST2A model for

the gas turbines, DEGOV1 model for the diesel generators andIEEEG1 general model for the steam plants. A detailed

description of the GAST2A built-in model in PSS/E for thegovernor used in the gas plant is described in [4], while details

on the corresponding standard IEEEG1 model for the governor

in the steam plant can be found in [3]. The parameters of these

models, validated both in Matlab and PSS/E software

packages, are presented in [4]. For the Automatic Voltage

Regulators (AVR), the built-in SEXS model of DIgSILENT is

used with adjusted parameters for each unit.

The electrical loads of the systems include typically various

kinds of electrical devices. The approach for the dynamic

modeling of the loads connected to Medium Voltage (MV)

feeders in this study is to assume constant impedance of theloads during dynamic simulations [5].

C. Wind farm models

In the reference year 2012, five wind farms will be

connected online in Rhodes power system. Three wind turbine

technologies are considered, namely Doubly Fed Induction

Generator (DFIG), Permanent Magnet Synchronous Generator

(PMSG) and Active Stall Induction Generator (ASIG) based

wind turbines. Table II below presents the sizes and type of

wind turbine technology used in these wind farms.

TABLE IIWIND FARMS IN RHODES POWER SYSTEM IN 2012.

Wind Turbine

Technology

Installed

Capacity (MW)

Wind Farm A1 DFIG 11.05

Wind Farm A2 DFIG 5.95

Wind Farm B1 PMSG 18

Wind Farm B2 PMSG 3

Wind Farm C ASIG 11.7

An aggregated method for modeling the wind farms isused, [6]-[7]. Such modeling approach is commonly used for

power system studies, as it reduces substantially both the

complexity of the system and the computation time, without

compromising the accuracy of the simulation results.

Models for all these different wind turbine technologies are

implemented in DIgSILENT, including the main components

of each wind turbine configuration:

Drive train and aerodynamics Pitch angle control system Control system Protection system

The DFIG wind turbine configuration stands nowadays as

the mainstream configuration for large wind turbines, [8]. Themain electrical components as well as the mechanical parts

and the controllers were considered in the model.The model

used in this study for the wind farms with DFIG wind turbines

is described in details in [9]. The DFIG system is essentially a

wound rotor induction generator with slip rings, with the stator

directly connected to the grid and with the rotor interfacedthrough a back-to-back partial-scale power converter [10].

Similar to the DFIG wind turbine configuration, the PMSG

wind turbine consists both of a wind turbine mechanical level

(i.e. aerodynamics, gearless drive-train and pitch angle

control) and an electrical level (i.e. multi-pole PMSG with a

full-scale frequency converter and its control). The

synchronous generator is connected to the grid through a full-scale frequency converter system, which controls the speed of

the generator and the power flow to the grid. The full-scale

frequency converter system consists of a back-to-back voltage

source converter (generator-side converter and the grid-sideconverter connected through a DC link), controlled by IGBT

switches, [11]-[12].

The sub-models for aerodynamics, mechanical components

and the squirrel cage induction generator for the ASIG wind

turbine were also included in the platform developed during

this study, [13]. The drive train is represented by a two-mass

model. The turbines power is controlled directly by the pitch

controller through the pitch angle defining two operating

modes for this wind turbine system power optimization, forlow wind speeds and power limitation, for high wind speeds.

III. ANCILLARY SERVICES PROVIDED BY WINDFARMS

In this section, some basic definitions on ancillary services

provided to power systems by modern wind turbine

technology are given, to introduce the main issues with

emphasis on isolated power systems. The technical

requirements set by the networks operators nowadays include

various aspects, such as fault ride-through capability of wind

turbines during faults, voltage-reactive power control and


3/8

overall control of the wind farms as conventional powerplants. The wind farms in this study are equipped with

advanced control features, which enable them to support the

grid during transient phenomena.

A. Fault ride through and voltage control

When the voltage drops at a grid fault instant, the wind

turbines are asked nowadays to stay online and provide withreactive power, supporting the voltage. Depending on the wind

turbine configuration, different control methods have to beimplemented.

The specific converter arrangement in the DFIG

configuration requires advanced protection system through the

so called crowbar, because of the high inrush stator and

rotor currents during grid faults. Ensuring that the grid-side

converter is not blocked at a grid fault but continues its

operation as a STATCOM, the controllability is improved

providing the DFIG with grid voltage support in uninterrupted

operation during the transient instant, [14]. Both wind farms in

the power system studied in this article equipped with DFIGturbines are designed including these control attributes. The

voltage control is also activated in the rotor side converter

control system, thus the wind farm does not maintain powerfactor of unity as usual, but regulates the reactive power in

order to support the voltage after the fault is cleared.

Respectively, PMSG wind turbines are connected via a full-

scale frequency converter to the grid and therefore in principle

can easily accomplish fault ride-through and support the grid

during faults. The generator side converter maintains the DC-

link voltage around the nominal value. The grid side

converter, which is subject to the grid fault, cannot deliver theexpected active power during the low voltage at the grid fault

instant. The chopper introduced in the DC-link ensures the

consumption of the surplus energy in the chopper resistance,

improving the response of the system during faults at the grid,

[15]-[16]. Voltage control is also included in the PMSG windfarms, providing with the additional capability to control the

reactive power exchanged with the grid after the fault

clearance.

For the ASIG wind farm, the method of power reduction at

the instant of a fault at the grid was implemented, [17]. The

normal power controller of the wind turbine is substituted byan auxiliary control system during the fault, which ramps

down the mechanical power of the rotor. The system goes

back to normal operation, as soon as the voltage at the Point of

Common Coupling (PCC) of the wind farm is re-established.

The significant result in the power system dynamic

security, when the wind farms are able to withstand low

voltages, is demonstrated through simulation results for all

three load scenarios developed.

B. Auxiliary frequency control

In case of sudden generation loss or large load connection,

the frequency of the frequency starts to drop. The two main

system functions that ensure return of an unbalanced system to

nominal frequency are the Primary Control, which occurs

during the first 30-40 sec after the event and prevents further

frequency drop, and the Secondary Control, which takes place

several minutes after the event and re-establishes nominal

frequency by adjusting the production in the conventional

units of the power system. In small power systems, likeRhodes, the secondary control is operated manual request by

the system operator.

The total angular momentum of the system, thus the sum of

the inertias of the rotating masses in the system, define the rate

at which the frequency changes. Fast frequency drops lead to

large load shedding, which is ordered by the Rate of Change

of Frequency (ROCOF) protection system, in order to avoid

frequency instability. In non interconnected systems, like the

one studied in this study, system inertia is reduced and

frequency deviations are often and quite severe. This state

becomes even more critical nowadays, when large wind

turbines substitute conventional generators in modern powersystems under increasing wind power penetration conditions.

The effect on the power system inertia and the availability of

inertia response from wind turbines have become key issues

for the secure integration of wind energy into the electrical

grids. Although sufficient spinning reserve is ensured to

overcome any frequency problems, increasing wind power

penetration is challenging the system security. Variable speed

wind turbines, like DFIG and PMSG, do not contribute to

system inertia, while fixed speed wind turbines, like ASIG,

have an inherent response when the system frequencydeviates, [18]. During the last years, auxiliary frequencysupport provided by modern wind turbines is considered to be

urgently needed. This means, that whenever the frequency in

the system deviates, due to imbalance between power

production and consumption, the wind turbines are asked to

contribute regulating their power accordingly.

In this paper, two different frequency control methods,

which were applied in the DFIG wind turbine models used in

the Rhodes power system model, are described:

a) Inertia Controlb) Droop Control

Results from frequency events in the Rhodes power system,

when these control methods are used in the wind farmsequipped with DFIG wind turbines, are given in Section IV.

The frequency controller is illustrated in Fig. 1.

Pgrid

Pgridref

Cascade

Power controller

rP ef

int

Maximum Power

Tracking Po

PI PIcascade

+

+

gen!

Frequency controller

Signals to Rotor Side

Converter

Pnormalref

Pauxref

Aux

inputsignal

Fig. 1. Frequency control scheme for DFIG wind turbines.

In Inertia control is used, the inertial response of the DFIG

is restored through an additional loop in the power reference

block providing the active power reference signal to the Rotor

Side Converter. The DFIG wind turbine adjusts is ordered to

adjust its power output when subjected to frequency

deviations, based on the measurement of the rate of change of

the frequency. This defines the additional power reference


4/8

signal, which is added to the normal power reference providedby the Maximum Power Tracking Controller, [19].

In Droop control, the auxiliary input signal that defines the

adjustment in the power production of the wind turbine is the

actual error between the measured frequency of the system

and the nominal frequency. This control method is based on

the primary frequency control applied to conventional

generators. Discussion on the results from each control

method proposed here is made in the next session.

IV. RESULTSIn this section, results from the Rhodes power system are

presented. The uninterrupted operation of all three wind

turbine technologies available in the power system studied

during faults and the frequency control capability of DFIG

wind turbines is investigated for the load scenario defined

above. The loss of the largest infeed in the system and a

typical short circuit are simulated and the frequency response

is under focus.

A. System response during faults

As it is mostly the case in many island grids, like Rhodes,

wind parks are allowed to trip during severe faults to avoiddestruction of the power electronics and mechanical stress.

The task to support the grid during faults, therefore, is

assigned to the conventional generators. When wind farms are

not able to withstand low voltages at the Point of Common

coupling, the frequency is decreased even more due to largewind power production loss, which is a result of the initial

fault. On the other hand, when wind farms are equipped with

fault ride through capability, they stay online and continue to

produce active power immediately after the fault clearance. As

it is the case in the control implemented in this paper for the

wind turbines, they can also contribute with reactive power to

support the voltage. This of course stands for the variable

speed wind turbines DFIG and PMSG wind turbines whereas in the case of ASIG, power reduction can avoid

extreme rotor acceleration, which may lead to instability and

disconnection of this type of wind turbines. In the following

graphs a three phase short circuit is simulated in the mediumvoltage network of Rhodes. Fig. 2-4, illustrate the response of

the DFIG wind turbines in WFA1 during the fault, with or

without damping controller. It should be noted that the active

and reactive power correspond to the whole wind farm

(WFA1), while the generator speed corresponds to one wind

turbine.

0 2 4 6 8 10 12

0

2

4

6

8

10

Time (sec)

GridPower(MW)

Fig. 2. Active power during the fault for wind farm with DFIG wind

turbines solid line, with damping controller, dashed line, without dampingcontroller.

The sudden drop of the voltage leads to drop in the statorand rotor flux, which result in decrease of the active power

delivered by the wind turbine. The electromagnetic torque is

also dropping, and as the drive train acting as torsion spring

gets untwisted during fault, the mechanical torque drops too.

However, the mechanical torque drops slower than the

electromagnetic torque and therefore the generator starts to

accelerate (Fig. 4). The high inrush currents, which follow the

voltage drop, trip the crowbar protection system. The rotor

side converter (RSC) is blocked and the generator behaves as a

conventional squirrel cage induction generator.

The effect of the damping controller, which acts directly on

the active power reference signal, is very crucial, [10]. Thiscontroller damps actively the torsional excitations in the drive

train system following the grid fault. When no damping

controller is used, the oscillations in the generator speed

remain undamped and could possibly lead to disconnection of

the wind turbine by the protection system. The comparison

reveals therefore a positive effect of the damping controller on

the response of the wind turbine.

0 2 4 6 8 10 12 0.95

1

1.05

1.1

Time (sec)

GeneratorSpeed

(pu

)

Fig. 3. Generator speed during the fault for DFIG wind turbine solid

line, with damping controller, dashed line, without damping controller.

During the fault, the voltage control of the grid side

converter (GSC) demands the wind turbine to deliver reactivepower to support the voltage at the PCC of the wind farm. The

wind farm manages to provide with large amount of reactive

power (Fig. 4). When the fault is cleared, the generator starts

to absorb reactive power, as it is still behaving as squirrel cage

induction generator as long as the crowbar is triggered. The

RSC is still blocked and this delays the quick restoration of the

voltage until the tripping of the crowbar protection system.

0 2 4 6 8 10 12 -4

-2

0

2

4

6

8

Time (sec

ReactiveP

ower(MVAR)

Fig. 4. Reactive power during the fault for wind farm with DFIG windturbines.

Respectively, Fig. 5-7 present the fault ride-throughcapability of the PMSG wind turbines in wind farm WFB2,

with and without chopper, [16].


5/8

0 2 4 6 8 -1

0

1

2

3

4

Time (sec)

GridPower(MW)

Fig. 5. Active power during the fault for wind farm with PMSG wind

turbines solid line, with chopper, dashed line, without chopper.

During the fault, the grid-side converter cannot deliver to

the grid the whole active power generated by the generator,

due to low voltage at the Point of Common Coupling. Due to

the imbalance between aerodynamic and electrical powerduring the fault, the generator starts to accelerate. Meanwhile,

the drive train gets untwisted and oscillates.

A PMSG wind turbine can ride through a grid fault without

any additional measure, i.e. a chopper. However, the

oscillations visible in the generator speed (see Fig. 6), aresignificantly reduced when a chopper is used. Besides the

faster damping of the oscillations, the chopper helps also to

decrease the rotor accelerations following the fault,

minimizing the mechanical stress of the wind turbine.

0 2 4 6 8 0.93

0.94

0.95

0.96

0.97

0.98

0.99

1

Time (sec)

GeneratorSpeed

(pu)

Fig. 6. Generator speed during the fault for a PMSG wind turbine solid

line, with chopper, dashed line, without chopper.

The wind turbine controls the voltage at the PCC and

delivers reactive power to support the voltage during the drop.

Fig. 7, shows the increase in the reactive power produced by

the wind farm equipped with PMSG wind turbines. As long as

the voltage is low, the wind farm supports the voltage, untilthe fault clearance.

0 2 4 6 8 -1

0

1

2

3

4

Time (sec)

ReactivePower(MW

)

Fig. 7. Reactive power during the fault for a wind farm equipped withPMSG wind turbines solid line, with chopper, dashed line, without chopper.

The wind farm equipped with ASIG wind turbines has towithstand the low voltage during the fault, and ensure

uninterrupted operation. Fig. 8-10 illustrate the behavior of

ASWG wind turbine during the fault.

After the clearance of the fault, the active power may still

be reduced for a few seconds. As explained in Section III,

during the voltage drop, the active power delivered by the

generator has to be reduced in order to make the turbine able

to ride through the fault.

0 5 10 15 -2

0

2

4

6

8

10

12

Time (sec)

GridPower(MW)

Fig. 8. Active power during the fault for a wind farm equipped with ASIGwind turbines

Although the inrush currents are high during the fault, the

thermal constants of the induction generator are also quite

high and the need for protection is reduced compared to the

sensitive power electronics of variable speed wind turbines.

Fig. 9 shows the rotor speed of the ASIG wind turbine,

which reflects the power system frequency behavior during thefault. When the fault occurs, the speed is initially increased

due to the acceleration of the conventional generators and

afterwards drops below nominal value.

0 5 10 15 0.97

0.98

0.99

1

1.01

1.02

1.03

1.04

Time (sec)

GeneratorSpeed(p

u)

Fig. 9. Generator speed during the fault for a ASIG wind turbine.

The reduction of the wind turbine mechanical power is

applied to assess the fault ride-through capability of the active

stall wind farms.

During the grid fault the ASIG wind turbine absorbsreactive power. After the initial peak in the reactive power,

shown in Fig. 10, the wind farm absorbs reactive power.

Reactive power is measured in the PCC, and includes thepower delivered to the grid by the capacitor banks installed at

the wind farm bus to reduce the negative effect on reactive

power-voltage control of the wind farm during severe faults.


6/8

0 5 10 15

0

5

10

15

Time (sec

Re

activePower(MVAR)

Fig. 10. Reactive power during the fault for a wind farm equipped with

ASIG wind turbines.

Fig. 11 shows the system frequency for the same fault when

FRT is or not available. When wind farms are not able to

withstand the low voltage during the fault, trip for protection

reasons (solid line in Fig. 11). In the load scenario under studyin this article, i.e. Maximum Wind Power Penetration

scenario, the frequency of the system drops down to 47.6 Hz.

The initial frequency drop is further increased due to the loss

of large amount of wind power in just few milliseconds after

the voltage drop. The under-frequency protection system actson the loads connected to the medium voltage substations and

65.5 % of the load, i.e. 54 MW, is immediately disconnected.

On the contrary, when the wind farms are able to remain

connected to the grid during fault, thus the uninterrupted

operation is fully ensured, the frequency minimum is 49.7 Hz

and the load shedding is totally avoided. The contribution of

the fault-ride through capability in the frequency stability butalso in the load shedding of power systems, like Rhodes,

appears to make FRT crucial for the secure operation of the

system.

0 2 4 6 8 10 47

48

49

50

51

52

Time (sec)

Systemf

requency(Hz)

Fig. 11. System frequency during the fault solid line, when wind farms

are tripping due to low voltage, dashed line, when FRT is available and windfarms stay online.

B. System frequency response for largest unit loss

The total wind power production is 28.2 MW (34 %) intotal 83 MW of demand. The system inertia is decreased due

to less conventional units producing, making the frequency

control task in the system more complex. The largest

conventional unit in the system produces 21 MW before the

protection system acts to take it out of operation this means

production loss equal to 25 % of the total demand. The fault is

severe and the power system stability is checked for all thefrequency control schemes designed in this study.

In Fig. 12 and 13 the system frequency for all the differentfrequency control schemes implemented in the wind farms A1

and A2 (see Table II) is shown.

0 2 4 6 8 10 1248

48.5

49

49.5

50

50.5

Time (sec)

SystemFr

equency(Hz)

(a)

(c)

(b)

Fig. 12. System frequency for largest unit loss when frequency control is

applied by DFIGs (a) No auxiliary control, (b) Droop control, (c) Inertia

control.

In case (a), when the wind farms do not have auxiliary

frequency control, the frequency drops below 48.5 Hz which

is the upper zone of the under-frequency protection relay

settings acting on the loads. This drop leads to disconnection

of 15.1 MW of load 18 % of the total demand. This loadshedding is not considered accepted in terms of dynamic

security terms, [1]. However, in all the other cases, where the

frequency control is activated in the DFIG wind farms, the

load shedding is avoided totally. The maximum frequency

drop appears in case (c), where the inertia controller is used.

The optimum frequency drop in terms of minimum frequency

is achieved in case (b), when Droop control is implemented.

The effect of auxiliary frequency control on the maximum

rate of change of frequency is very crucial. As illustrated in

Fig. 12, where the initial drop of the frequency for all cases is

zoomed in, this rate is very high. The inertia of the system inthis case is low because the number of the conventional

generators connected to the system is reduced. The rate of

change of frequency is close to 2.8 Hz/sec (in absolute value)

in case (a). Inertia control manages to reduce the rate to 1.8

Hz/sec, which is the lowest rate (see Table IV).

1 1.2 1.4 1.6 1.849.2

49.4

49.6

49.8

50

50.2

Time (sec)

SystemF

requency(Hz)

(a)

(b)

(c)

Fig. 13. System frequency for largest unit loss when frequency control isapplied by DFIGs Zoom in the first seconds after the event (a) No auxiliary

control, (b) Droop control, (c) Inertia control.

TABLE IVRESULTS FOR SCENC LOSS OF LARGEST INFEED

Frequency

Control

Scheme

Minimum

Frequency

(Hz)

Maximum

Rate of

change of

frequency

Load

Shedding

(MW)


7/8

(Hz/sec)

(a)

No

auxiliarycontrol

48.28 -5 15.1

(b)Droop

control48.69 -5 0

(c)Inertia

control48.50 -3.8 0

V. CONCLUSIONSThis article demonstrated power control issues based on a

full detailed model of an autonomous power system with high

wind power penetration. Three different wind turbine

technologies were modeled and the dynamic security of the

power system during faults was investigated. Based onoperational data, the scenario studied was the maximum wind

power penetration. The protection system in the power system

was also included providing more accurate results regarding

the load shedding. The effect of uninterrupted operation of

wind farms during low voltages was demonstrated. As wind

power penetration is increasing in modern power systems, the

wind turbines have to contribute to the frequency stability ofthe system, acting similar to conventional power plants. In this

article, two different frequency control schemes were

investigated to enhance the primary frequency support of

DFIG wind turbines. Simulation ended up to the followingconclusions:

1. System inertia in non interconnected power systems,like Rhodes, is significantly decreased, especially

under high wind power penetration conditions. The

load shedding following events at these systems is

often and severe. Due to lack of specific grid code

defining these aspects, it is often the case that still

wind farms are allowed to trip when low voltage is

detected at the Point of Common Coupling. This

means that, whenever there is a short circuit at the

grid, the system encounters probable loss of

production. This leads to further stress in thedynamic security of the system in terms of frequency

stability.

2. When wind farms are able to withstand low voltagesand remain on line, the frequency stabilization is

ensured through the primary control service of the

conventional units. The load shedding following

frequency drops is decreased in this case and the

wind farms can continue their operation contributing

to the system security.

3. DFIG and PMSG wind turbines can contribute tosystem stability by adjusting their reactive powerproduction during low voltage instances. Modern

power electronic technology can provide with

sophisticated ride through capability, which allows

wind farms to operate as a conventional power plant

in terms of voltage control.

4. DFIG wind turbines, when equipped with auxiliaryfrequency control, can provide with valuable active

power response during the first seconds following the

frequency event. This control makes it possible for

the wind turbines to participate in the primary control

service of the system, stabilizing frequency afterdeviations due to i.e. loss of generation.

5. Inertia and Droop control were demonstrated in thisarticle and were proved to be efficient in improving

the system frequency response. When Inertia control

is used, the rate of change of frequency is

significantly reduced, while Droop control seems to

benefit more the power system when looking at the

minimum frequency after the event.

6. Although wind power production is higher than 30 %of the total demand, the auxiliary frequency control

implemented in two wind farms in the Rhodes power

system, manages to avoid load shedding totally.Therefore, the rule of thumb of 30 % penetration,

which is often used in autonomous power systems,

can be further expanded as long as auxiliary

frequency control is provided by wind farms.

7. The benefits of the primary frequency support frommodern wind turbines increase as the number of the

turbines with this capability rises. This means that, if

all new wind farms installed in autonomous power

systems are equipped with primary frequency control

capability, the frequency stability can be ensuredeven for penetration levels that today are hard toconsider.

8. The review of the frequency and voltage protectionsystem settings can be done, as long as the stability

of the system is ensured. In many cases, the

protection settings are quite sensitive and large

amounts of load are cut off. The review of the

protection system in modern power systems has to

follow the progress made in the wind farms

capability to support the grid during disturbances.

VI. REFERENCES[1] I. D. Margaris, J. C. Mantzaris, M. E. Karystianos, A. I. Tsouchnikas, C.

D. Vournas, N. D. Hatziargyriou and I. C. Vitellas, Methods for

evaluating penetration levels of wind generation in autonomoussystems, accepted for presentation in IEEE PowerTech Conf.,

Bucharest, June 2009.[2] P. Kundur, Power System Stability and Control, Ed. McGraw-Hill,

1994.

[3] DIgSILENT GmbH. DIgSILENT technical documentationPowerFactory, 2006.

[4] J. Mantzaris, M. Karystianos, C. Vournas, Comparison of Gas Turbineand Combined Cycle Models for System Stability Studies, presented atthe 6th Mediterranean. Conf. MedPower, Thessaloniki, Greece, 2008.

[5] T. V. Cutsem, C. Vournas, Voltage Stability of Electric Power Systems,Ed. Springer, 1998.

[6] V. Akhmatov, Analysis of dynamic behavior of electric power systemswith large amount of wind power, PhD thesis, 2003, rsted DTU.

[7] M. Poeller, S. Achilles, Aggregated wind park models for analyzingpower system dynamics, presented in Fourth international workshop onlarge-scale integration of wind power and transmission networks,

October 2003, Billund, Denmark, DIgSILENT, 10pp.

[8] A.D.Hansen, L.H.Hansen, Market penetration of different wind turbineconcepts over the years, EWEC 2007, Milano,6pp.

[9] A.D. Hansen, P. Srensen, F. Iov, F. Blaabjerg,, Centralised powercontrol of wind farm with doubly-fed induction generators. RenewableEnergy, vol 31 (2006), 935-951.

[10] A.D. Hansen, G. Michalke, Fault ride-through capability of DFIGwind turbines, Renewable Energy, vol 32 (2007), pp 1594-1610.

[11] M.H. Hansen, A.D. Hansen, T.J. Larsen, S. ye, P. Srensen, Controldesign for a pitch-regulated variable speed wind turbine, Ris-R-1500(EN), 2005, 84pp.


8/8

[12] A.D. Hansen, G. Michalke, Modelling and control of variable speedmulti-pole PMSG wind turbine, Wind Energy, 2008, Vol.11(5), pp 537-554.

[13] C. Jauch, A. D. Hansen, P. Srensen and F. Blaabjerg, SimulationModel of an Active-stall Fixed-speed Wind Turbine Controller, WindEngineering, Vol. 28, no.2, pp. 177-195, 2004.

[14] A.D. Hansen, G. Michalke, P. Srensen, T. Lund, F. Iov, Co-ordinatedvoltage control of DFIG wind turbines in uninterrupted operation duringgrid faults, Wind Energy, Vol. 10, No. 1, 2007, pp.51-68.

[15] A.D. Hansen, G. Michalke, Modelling and control of variable speedmulti-pole PMSG wind turbine, Wind Energy, 2008, Vol. 11(5), pp

537-554.[16] A.D. Hansen, G. Michalke, Multi-pole PMSG wind turbines grid

support capability in uninterrupted operation during grid faults,

submitted to IET Renewable Power Generation, June 2008, accepted in

Nov. 2008.[17] Akhmatov V., Knudsen H., Nielsen A.H., Pedersen J.K., Poulsen N.K.,

Modeling and transient stability of large wind farms, Electrical Power

and Energy Systems 25 (2003), 123-144.[18] J. Ekanayake, and N. Jenkins, Comparison of the response of doubly

fed and fixed-speed induction generator wind turbines to changes in

network frequency, IEEE Trans. Energy Conversion, 2004, 19, (4), pp.

800802.[19] L. Holdsworth , J.B. Ekanayake and N. Jenkins, Power system

frequency response from fixed speed and doubly fed inductiongenerator-based wind turbines, Wind Energy 2004; 7 : 21-35.DOI:10.1002/we.105.

[20] G. Ramtharan, J.B. Ekanayake and N. Jenkins, Frequency support fromdoubly fed induction generator wind turbines, IET Renew. Power

Gener., 2007, 1, (1), pp. 39.

VII. BIOGRAPHIESIoannis Margarisreceived Dipl. Eng. Degree and Master degree in Electrical

and Computer Engineering, Power Systems, from the National Technical

University of Athens, in 2006. He is currently pursuing his PhD thesis atNational Technical University of Athens, School of Electrical and Computer

Engineering. His research interests are dynamic modeling and control of windturbines and wind farms, power electronics, FRT, power system integration ofwind power. During 2008-2009 he was visitor PhD student for a six month

period at Ris National Laboratory in Roskilde, Denmark.Anca D. Hansen received her Ph.D. in modelling and control engineering

from Denmark Technical University in 1997. Since 1998 she has been

employed at Ris-DTU National Laboratory in the Wind Energy Department first as Post Doc., scientist and afterwards as senior scientist. Her workingfield and research interests are on the topics of dynamic modelling and control

of wind turbines, as well as dynamic modelling and control of wind farms andon wind farm grid interaction. She is author or co-author of more than 80

journal/conference papers, several research reports in her research fields.

Poul Srensen (M04) was born in 1958. He received M.Sc. in electrical

engineering from the Technical University of Denmark in 1987. Since 1987he has been employed at Ris National Laboratory in Roskilde, presently as a

Senior Scientist. His main technical interest is integration of wind power intopower systems, involving a variety of technical disciplines including powersystem control and stability, dynamic modeling and control of wind turbines

and wind farms, and wind fluctuation statistics.Nikos D. Hatziargyriou was born in Athens, Greece. He received the

Diploma in Electrical and Mechanical Engineering from NTUA and MSc and

PhD degrees from UMIST, Manchester, UK. He is currently executive Vice-

Chair and Deputy CEO of the Public Power Corporation of Greece and part-

time professor at the Power Division of the Electrical and ComputerEngineering Department of NTUA. His research interests include Dispersedand Renewable Generation, Dynamic Security Assessment, and application ofArtificial Intelligence Techniques to power systems. He is fellow IEEE

member, chair of the PSDP Committee, convener of CIGRE SCC6 andmember of the Technical Chamber of Greece.

Documents

Bremen Workshop 2009