ewis - european wind integration study - contract number:...

Contract number: TREN/07/FP6EN/S07.70123/038509

European Transmission System Operators

European Wind Integration Study (EWIS) Towards a Successful Integration of Wind

Power into European Electricity Grids

EWIS Wind Turbine Model Validation Report

Revision [Final Version – Including EWEA meeting outcome: 15.12.2008 – WG3]

[Version 3.0 - Including EWEA Feed Back: 04.12.2008 – WG 3] [Draft Version 2.0: 02.06.2008 – WG3]

[Draft Version 1.0: 15.04.2008 – WG3]

Content EWEA Feedback on the ‘EWIS-Wind Turbine Model Validation Report’ .....................................4 1. Validation of Wind Turbine models ....................................................................................7 1.1. Introduction .............................................................................................................................7 1.2. Current wind power supply and harmonised models for wind turbine....................................8 1.3. Test System............................................................................................................................9 1.3.1. Generator Types ..............................................................................................................11 1.3.2. Type of Investigations ......................................................................................................11 1.3.3. Grid Connection Requirements........................................................................................12 1.4. Visualisation of the results....................................................................................................13 1.5. Results..................................................................................................................................13 1.6. System parameters ..............................................................................................................14 1.7. Exemplary Results................................................................................................................18 2. Wind turbine models based on grid code requirements in Germany...........................24 2.1. Wind Turbine Protection .......................................................................................................26 2.2. Wind farm based on double fed induction generators..........................................................27 2.3. Wind farm based on squirrel cage induction generators......................................................34 2.4. Wind farm based on full size inverter model ........................................................................35 3. Wind turbine models based on grid code requirements in Spain.................................36 3.1. Objectives .............................................................................................................................36 3.2. Grids used for the test ..........................................................................................................37 3.3. REE grid used to test wind turbines .....................................................................................37 3.4. Wind turbine models.............................................................................................................40 3.5. Doubly fed induction generator model..................................................................................45 3.6. Squirrel cage induction generator model..............................................................................57 4. Grid code requirements and wind turbine models used in Czech Republic................81 4.1. Grid code requirements ........................................................................................................81 4.2. Squirrel cage induction generator model..............................................................................82 4.3. Double Fed Induction Generator model ...............................................................................85 5. Validation of Wind Turbine Models - SCIG - in Denmark................................................89 5.1. Introduction ...........................................................................................................................89 5.2. Test System..........................................................................................................................89 5.3. Wind Turbine Model with SCIG ............................................................................................91 5.3.1. Electric Parameters ..........................................................................................................91 5.3.2. Mechanical System Representation.................................................................................92 5.3.3. Compensation of Reactive Power....................................................................................94 5.3.4. Protection System ............................................................................................................95 5.4. Analysis of the Test System with SCIG Wind Turbine .........................................................96 5.4.1. Three Phase short circuit, UPCC = 0.2 p.u. .......................................................................96 5.4.2. Three Phase short circuit, UPCC = 0.8 p.u. .......................................................................99 5.5. Influence of Two – Mass Model..........................................................................................102 5.6. Conclusions and Recommendations..................................................................................105 5.6.1. Recommendations for Simulation Settings ....................................................................106 5.7. Appendix - SCIG, Denmark ................................................................................................107 5.7.1. Simulation Results with PowerFactory – Load Flow......................................................107 5.7.2. Simulation Results with Two-Mass Model in PowerFactory ..........................................109 6. Validation of Wind Turbine Models - DFIG - in Denmark..............................................112 6.1. Introduction .........................................................................................................................112 6.2. Test System........................................................................................................................112 6.3. Wind Turbine Model with DFIG ..........................................................................................114

2

6.3.1. General Structure of the Model ......................................................................................114 6.3.2. Electric Parameters ........................................................................................................114 6.3.3. Mechanical System Representation...............................................................................116 6.3.4. Control System of the Wind Turbine ..............................................................................117 6.3.4.1 Generator and Rotor Side Frequency Converter ...........................................................117 6.3.4.2 Grid-Side Frequency Converter .....................................................................................121 6.3.5. Protection System - Crowbar .........................................................................................121 6.3.4.3 General Remarks ...........................................................................................................121 6.3.4.4 Influence of Crowbar Resistance Size on the Machine Behaviour ................................124 6.4. Analysis of the Test System with DFIG Wind Turbine........................................................124 6.5. Conclusions and Recommendations..................................................................................125 6.5.1. Recommendations for Simulation Settings ....................................................................126 6.6. Appendix DFIG - Denmark .................................................................................................127 6.6.1. Simulation Results with PowerFactory – Load Flow......................................................127 6.6.2. Block Diagrams of the Enhanced Wind Turbine Model .................................................128 6.6.3. Simulation Results for Dynamic Analysis.......................................................................131 6.6.4. Influence of Crowbar Resistance on Behavior of the Machine ......................................135 7. References ........................................................................................................................137

3

EWEA Feedback on the ‘EWIS-Wind Turbine Model Validation Report’

⇒ EWIS received EWEA Feed Back to

Wind Turbine Model Validation Report

⇒ Provision of Real Measurements

⇒ WT-Model Validation discussed with

EWEA adds more value to EWIS

investigations.

• EWEA - Time duration for the grid disturbance (150 ms) is not sufficient to

demonstrate the reactive current support and active power production during the

voltage dip-

EWIS have used even longer time duration like 500 ms and the WT models fulfill

the requirements like reactive current support and active power production during

voltage dip. (See Figure 0-1)

• EWEA - EWIS WT models crowbar operation time:100ms-

With the available experience of the TSOs the average timing for crowbar firing is

100 ms, still EWIS WT models are capable of using 40ms time for the crowbar

firing and shows satisfactory performance of the reactive power infeed during fault

duration. (See Figure 0-1)

4

Fault incident

Voltage support

Post fault voltage supportCrow BarFiring

EWEA – Feed BackABB – Presentation

- higher short circuit ratio- faster reactive

current provision- better voltage recovery after

fault clearance- post fault voltage support

Green Curve: Ir (measurement)Red Curve: EWIS DFIG Ir

Voltage drop to 20%Fault duration: 500 ms

Crow bar firing (40ms)

nom. voltage

TransmissionCode (Germany) requirements

Figure 0-1: EWIS Results compared with the Real measurement provided by EWEA

• EWEA – Capability of EWIS WT models for using 3rd order or 5th order transient

model for DFIG

The 5th order transient model is useful for local behaviour (regional level). It will

have no influence on global system behavior (EWIS - Global level project).

EWIS WT models are capable of using both 3rd order and 5th order transient

models. It was agreed for stability studies the 3rd order model is appropriate.

EWIS models presented in the EWEA-meeting shows satisfactory response for

each technology of WT – evidence by EWEA feed back after the EWIS – EWEA

wind turbine model feed back meeting on 12th of December 2008, Renewable

Energy House in Brussels.

5

Conclusion

• EWIS is thankful to EWEA for their feedback on the EWIS-Wind Turbine Model

Validation Report.

• EWIS WT models are updated to fulfill additional requirements asked by EWEA

such as crowbar firing time, fault duration and the 3rd and 5th order transient

models.

• EWIS WT models are simplified for the Extra High Voltage level but detailed

enough to show the desired WT behaviour on the Extra High Voltage level for the

EWIS study.

• The simulation results shown in the presentation confirms that EWIS WT models

fully satisfy the criteria ‘simplification and accuracy’ required for the Global level

investigation such as EWIS.

• All meeting participants agreed that wind turbines are well represented by

the presented EWIS standard wind turbine model behaviour, all the open questions

were discussed and solved.

• The meeting participants agreed that EWIS can start the investigation based on the

represented wind turbine behaviour presented during the EWIS-EWEA wind

turbine model feed back meeting on 12th of December 2008, Renewable Energy

House, Brussels.

6

1. Validation of Wind Turbine models

1.1. Introduction

The European electricity network is the route for the efficient transport of wind power from

turbines to consumers. It also provides the means for managing the variability of wind

generation by harnessing diversity and backup energy sources. To meet Europe’s

renewable energy targets, changes are needed in order to most efficiently integrate wind

power. In order to better analyse the future development of RES generation in Europe,

thorough examination are performed on the European level to find cost effective solutions

to maintain system stability and security, providing impact on the electric infrastructure as

a whole and the future tasks for infrastructure development.

For the purpose of this study wind turbine equivalent models are used based on publicly

available data, endorsed by wind turbine manufacturers, validation and information within

simulation tool packages. Some of these sources are listed in the List of References. The

report shows an overview over the model representation in all relevant software packages

for power system simulation as MODES, Netomac, PowerFactory, PSD, PSS-E, etc.

The complexity of the models used for the actual study has on the one hand to be exact

enough with respect to the dynamic behaviour of the turbine, fulfilling all grid code

requirements in Europe but on the other hand be suitable for large-scale grid studies.

Therefore simplified wind turbine models are used in EWIS that perform the typical

response of each technology (See 1.3).

Measurements for wind turbine models in form of certificates and bilateral analysis

together with manufacturers are available, but related to specific wind turbine architecture

and almost confidential. The described standard models are derived from specific WT-

models given by manufacturers with respect for power system analysis. For the

completion of the report the cooperation to TradeWind is an opportunity to implement real

measurements for the general types of wind turbine models.

7

1.2. Current wind power supply and harmonised models for wind turbine

For the regional and pan-European investigations it is necessary to find out the most

convenient and realistic distributions for wind power supply. Steady state analysis and the

grid expansion/re-enforcement measures will depend on this subject.

As wind power has the highest growth of all RES, transient analysis is focused on the

impacts of introducing wind power into the power system. Therefore special wind turbine

models are used, e.g. the following aspects are taken into consideration:

• Wind turbines (WT) equivalents for grid connections in distribution and high voltage

systems.

• Regional distribution of different WT-technology related to the installed capacity

• Different operation modes of WT-models in accordance to grid requirements

• Representation and distribution of old installations which can not or through

retrofitting fulfil new grid code requirements (e.g. fault-ride-through behaviour)

As the use of wind power plants increases worldwide, it is important to understand the

effect these power sources have on the operation of the grid. Transmission System

Operators (TSOs) focus their attention on the system interaction of the various wind

turbine types and the differences compared to conventional power plants as power plants

have to provide system services to enable the steady operation of a grid. There are

different wind power technologies. The three most commonly used technologies will be

taken into consideration for this study:

o The squirrel cage induction generator (with or without fast voltage support)

o The doubly fed induction generator and

o The full size inverter model.

8

The required behaviour of different WT-technologies is considered and implemented in the

dynamic equivalents:

• Reactive power generation and voltage control capability

• Behaviour in case of system faults

• Frequency control and others

• Damping of inter-area oscillations (control capability of global swings).

The WT equivalents cover steady state, short circuit and stability calculations for the

different operation modes of WT-models dependent on the national grid code

requirements (e.g. FRT-capability, disconnection).

1.3. Test System The wind turbine models used within the study have been verified using a simple test

system with an infinite bus, a voltage source transformer and the respective wind turbine,

see Figure 1-1, where the tap changer is fixed:

1

2 3

Figure 1-1: Test system

Additional parameters for this test system can be found in and Table 1.

Assumptions:

MVASk 3000=′′

1.0=n

n

XR

fZ

kV110

kV20 kV690.0

Fault

%650==

cc

n

UMVAS

%1150==

cc

n

UMVASGrid Equivalent Wind Turbine

gZ wZ

MWPw 45=20/11012 =r 69.0/2023 =r

Pcc

1

2 3

MVASk 3000=′′

1.0=n

n

XR

MVASk 3000=′′

1.0=n

n

XR

fZ

kV110

kV20 kV690.0

Fault

%650==

cc

n

UMVAS

%1150==

cc

n

UMVASGrid Equivalent Wind Turbine

gZ wZ

MWPw 45=20/11012 =r 69.0/2023 =r

Pcc

9

• Transformers: Pcu = PFe = 0 MW

• Grid equivalent: voltage factor c = 1.0 .(although 1.1 is also correct [10])

In order to transform the equivalent grid parameters into the p.u. system, some

calculations are necessary:

( )Ω=== 033.4

30001100.1 2

''

21

MVAkVx

S

UxcZ

k

g

Ω=Ω=⇒=+= 4013,0013.41.0222

ggg

gggg RandX

X

RandXRZ

10

1.3.1. Generator Types

It has been decided, that three types of wind turbine should be taken into account, these

are

a) The squirrel cage induction generator (with or without fast voltage support),

b) The doubly fed induction generator and

c) The full size inverter model.

For the two latter ones the reactive power exchange at the point of common coupling is set

to zero.

1.3.2. Type of Investigations

To compare the behaviour of the wind turbine models, several investigations have been

made connecting the different wind turbine types to the test system.

A 150 ms symmetrical three-phase short circuit at the point of common coupling (PCC)

bus has been simulated with a voltage dip at the 110 kV node to

• 80%,

• 50% and

• Less than 30% of nominal voltage

To calculate the fault impedance to reach the desired voltage dip it has been used the

methodology of IEC [10].

• To obtain the voltage of 0.2 p.u. at the PCC, the fault reactance is set to 0.965 Ω.

• To obtain the voltage of 0.8 p.u. at the PCC, the fault reactance is set to 15.45 Ω.

Asymmetrical faults have not been investigated because no influence on the overall

system security is expected.

11

1.3.3. Grid Connection Requirements

Generally, there are several grid connection requirements for wind turbines valid in several

TSO regions, see some examples shown in Figure 1-2.

Special characteristics, such as fault ride through capability; voltage support or voltage

control depend on the national grid codes.

One of the study´s objectives is to investigate the effect of grid codes changes or their

adaptations on the overall system behaviour.

Figure 1-2: Present Situation: Different Grid Codes in Europe Thus, to achieve this goal with the models used in this study, a calibration possibility is

required which enables the turbine models to fulfil the European grid code requirements

e.g. with respect to

• Fault Ride Through Capability,

• Contribution to voltage support characteristics,

• Contribution to frequency maintenance.

12

1.4. Visualisation of the results

In terms of comparison the following variables have been visualized and compared:

• The generator voltage at the 20 kV generator bus bar and the 110 kV point of

common coupling (PCC) in [p,u.],

• Active and reactive power in [MW] resp. [MVar] at the 110 kV PCC,

• Active and reactive current in [kA] or [p.u.] at the 110 kV PCC,

1.5. Results

The results of the different wind turbine models connected to the test system have been

compared and evaluated. It was concluded that the respective wind turbine models, which

then are used in the study, behave in a way they are expected to do after initiated faults in

the test system.

Some exemplary graphics calculated with several software tools are presented for the

Doubly Fed Induction Generator WT Model and for the remaining two types of the WT

models namely squirrel Cage Induction Generator and Full Size Inverter Model. The

detailed results of all types of WT Models (model description and behavior) based on

various grid code requirements.

The example of the DFIG during a symmetrical three phase fault at the 110 kV PCC with

UFault = 0.2 p.u. for a wind farm with P = 30 *1,5 MW with voltage support is shown,

visualizing a simulation period of 1 second. The other respective investigations on different

turbine types and grid faults have been executed accordingly, but are not documented in

this report.

13

1.6. System parameters

System Component Parameter Value

c – Factor 1.

Voltage Magnitude Setpoint

1 pu

Equivalent Grid

Voltage Angle Setpoint

0°

Copper Losses – PCu 0 kW

No Load Current – I0 0 %

No Load Losses – PFe 0 kW

Vector Group HV-Side

YN

Vector Group LV-Side YN

Transformer 110/20 kV

Phase Shift 0°




Vector Group HV-Side

YN


Transformer 20/0.69 kV

Phase Shift 0°

Table 1 Additional System Parameters

The fault reactance in case of URES=0.2pu is equal to 1Ω and in case of URES=0.8pu is

equal to 15.5Ω.

14

Parameter Value Unit

Pn 1500 [kW] Un 690 [V] xn Un

2/Pn [Ω] ln xn/ω1 [H] p 2 ω1 2πfn [rad/s] fn 50 [Hz] r1 0.008 xn [Ω] r2’ 0.008 xn [Ω] l1s 0.080 ln [H] l2s’ 0.080 ln [H] lh 3.0 ln [H] l1 lh +l1s [H] l2 lh + l2s’ [H]

kP and kQ 3.56 e-5 [V/W] TiP and TiQ 0.0128 [s]

W21 2.7 Table 2: Parameter of DFIG

Parameter Value

Rs 0.008 p.u. Xs 0.080 p.u. Xm 3.0 p.u. Rr’ 0.008 p.u. Xr’ 0.080 p.u.

Table 3: Parameter of DFIG transformed into p.u.-system

15

The following settings are recommended:

Parameter Value Sn 51.379 MVA Un 0.69 kV n0 1500 rpm Rs 0.008 p.u. Xs 0.080 p.u. Xm 3.0 p.u. Rr’ 0.008 p.u. Xr’ 0.080 p.u.

Slip Range ± 20% UROTrat 1939V IROTmax 32.45kA *) RCROW 0.16pu

PSET-POINT_PCC 45 MW QSET-POINT_PCC 0 Mvar

PSET-POINT_GEN_STATOR 38 MW QSET-POINT_GEN_STATOR 7 Mvar

CDC 144412.8 µF

SGridFilter 20MVA UN_GridFilter 0.69 kV uK_GridFilter 31.8%

Copper Losses Grid Filter

21 kW

Table 4: Summary of electrical parameter settings *) Such a high value results from the assumption of the basis current for the rotor one of

the software tools

16

Parameter Value Base Power – PBASE 5 MW Turbine Damping - DTUR 0 Nms/rad Inertia of the Turbine – JWTR 6.1·106 kg m2 Inertia of the Generator – JGEN 101.72 kg m2 Shaft Damping - DSHAF 1.4·106 Nms/rad Stiffness Constant – kms 8.3·107 Nm/rad Nominal Angular Speed of

Turbine 18 rpm

Inertia of Equivalent Generator - JEQV=30xJGEN

3051.6 kg m2

Table 5: Summary of mechanical parameter settings

17

1.7. Exemplary Results

Exemplary results for the DFIG, exposed to a 150 ms three phase Test Fault with UFault =

0.2 p.u. - Crowbar operation in the range of 0.4 … 0.1 sec adjustable. With compliance to

the EWEA feedback and measurement a crow bar operation of 0.4 sec. will be adjusted in

the model.

Measurement provided by EWEA

18

Voltage profile:

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

pu

0,00

0,25

0,50

0,75

1,00

Software 1, Grid Code 1, Protection Scheme 1

Pcc (110kV bus) voltage (p.u.)

20kV bus voltage (p.u.)

Pcc (110kV bus) voltage (p.u.)

20kV bus voltage (p.u.)


0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Volta

ge [p

u]

U_pcc

U_20kV


19

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t[s]

/U/_PCC[p.j.] /U/_20KV[p.j.]

Software 4 with crowbar

20

Active Power and Reactive Power:

Act. Power Reac. Power

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

-1

0

1

2pu


P (MW) from 20kV node to Pcc (110kV node)*

Q (MVAr) from 20kV node to Pcc (110kV node)*


(*) Represented in the 20kV side





(*) Represented in the 20kV side



-100

-50

0

50

100

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]


21

-120

-100

-80

-60

-40

-20

0

20

40

60

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

t[s]

PV_PCC[ MW] QV_PCC[MVAr]

Software 4 with crowbar

22

Active Current and Reactive Current:

Act Current App. Current Reac. Current

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-3-2-10123

puVoltage support

Crow BarFiring

Post fault voltage support

Fault incident Software 1, Grid Code 1, Protection Scheme 1

Reactive current (kA) from 20kV bus to Pcc (110kV bus)

Active current (kA) from 20kV bus to Pcc (110kV bus)




-3

-2

-1

0

1

2

3

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [p

u]

RealImag


23

2. Wind turbine models based on grid code requirements in Germany

Figure 2-1: Grid code requirements [IEEE 2006, Montreal] Wind farm system control automatics

WF

NET

SSaaffeegguuaarrdd aatt ccoonnnneeccttiioonn ppooiinntt

Stepwise tripping of WT after 1,5 … 2,4

Tripping of WT after 0.5 s

SSaaffeegguuaarrdd aatt WWTT tteerrmmiinnaall nnooddeess

GGrriidd rreeqquuiirreemmeennttss

WT

Conditions: • Reactive power flow directed to wind farm • Voltage dip below 85%

Condition: voltage dip below 80%

Tripping of WT within 100 ms

Conditions: voltage about 120% or frequency exceeds limits 47.5 Hz – 51.5 Hz

Power reduction about 50.2 Hz

CCoonnnneeccttiioonn PPooiinntt

Fault Ride Through Contribution to voltage maintenance Contribution to frequency maintenance

Figure 2-2: [IEEE 2006, Montreal] Wind Farm Frequency Characteristics

24

PΔfNet

Power reduction

Frequency

P M Available power

P Δ

Hz50fHz50.2P20ΔP Netz

M−

=

fNet 50.2 Hz

ΔPΔP=40% PM / Hz

At fNet ≤ 47.5 Hz and fNet ≥ 51.5 Hz separation from grid

Between 47.5 Hz ≤ fNet ≤ 50.2 Hz no limitation

fNet

at 50,2 Hz ≤ fNet ≤ 51,5 Hz

Figure 2-3: [IEEE 2006, Montreal] DFIG and full size inverter model are fulfilling all grid code requirements in Germany

• FRT capability • Voltage support / voltage control • Power reduction in case of over frequency (Figure 2-3)

Squirrel cage induction generator (SCIG) model represents the old WT-installations and cannot fulfil the grid code requirements in Germany [7]. For this model no voltage support and FRT capability are adjusted. It is proposed to change the connection requirements for old installations to reduce the outage of wind power generation in case of voltage drops in Germany as followed:

• Delay of the trip signal in case of low voltage for 250 ms • No voltage support • Power reduction in case of over frequency (Figure 2-3)

25

2.1. Wind Turbine Protection Each wind turbine model is equipped with under voltage protection. The settings for the excitation, trip delay and tripping of the generator is adjustable. The following settings are adjusted within the models: DFIG, DDSM: U<< 0.15 pu (FRT capability, trip below 0.15 pu) U>> 1.20 pu Trip f>> 50.2 Hz (Reduction of active power generation) f<< 47.5 Hz Trip f>> 51.5 Hz Trip SCIG U<< 0.8 pu Trip U>> 1.10 pu Trip f>> 50.2 Hz (Reduction of active power generation) f<< 47.5 Hz Trip f>> 51.5 Hz Trip

26

2.2. Wind farm based on double fed induction generators The model which was developed jointly with the manufacturers, the FGW and the

transmission system operators and which also includes voltage support was used in the

investigations. The wind turbine model was adapted to the German Transmission code

requirements (TC 2007) (Fault Ride Through, voltage support, system automatics, active

power reduction in case of over frequency). For the most part the currently commissioned

wind turbines do not have to meet any extended connection conditions such as voltage

support due to the time of application. Nevertheless all wind turbines with voltage support

connected to the interconnected power system after 2004 were taken into account for the

dynamic investigations of the 2008 and 2015 variants.

WT-Control-Unit

win

dsp

eed

mec

h.to

rque

spee

d

Lim

its

Windspeed

Win

d m

odel

WEA-Model

Grid

Mod

el

Act

ive

pow

er

Rea

ctiv

e po

wer

Voltage

Current

U-Control

Figure 2-4: WT block diagram

27

active power

reactive power

speed

Figure 2-5: DFIG generator model

28

GeneratorModel

TransformationField to stator

Figure 2-6: DFIG simplified control model

r1 = 0.008 xnr2‘ = 0.008 xnl1s = 0.080 lnl2s‘ = 0.080 ln

un = 690. Vpn = 1500. kWxn = un2/pn (Ohm)ln = xn/w1 (H)

w21 = 2.7p = 2 J = 19800 kgm2

lh = 3.000 lnl1 = lh + l1sl2 = lh + l2s‘fn = 50 Hzw1 = 2p fn

ControllerkP = 3.56e-5 V/WTiP = 0.0128 s

kQ = 3.56e-5 V/WTiQ = 0.0128 s

Figure 2-7: DFIG model settings

29

The verification of the model is carried out with the terminal behaviour which is carried out

for the variants agreed on with the expert assessor. The active and reactive power, the

amount of current, the voltage at the 110 kV connection point, the voltage at the 20 kV

busbar, the rotor currents and the absolute value of the space vector of the rotor current

are depicted in the following figures.

Figure 2-8 shows a voltage dip to 20% UN at the grid connection point for the double fed

induction machine, crow bar operation for 100ms (Figure 2-9: crow bar operation for only

40ms). For the investigation crow bar firing operation of 100 ms will be taken into account.

Figure 2-14 shows that crow bar operation is a local phenomenon and has no influence on

the global system behaviour. The wind farm fulfils the German TransmissionCode 2007

(voltage support, post fault voltage support …). In Figure 2-10 to Figure 2-13 the dfig

model behaviour for different fault incidences (different fault duration time and voltage

drops) are shown.

Voltage drop to 25% nom. voltageFault duration: 150 msTransmissionCode (Germany) requirementsCrow bar firing (100ms)

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

pu

0,00

0,25

0,50

0,75


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

-1

0

1

2


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

0

2

4

pu

pu Voltage supportCrow BarFiring


Fault incident Figure 2-8: DFIG voltage dip to 25% UN at the PCC – Crow bar operation time: 100ms

30

Voltage drop to 25% nom. voltageFault duration: 150 msTransmissionCode (Germany) requirementsCrow bar firing (40ms)

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

pu

0,00

0,25

0,50

0,75

1,00


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

-1

0

1

2


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-3-2-10123

pu

puVoltage support

Crow BarFiring


Fault incident Figure 2-9: DFIG voltage dip to 25% UN at the PCC- Crow bar operation time: 40ms

Voltage drop to 30% nom. voltageFault duration: 150 msTransmissionCode (Germany) requirementsNo crow bar firing

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

pu

0,00

0,25

0,50

0,75

1,00


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

-1

0

1

2


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-2

-1

0

1

2

3

pu

puVoltage support (150 ms fault duration)


Fault incident Figure 2-10: DFIG voltage dip to 30% UN at the PCC

31


UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

pu

0,00

0,25

0,50

0,75

1,00


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

-1,0

0,0

1,0


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

-1

0

1

2

pu



Fault incident Figure 2-11: DFIG voltage dip to 50% UN at the PCC – Fault duration time: 300 ms


UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

pu

0,00

0,25

0,50

0,75


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

-0,5

0,0

0,5

1,0


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 1,3 1,4

-0,5

0,0

0,5

1,0

1,5

pu

pu Voltage support (1500 ms fault duration)


Fault incident Figure 2-12: DFIG voltage dip to 70% UN at the PCC – Fault duration time: 500 ms

32


UGEN UPCC

t/s0,25 0,50 0,75 1,00 1,25 1,50 1,75 2,00 2,25

pu

0,00

0,25

0,50

0,75


t/s0,25 0,50 0,75 1,00 1,25 1,50 1,75 2,00 2,25

-0,25

0,00

0,25

0,50

0,75

1,00


t/s0,25 0,50 0,75 1,00 1,25 1,50 1,75 2,00 2,250,0

0,5

1,0

1,5

pu



Fault incident Figure 2-13: DFIG voltage dip to 80% UN at the PCC – Fault duration: 1500 ms

ResidualVoltage [%]

Involved WTPower [MW]

≤ 20 % UN 80 MW

≤ 30 % UN 150 MW

≤ 40 % UN 200 MW

≤ 50 % UN 1200 MW

≤ 60 % UN 2500 MW

≤ 70 % UN 2900 MW

Crow Bar Firing

Figure 2-14: Potential gradient area: wind farms affected by crow bar operation

33

2.3. Wind farm based on squirrel cage induction generators

On the recommendation of the expert assessor, the model of the squirrel cage induction

generator was compared with the reference model of the 2 MW wind turbine with a fixed

speed of the following reference http://www.risoe.dk/rispubl/VEA/veapdf/ris-r-1331.pdf.

The terminal behaviour of the implemented model is visualised for the voltage dips of 25%

UN in Figure 2-15.

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,90,00

0,25

0,50

0,75


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-1,0

-0,5

0,0

0,5

1,0

1,5


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

-1

0

1

2

Voltage drop to 25% nom. voltageFault duration: 150 msTransmissionCode (Germany) requirements

pu

pu

pu

Fault incident

Figure 2-15: IG voltage dip to 25% UN at the PCC – Fault duration time: 150ms

34

http://www.risoe.dk/rispubl/VEA/veapdf/ris-r-1331.pdf

2.4. Wind farm based on full size inverter model

The employed model of the synchronous generator is based on a full converter model,

whose terminal behaviour was visualised. It is a 2.0 MVA turbine.

The single turbine model can be regarded as a wind farm model under the following basic

conditions:

the same wind conditions apply to all machines, i.e. the same internal operating state

exists

all machines have the same terminal voltage

residual current must not be smaller than 15% in case of a fault

settling time of the model is 5s

The simulation of selected faults yields the following results which have been visualised in

Figure 2-16.

UGEN UPCC

t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,90,00

0,25

0,50

0,75

1,00


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,90,0

0,5

1,0

1,5


t/s0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,90,0

0,5

1,0

1,5

pu

pu

pu

Voltage drop to 25% nom. voltageFault duration: 150 msTransmissionCode (Germany) requirements

Fault incident Figure 2-16: Full size inverter model during a voltage dip to 25 % UN at the PCC

35

3. Wind turbine models based on grid code requirements in Spain

3.1. Objectives A test procedure was proposed for the validation of wind turbine models to be used in the

EWIS study:

• Wind turbine models will be validated by matching the simulation of the model with

the simulation of most of the existing wind turbines and will be checked with the

actual measurement of the existing wind turbine developments in different

countries.

• Based on the approved minutes of the first EWIS WG3 meeting the behaviour of

the dynamic windfarm models must fulfil the sharpest connection requirements in

Europe. The proposal covers the requirements for wind farms based on [8].

• The tests are based on the parameters of the test model proposed

Red Eléctrica de España (REE) has developed, in FLECS language for PSS/E, a

simplified wind generator user model called SWGREE (simplified wind generator of REE),

oriented for planning studies, avoiding the complex manufacturer models. Consequently,

in this document, our objective has been to show how our model works, not to reproduce

accurately the behaviour of actual wind turbine models under any circumstance.

Several wind technologies have been simulated, as it was proposed initially . Changing the

value of the SWGREE model parameters, it is possible to simulate the typical behaviour of

those technically adapted wind technologies:

• SCIG (squirrel cage Induction generator)

• DFIG (doubly fed induction generator)

• DDG (Direct drive converter generator)

On the contrary, for SCIG turbines without voltage control, we have used a different PSS/E

user model, which fits more with its behaviour.

36

3.2. Grids used for the test

The wind turbine models have been tested with two different grids, the first of them is the

grid that REE always uses to test if the wind turbine model fulfils our grid code [16], and

the second is the grid proposed for the EWIS study . However, in this document, only the

results obtained with the EWIS grid test are presented (from 3.1 to 3.3.)

3.3. REE grid used to test wind turbines

In order to simulate the equivalent electrical grid, a dynamic system has been chosen

consisting of one bus in which the equivalent dynamic model of the UCTE is modelled,

along with another bus in which an equivalent model reflecting the dynamic characteristics

due to the hypothetical closest electrical grid (RED bus) and a third that represents the

grid connection point (PCR bus). Those buses are separated by impedances of

predetermined values in a way in which the typical voltage profile of the Spanish Electrical

System is reproduced. Then, it is guaranteed that all of the wind farms are tested, by

simulation, for short-circuits with equal characteristics.

Further explanations about this verification procedure can be consulted in [17]. The

purpose of this document is to provide a procedure for measuring and assessing the

response of wind farms in the event of voltage dips. This procedure must ensure the

uniformity of tests and simulations, the precision of measurements and the assessment of

the response of wind farms in the event of voltage dips.

37

UCTE equivalent400.000 MVA285.000 MW34.056 MVArH = 4,5 sZ’’ j 0,22 puVsched=1,042

Gen2Red Load 1600 MW700 Mvar

IZ

X = 0,004YNd11

R = 0,003 puX = 0,03 pu

R = 0,001 puX = 0,01 pu

X = 0,006YNd11

Red generator2.000 MVA1.600 MW700 MVArH = 3 sZ’’ j 0,22 puVsched=1,042

PCRX = 0,000025YNd11

Gen1

REDUCTE

B = 0,4 puB = 0,4 pu B = 0,1 pu B = 0,1 pu

UCTE load285.000 MW15.000 MVArIZ modelled

Bus Gen120 kV

Bus Gen220 kV

BusLoad Red

20 kV

UCTE equivalent400.000 MVA285.000 MW34.056 MVArH = 4,5 sZ’’ j 0,22 puVsched=1,042

Gen2Red Load 1600 MW700 Mvar

IZ

X = 0,004YNd11

R = 0,003 puX = 0,03 pu

R = 0,001 puX = 0,01 pu

X = 0,006YNd11

Red generator2.000 MVA1.600 MW700 MVArH = 3 sZ’’ j 0,22 puVsched=1,042

PCRX = 0,000025YNd11

Gen1

REDUCTE

B = 0,4 puB = 0,4 pu B = 0,1 pu B = 0,1 pu

UCTE load285.000 MW15.000 MVArIZ modelled

Bus Gen120 kV

Bus Gen220 kV

BusLoad Red

20 kV

Figure 3-1 REE grid proposed to test the grid code in wind turbines

The ride-through requirement (uninterrupted operation) and the reactive power

consumption limits regarding the Spanish Grid Code [16], for a three-phase balanced fault,

are described in figure 3-2.

0 0.5 1 15

nUU

10.95

0.8

0.2

Time (s)

Fault clearance

Fault starting

0.85

0.5

1 0.9 0

Voltage (p.u.)

0.5

tIIr

nUU

Q generation

0.85

Q consumption

(1) fault and recovery

(2) normal operation

(2)

(1)

Situations where generators must

remain connected - blue area for any type of grounded fault

Figure 3-2: Spanish grid code requirements

0.6

- Red dashed area for 2 ph isolated faults

0 0.5 1 15

nUU

10.95

0.8

0.2

Time (s)

Fault clearance

Fault starting

0.85

0.5

1 0.9 0

Voltage (p.u.)

0.5

nUU

Q generation

0.85

tIIr

Q consumption

(1) fault and recovery

(2) normal operation

(2)

(1)

(2)

(1)

Situations where generators must

remain connected - blue area for any type of grounded fault

0.6


0.60.6


38

In case of a disturbance in the grid, if the voltage at the grid connection point is inside the

dashed zone, the wind turbine generator is not allowed to be tripped.

In the REE EHV grid, 0.5 s is the longest planned time for fault clearance, assuming a

malfunction of the nearest breaker protection, then, it is necessary the actuation of the rest

of the back-up elements, as the breaker failure protection.

During the voltage dip, from the starting to the clearance of the fault, the reactive power

supplied by the wind turbine generator is severely reduced. Once the fault is cleared, at

0.5 s, the consumption of reactive power increases.

39

3.4. Wind turbine models

The increasing penetration of wind energy in power systems implies that Transmission

System Operators (TSOs) need to use dynamic models for transient stability studies. The

models developed by the wind turbine manufacturers reproduce the behaviour of their

machines with a great level of accuracy. Such level of detail is not relevant to the overall

standpoint of the TSO, and results in dynamic models not suitable for the simulation of

large power systems for two reasons. First, the use of these models usually requires prior

macros to obtain the dynamic data. Second, due to the large number of state variables to

perform, most of them out of the TSO concerning, the dynamic simulation length and

complexity are hugely increased.

From this experience, and in order to take into account the behaviour of wind generation in

the transient stability studies carried out by the TSOs, it has been developed a simplified

model of wind generator (SWGREE) [21], to be used with the PSS/E software. This model

is primarily aimed for planning studies, to evaluate the needs of the power system, and

has the same treatment and flexibility than any other PSS/E conventional generator

model. While it is indicated mainly for the technically adapted variable speed wind

turbines, which fulfil the Spanish Grid Code for voltage dips [16], it reproduces the typical

response of the known variable speed wind technologies, only adjusting its parameters. In

addition, the simplified model incorporates the possibility of adjustable modelling

of: power-frequency regulation, voltage dynamic regulation, and active power response

proportional to the derivative of frequency in order to avoid the effects of the increasing

loss of inertia of the system (see figure 3-3). These control capabilities qualify the model to

be used either for isolated system or large interconnected systems.

40

Some assumptions have been necessarily made to make the model such simplified. We

are not concerned about the inside wind turbine behaviour, then, monitoring many internal

variables of the machine is not useful for TSOs planning studies:

• It has been removed the need of the drive model. While the wind generator is not

tripped, the wind turbine will not show any mechanical problem which could modify its

usual response. The model incorporates a parameter which corresponds to the

maximum torque of the wind turbine. A detailed model of a wind turbine, including the

mechanical system, could be suitable for real time studies, i.e., if the intention is to

repeat the real response of each machine. For planning studies, where the target is to

decide which requirements have to be fulfilled by the wind turbines, it is more suitable to

use a simplified model, with a typical response for each technology.

• The converters keep their usual response while the wind generator is not tripped.

• The power electronics is able to follow the external frequency all the time.

Power-frequency

control module

Voltage control

module

(de)magnetizing

current module

Converter limiter

module

activeI

reactiveImeasureV

converterI

magnetI

sorceI

busI

sorceZ busV

measure f

frequencydeadband

Power-frequency

control module

Voltage control

module

(de)magnetizing

current module

Converter limiter

module

activeI

reactiveImeasureV

converterI

magnetI

sorceI

busI

sorceZ busV

measure f

frequencydeadband

Figure 3-3: Simplified block diagram of the REE wind generator model (SWGREE)

41

The input parameters that this model needs are:

• Voltage control module: gain (Kv) and time constant (Tv). Maximum transient reactive current with respect to the total amount of current (Ir/It)

• Coordinates describing right diagram of figure 3-2 (U/Un vs. Ir/It)

• Magnetizing and demagnetizing current time constant

• Power-Frequency control: Droop and time constant of the control

• Limiters: mechanical output power, steady state and transient current

• Power response proportional to the derivative of frequency: gain and time constant

• Converter time constant

Depending on the wind generator technology, different parameters have to be set. Besides

those model parameters, the source impedance (Zsorce for PSS/E) may be changed to

adjust, mainly, the reactive power peaks of those different technologies:

• for DFIG or SCIG Zsorce is set to 0.25 p.u.

• for DDG the Zsorce is set to 100 p.u., (so high to avoid the (de)magnetizing discharge peak)

The power-frequency control has been disabled in this test. It does make more sense for

islanded systems. It has been proved for wind penetration studies in islanded systems and

it worked properly.

The following variables are monitored for the description of the terminal behaviour of the

wind farm models:

• Active and reactive power (in MW/MVAr) at the connection point.

• Active and reactive current (in kA) at the connection point.

• Amount of current (in kA) at the connection point.

• Generator voltage at the generator bus bar in p.u.

• Frequency (in Hz) at the connection point (only with frequency variation)

In the loadflow there has to be no power reactive flow through the PCC bus. In order to

cope with that, the wind generator at the 690 V bus has to supply the necessary reactive

power, either with a capacitor, in the case of SCIG or, in the case of DDG or DFIG, by

42

reducing its power factor. In the figures, the active and reactive power flow from 20 kV bus

to Pcc (110 kV bus) in the 20 kV side of the bus are represented. It has to be taken into

account the reactive power consumption by the transformer, so this flow is not zero in the

20 kV side of the transformer, but it is zero in the 110 kV side.

From the bus voltages and the active and reactive power values, we calculate the currents

(amount of current and active/reactive current). The value of those currents, from 20 kV

bus to 110 kV bus, in the 110 kV side of the transformer is represented. We can

appreciate that the initial reactive current in the loadflow is different to zero. This happens

because of the initial value of the reactive power, as explained in last paragraph.

The shape of the responses of the grid machine equivalent, at Pcc (110 kV bus), is not very

realistic. The dynamic model used is a classical constant voltage behind transient

reactance generator model (GENCLS in PSS/E). This is the main reason to consider the

variables in the 20 kV side and make the conversion to the 110 kV side instead of

considering the variables of the grid equivalent.

To represent the mechanical coupling of the non adapted SCIG (without voltage control) it

has been used a two masses model (TUREOL).

The electric part of the SCIG is modelled by a user model called GENIND (induction

generator). The model adds, in parallel with the generator terminals, a compensation

reactance to adjust the reactive power injected assigned in the loadflow.

43

Its input parameters and values to set are:

- Stator resistance: Rs=0.00779 p.u. (machine base)

- Stator reactance: Xs=0.07937 p.u. (machine base)

- Rotor resistance: Rr=0.00779 p.u. (machine base)

- Rotor reactance: Xr=0.1158 p.u. (machine base)

- Magnetizing reactance: Xm=4.1039 p.u. (machine base)

- Inertia: H=0.423 s

- Damping: D=0

To model the DFIG and DDG without voltage support it has been used the simplified

model, SWGREE, changing the value of the reactive current that the generator has to

supply during the voltage dip. In the model, “voltage control” does not refer to the fault ride

through capability but to the capability of supplying reactive current during the voltage dip

and once the fault is cleared, in the same way as an AVR. The parameters have to be set

like this:

- Voltage control enabled: Ir/It=1 and the gain of this control has to be adjusted

properly

- To disable the voltage control, Ir/It=0 and the gain of this control is set to 0.

In order to trip the wind generator, according to the voltage dip curve of our Grid Code

(see figure 3-2), we have modelled another PSS/E user model, called PRMTCR,

(undervoltage delayed relay). The input parameters of this model are, besides the points

that describe the voltage dip curve, a voltage slope threshold at the fault starting.

Sometimes the DFIGs trip due to a huge reduction, with a very sharp slope, of the bus

voltage at the fault starting. Besides, the model incorporates an overspeed protection

relay.

In the following sections, 3.1 to 3.3, the results are presented. The figures presents the

results obtained with REE’s user models.

44

3.5. Doubly fed induction generator model

• DFIG. Voltage dip to 20% UN at the PCC without voltage support

Pcc (110kV bus) voltage (p.u.) 20kV bus voltage (p.u.)

Pcc (110kV)

20kV 0.69kV

Figure 3-4: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.2pu). Voltage at 20 and 110 kV bus.

Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

45

Active power (MW) from 20kV bus to Pcc (110kV bus) Reactive power (MVAr) from 20kV bus to Pcc (110kV bus)

Pcc (110kV)

20kV 0.69kV

Figure 3-5: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.2pu). Active and reactive power.

Figure 3-6: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.2pu). Total current.

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Current (kA) from 20kV bus to Pcc (110kV bus) referred to the 110kV side of the transformer

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)

46

Reactive current (kA) from 20kV bus to Pcc (110kV bus) Active current (kA) from 20kV bus to Pcc (110kV bus)

Pcc (110kV)

20kV 0.69kV

Figure 3-7: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.2pu). Active and reactive current.

Figure 3-8: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.2pu). Frequency.

Active current (kA)

Reactive current (kA)aIrI

Pcc (110kV)

20kV 0.69kV

Active current (kA)


Frequency deviation (Hz) respect 50 Hz at Pcc bus (110 kV)

47

• DFIG. Voltage dip to 80% UN at the PCC without voltage support


Pcc (110kV)

20kV 0.69kV

Figure 3-9: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.8pu). Voltage at 20 and 110 kV bus.

Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

48


Pcc (110kV)

20kV 0.69kV

Figure 3-10: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.8pu). Active and reactive power.

Figure 3-11: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT =0.8pu). Total current

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)


Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)

49

Pcc (110kV)

20kV 0.69kV

Figure 3-12: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT =0.8pu). Active and reactive current

Figure 3-13: Simulation of 3-phase 150ms fault at 110 kV node (UFAULT=0.8pu). Frequency

Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)




50

• DFIG. Voltage dip to 20% UN at the PCC with voltage support

20kV bus voltage (p.u.) Pcc (110kV bus) voltage (p.u.)

Pcc (110kV)

20kV 0.69kV

Figure 3-14: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Voltage at 20 and 110 kV bus.

Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

51


Pcc (110kV)

20kV 0.69kV

Figure 3-15: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Active and reactive power.

Figure 3-16: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Total current.

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


52


Pcc (110kV)

20kV 0.69kV

Figure 3-17: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Active and reactive current.

Figure 3-18: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Frequency

Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



53

• DFIG. Voltage dip to 80% UN at the PCC with voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

54

Reactive power (MVAr) from 20kV bus to Pcc (110kV bus) Active power (MW) from 20kV bus to Pcc (110kV bus)

Pcc (110kV)

20kV 0.69kV


Figure 3-21: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu).

Total current.

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)


Pcc (110kV)

20kV 0.69kV

Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)


55

Pcc (110kV)

20kV 0.69kV


Figure 3-23: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu). Frequency.

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)



56

3.6. Squirrel cage induction generator model

• SCIG. Voltage dip to 20% UN at the PCC without voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

57


Pcc (110kV)

20kV 0.69kV

Figure 3-25: Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Active and reactive power.

Figure 3-26: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Total current

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


58


Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



59

• SCIG. Voltage dip to 80% UN at the PCC without voltage support


Pcc (110kV)

20kV 0.69kV

Figure 3-29: Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu). Voltage at 20 and 110 kV bus.

Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

60


Pcc (110kV)

20kV 0.69kV



P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


61


Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



62

• SCIG. Voltage dip to 20% UN at the PCC with voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

63


Pcc (110kV)

20kV 0.69kV



P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


64


Pcc (110kV)

20kV 0.69kV



Frequency.

Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



65

• SCIG. Voltage dip to 80% UN at the PCC with voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

66


Pcc (110kV)

20kV 0.69kV



Total current.

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


67

Active current (kA) from 20kV bus to Pcc (110kV bus) Reactive current (kA) from 20kV bus to Pcc (110kV bus)

Pcc (110kV)

20kV 0.69kV



Frequency.

Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



68

Direct drive generator model

• DD. Voltage dip to 20% UN at the PCC without voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

69

Pcc (110kV)

20kV 0.69kV



P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)


Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


70

Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)


Active current (kA) from 20kV bus to Pcc (110kV bus) Reactive current (kA) from 20kV bus to Pcc (110kV bus)


71

• DD. Voltage dip to 80% UN at the PCC without voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

72


Pcc (110kV)

20kV 0.69kV


Figure 3-51: Simulation of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu). Total current .

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


73

Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)




74

• DD. Voltage dip to 20% UN at the PCC with voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

75


Pcc (110kV)

20kV 0.69kV



P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


76



Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



77

• DD. Voltage dip to 80% UN at the PCC with voltage support


Pcc (110kV)

20kV 0.69kV


Voltage(p.u.)

Pcc (110kV)

20kV 0.69kV

Voltage(p.u.)

78


Pcc (110kV)

20kV 0.69kV



P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

P (MW)

Q (MVAr)

P (MW)

Q (MVAr)

Pcc (110kV)

20kV 0.69kV

Current(kA)

Pcc (110kV)

20kV 0.69kV

Current(kA)


79


Pcc (110kV)

20kV 0.69kV



Active current (kA)


Pcc (110kV)

20kV 0.69kV

Active current (kA)



80

4. Grid code requirements and wind turbine models used in Czech Republic

4.1. Grid code requirements A special chapter in the 4th part of the Czeck Republic Grid code (Development planning of

the transmission system) deals with rules for connecting of wind turbine (WT) to the

ntation only.

transmission system. There are several figures for prese

igure 4-1 Fault Ride Through capability –behaviour of WR during near short circuit

U/Un [%]

t [ms]

100

50

0 150 1000 2000 3000650 1600

WT can be disconnected

WT must not be disconnected

85

15

F IR [p.u.]

-0,5

U [p.u.]1,00,5 0,9 Recommended

-1,0

Required

Figure 4-2 Voltage support - additional reactive current during decreasing of voltage

81

Active power PWT

80 %

60 %

40 %

20 %

50 f [Hz]

100 %

51 52 53494847

A

Switching off

BC

E

D

Power reduction

Normal operation Switching off

Figure 4-3 WT behaviour during frequency deviations - power reduction

4.2. Squirrel cage induction generator model The model of Squirrel Cage Induction Generator (SCIG) used in the MODES network

simulator was tested on simple Test System (corresponding to the chapter 1.3).

The following figure depicts the one line scheme.

Figure 4-4 One line scheme of the test system The capacitor bank C was selected to compensate approximately 65% of reactive power

consumption of the SCIG. The short circuit 0.15 s near PCC was simulated for voltage dip

80% (XF=1.05 Ω).

Parameter nSYN(rpm) Tm(s) X1(-) X2’(-) Xm(-) R1(-) R2’(-) SCIG 1500 8 0.125 0.125 4. 0.01 0.01 DFIG 1500 14 0.08 0.08 3 0.008 0.008 Table 6: Parameters of asynchronous generators

Following figures show calculation results.

82

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t[s]

/U/_COL_20KV[p.j.] /U/_PPC[p.j.]

Figure 4-5 Time courses of network voltages in per units

-100

-80

-60

-40

-20

0

20

40

60

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

t[s]


Figure 4-6 Time courses of power flows to the PCC bus

83

200

250

300

350

400

450

500

550

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t[s]

IV_PCC[A]

Figure 4-7 Time course of current to the PCC bus

-0.026

-0.024

-0.022

-0.020

-0.018

-0.016

-0.014

-0.012

-0.0100.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

t[s]

SR_WTG[p.j.]

Figure 4-8 Time course of the asynchronous generator slip (speed deviation)

84

4.3. Double Fed Induction Generator model The model of Double Fed Induction Generator (DFIG) used in the MODES network

simulator was tested on the same Test System like SCIG but without condensator batery -

there were 30 units 1.5 MW each (with parameters in the Table 6) .

The DFIG can operate in two regimes depending on the terminal voltage. Both regimes

are indicated in the following figure (corresponding models in the lower part):

I=(PG-jQ)/UG*

for UG>Umez for UG<Umez0

C CR

UG

E’

CX’

C

X’

UG

PG=PZ Q=QG +UG2/X’

IG

IG IG

jX’=j[X1+(X2+Xmi)/X2/Xmi] Figure 4-4-9 Two regimes of the as. generator with converter and corresponding models

The rotor winding is supplied from the converter in normal operation (with terminal voltage

above defined threshold value Umez). Then the DFIG is modelled simply by current

injection with parallel transient reactance connected between terminals and ground.

Components of the current (active and reactive power injections) are controlled by the

converter. This control is presented on and is described later.

The rotor part of the converter is blocked and rotor circuit is closed (shorted) through a

resistance R=0.16 p.u. when terminal voltage decreases under threshold value Umez.

Network part of the converter can supply a reactive power form a capacity C. Then the

DFIG is modeled by conventional induction machine with squirrel-cage rotor and additional

resistance R. If the voltage increases the defined threshold value Umez0 the model is

switched into the first regime model and DFIG is modelled by injected current again.

In both regimes rotor acceleration is calculated by motion equation. The turbine and

generator per unit speed ω is the same, so that torsion of the shaft is neglected.

85

The wind turbine model implemented in the MODES is depicted in the upper part of the

following figure. It consists of static model of the turbine output, speed and pitch control.

Turbine output (mechanical power) depends on the following variables:

• ω turbine speed

• β pitch angle

• vE equivalent wind speed.

Figure 4-10 Block scheme of connection the wind turbine and DFIG models

86

The model of the active power control was created according reference [12] and it is

compatible to the PSLF model with two exceptions. Power is controlled instead of torque

and generated power PG is equal directly power order (converter lags are neglected).

Following figures show calculation results.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t[s]

/U/_PCC[p.j.] /U/_20KV[p.j.]

Figure 4-11 Time courses of network voltages in per units

-120

-100

-80

-60

-40

-20

0

20

40

60

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

t[s]


Figure 4-12 Time courses of power flows to the PCC bus

87

0

100

200

300

400

500

600

700

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t[s]

IV_PCC[A]

Figure 4-13 Time course of current to the PCC bus

-0.208

-0.207

-0.206

-0.205

-0.204

-0.203

-0.202

-0.201

-0.200

-0.1990.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

t[s]

SR_WTG[p.j.]

Figure 4-14 Time course of the asynchronous generator slip (speed deviation)

88

5. Validation of Wind Turbine Models - SCIG - in Denmark

(Wind Turbines with Squirrel Cage Induction Generator)

5.1. Introduction This report describes the validation process of the wind turbine (WT) with squirrel cage

induction generator (SCIG) in the PowerFactory simulation environment. The validation

process was performed according to the procedure included in [15]. Also the available

information about the parameters of the test system as well as about the machine

parameters has been taken from [15].

5.2. Test System The test system defined in [15] has been implemented into PowerFactory as shown in

Figure 5-1 .

Figure 5-1 Test System Implementation into PowerFactory

89

The parameters given for the system components are summarized in Table 7.

System Component Parameter Value SK’’ 3 GVA Equivalent Grid RN/XN 0.1 U1rat 110 kV U2rat 20 kV UK 11%


SN 50 MVA U1rat 20 kV U2rat 0.69 kV UK 6 %


SN 50 MVA

Table 7: Parameters of the Test System The description of the test system given in [15] does not define all parameters required in

the simulation software. Thus, the values of missing parameters have been assumed as

given in Table 8.

System Component Parameter Assumed Value

c – Factor 1.

Voltage Magnitude Setpoint 1 pu

Equivalent Grid

Voltage Angle Setpoint 0°




Vector Group HV-Side YN



Phase Shift 0°

Cooper Losses – PCu 0 kW






Phase Shift 0°

Table 8: Assumed Values of Additional System Parameters

90

5.3. Wind Turbine Model with SCIG

5.3.1. Electric Parameters

The model of the wind turbine with SCIG has been in general parameterized according to

[16] since it was the given reference in [15]. The original data of the SCIG are given in

Table 9. Parameter Value

Sn 2.3 MVA Un 0.96 kV n0 1500 rpm Rs 0.004 Ω Xs 0.05 Ω Xm 1.6 Ω Rr’ 0.004 Ω Xr’ 0.05 Ω

Table 9: Parameter of SCIG given in [16] The parameters of the generator equivalent circuit has been transformed into the per unit

system using a basis impedance as defined by Eq. (1).

Ω=== 4007.03.2

)96.0( 22

MVAkV

SUZ

n

nBASE (1)

The resulting per unit parameters are given in Table 3.

Parameter Value


Table 10: Parameter of SCIG given in [16] transformed into p.u.-System However, the SCIG model analyzed in [16] was connected to a grid at the voltage level

0.96 kV. In order to connect this generator to the defined test system from [15], the rated

voltage of the generator was changed to Un=0.69 kV in the simulation software. The p.u.-

parameters of the machine have not been changed. Moreover, the rated power of the

machine has been modified according to [15] and set to Sn_FARM=22*Sn=50.6 MVA. As

91

setpoint for the simulation the active power of the machine has according to [15] been set

in PowerFactory to P=44 MW. The slip value and reactive power drawn by the machine

from the grid has been automatically determined by PowerFactory using the internal

machine parameters.

5.3.2. Mechanical System Representation

The mechanical system of the wind turbine has in a first step been represented as

lumped-mass model. The equivalent inertia for the whole farm with Sn_FARM=50.6 MVA has

been calculated using the parameters of the two-mass system given in [16] for the wind

turbine with Sn=2.3 MVA, which are summarized in Table 11.

Parameter Value

Inertia of the Turbine – JWTR 4.176·106 kg m2 Inertia of the Generator– JGEN 93.22 kg m2 Stiffness Constant – kms 8.949·107 Nm/rad Gear Box Ratio – fGB=nGEN/nWTR 80

Table 11: Parameter of Mechanical Parameters of WT with Sn=2.3 MVA [16]

Using given parameters of the mechanical system the inertia constant H has been

calculated for the 2.3 MVA wind turbine using Eq. (2) according to [17].

n

m

SJH

205.0 ω⋅

= (2)

Where J – is the inertia in kg m2, ω0m – is the nominal mechanical angular speed in rad/s

and Sn – is the rated apparent power in MVA.

Since the nominal angular speed of the turbine and the generator are different and are

related to each other by the gear box ratio – fGB, the right value has to be used for the

calculation of inertia constants of both elements. With this in mind the following values of

the angular speed have been used for the generator and the turbine – Eq. (3) and Eq. (4),

respectively.

⎥⎦⎤

⎢⎣⎡=⋅=⋅=− sradnGENm 0796.157

6021500

602

00ππω (3)

92

⎥⎦⎤

⎢⎣⎡=⋅⋅=⋅⋅=− srad

fn

GBWTRm 9635.1

801

60215001

602

00ππω (4)

Moreover, it was assumed that for the calculation of the inertia constant of the turbine the

same value of Sn=2.3 MVA as in case of the generator is used.

The resulting inertia constants for the generator and the turbine obtained with parameters

from Table 11 are given by Eq. (4) and Eq. (5), respectively.

5.0=GENH (5)

4999.3=WTRH (6)

For the lumped-mass model the equivalent inertia constant is treated as a sum of the

inertia constant for the generator and for the turbine [18] and equals for the considered

wind turbine HOneM=3.9999.

In the analysis at hand the whole farm is represented as one single large equivalent unit

with Sn_FARM=50.6 MVA. It has been assumed that the inertia constant H has the same

value as in case of the small unit (Sn=2.3 MVA), which has been scaled up in order to

represent the farm equivalent, since it is in per unit [17].

This constant HOneM can directly be applied for the simulation of the farm behavior in some

software packages. However, the tool PowerFactory requires the definition of the inertia J

in [kg m2]. For this purpose the value of inertia for the whole farm JFARM has been

calculated with Eq. (7).

262_2

0

52.16405106.500796.1579999.322 mkgSHJ FARMn

GENm

OneMFARM =⋅⋅

⋅=⋅

⋅=

−ω (7)

The resulting equivalent inertia of lumped-mass system of the farm can be also calculated

according to Eq. (8).

TURBGB

WTRGENFARM N

fJ

JJ ⋅⎟⎟⎠

⎞⎜⎜⎝

⎛+= 2 (8)

Where the JGEN and JWTR are inertias for the original turbine with Sn=2.3 MVA, fGB – is the

gear box ration and NTURB is the number of wind turbines collected to one single equivalent

unit. In this case NTURB=22.

93

The stiffness of the shaft given in Table 11 was transformed into the per unit system using

Eq. (9) according to [19].

47748.0280502

103.210949.8

226

7

220 =

⋅⋅⋅

⋅⋅⋅

=⋅

⋅=πω

pfSkK

GBBASE

msms (9)

Where p – is the number of pole pairs and the obtained value of stiffness - Kms is

expressed in pu/el. rad.

5.3.3. Compensation of Reactive Power In the test system defined in [15] there is no information about the compensation of

reactive power drawn by the SCIG.

However, test simulations in PowerFactory performed without any compensation device

show that the reactive power value in the point of common coupling on the 110 kV level is

significantly higher than the one given on the curves from [15].

The value given in [15] is approximately equal to 14 MVar (Figure 6, [15]), while the value

obtained in the simulation with PowerFactory without compensation is equal to 34.4 MVar,

see Figure 5-8.

Thus, in order to get the same reactive power value on the 110 kV level an additional

capacitor has been connected to the connection node of the SCIG at 0.69 kV level.

The size of the capacitor has been adjusted to obtain the desired reactive power level at

the 110kV-PCC and was equal to 39326 µF – which corresponds to 17.65 MVar.

According to [16] the capacitor bank has been represented as Delta – circuit. The

respective results, which are similar to the results from [15] can be found in Figure 5-9

94

5.3.4. Protection System

Wind turbines are generally equipped with a complex protection system. This system

controls various parameters of the wind turbine, as e.g. the mechanical, the thermal and

the electric parameters. In case of exceeding a defined limit concerning some parameter,

the wind turbine is disconnected from the grid. In case of electric parameters the voltage

and the frequency as well as the current are observed. The protection scheme used in the

wind turbines with SCIG causes immediate disconnection from the grid in case of under-

voltage in case of a grid fault.

The default behavior of the protection system can be different in each country according to

the TSOs' requirements defined in the respective GridCode.

The Danish grid code for Wind Turbines above voltages of 110 kV [20] states:

"The wind farm shall be equipped with voltage and frequency relays for disconnection of the wind farm at abnormal voltages and/or frequencies. The relays shall be set according to agreement with the regional grid company and the system operator. The protective functions of the wind turbine shall have settings and time delays meeting the requirements in section 8.2. [..] By means of simulation the plant owner shall document the behaviour of the wind farm by application of a fixed voltage profile. […] The turbine test shall be done with the voltage profile shown in (the figure below) and shall show the behaviour of the wind farm in the case of a three-phase fault with a slowly recovering voltage"

Figure 5-2 Voltage Profile for Simulation of Three Phase Fault. Source: [20]

95

In the model at present used for this validation of turbine models similar to [15] no

protection system has been implemented.

During the system analysis a respective protection scheme will be implemented into the

wind farm equivalent.

5.4. Analysis of the Test System with SCIG Wind Turbine

5.4.1. Three Phase short circuit, UPCC = 0.2 p.u. In the test system a three-phase short circuit at the 110 kV bus has been simulated. The

duration of the fault was set to 150 ms. In the first case the reactance of the short circuit

was adjusted in order to obtain the desired level of the residual voltage at the 110 kV node

equal to 0.2 p.u., which was the requirement from [15]. For this purpose the reactance

XFAULT has been set to XFAULT = 1 Ω. The simulation was performed for the following

system configurations:

- without reactive power compensation,

- with reactive power compensation.

The results of the simulation for both configurations are shown in Figure 5-3 and in Figure

5-4, respectively.

From the obtained results it can be seen that the results without a reactive power

compensation unit in the test system modelled in PowerFactory differ significantly from the

ones given in [15]. If the above described assumed compensation unit has been

connected to the test system, the results are similar to the ones presented in [15].

However, some differences between the results in [15] and Figure 5-3 concerning the

profiles of current, voltage and reactive power after fault-clearing can be recognized.

These differences can be the consequence of different parameters used for system inertia

and for the compensation units in both compared systems.

This has been proven by variant calculations which have shown that a change of the

inertia to a significantly higher value, (e.g. to J=100000 kg m2) allows to obtain the

simulation results more similar to the ones given in [15], but this inertia is not realistic.

96

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9Time [s]

Volta

ge [p

u]

1

U_pcc

U_20kV

-100-75-50-25

0255075

100

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0,6

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

-3

-2,5

-2

-1,5

-1

-0,5

0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Slip

[%]

Figure 5-3 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu) – without reactive power compensation

97

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9Time [s]

Volta

ge [p

u]

1

U_pcc

U_20kV

-100-75-50-25

0255075

100

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0,6

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

-3

-2,5

-2

-1,5

-1

-0,5

0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Slip

[%]

Figure 5-4 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu) – with reactive power compensation

98

5.4.2. Three Phase short circuit, UPCC = 0.8 p.u.

In the second part of the analysis a three-phase short circuit with residual voltage equal to

0.8 pu was simulated at the 110-side of the PCC. For this purpose the fault reactance was

set to 15.5 Ω. The results are given in Figure 5-5 and Figure 5-6 for a case without and

with a reactive power compensation unit, respectively. The comparison of these results

with the ones presented in [15] shows the same trend as in case of the fault with residual

voltage UFAULT=0.2 pu (see above).

The inertia of the wind turbine system has a crucial influence on the behaviour during and

after the fault, since the reactive power drawn by the SCIG depends strongly on the

machine slip and this, in turn, depends on the inertia.

99

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9Time [s]

Volta

ge [p

u]

1

U_pcc

U_20kV

-75

-50

-25

0

25

50

75

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

-2

-1,6

-1,2

-0,8

-0,4

0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Slip

[%]

Figure 5-5 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu) – without reactive power compensation

100

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9Time [s]

Volta

ge [p

u]

1

U_pcc

U_20kV

-75

-50

-25

0

25

50

75

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

-1,6

-1,2

-0,8

-0,4

0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Slip

[%]


101

5.5. Influence of Two – Mass Model Since the analysis from section 5.4 have been performed using a lumped – mass model

for the whole wind turbine (respectively upscaled wind farm equivalent), the characteristic

oscillations of the angular speed and therefore of the active and reactive power can be

neglected. In order to check the influence of a two-mass representation of the mechanical

system on the behaviour of the wind turbine the previously introduced model has been

enhanced according to [18].

The model is described with the following set of equations Eq. (10) – Eq. (14).

MMGMM

WTR DTTdt

dH ωω⋅−−=2 (10)

MGEGG

GEN DTTdt

dH ωω⋅−−=2 (11)

( )GMS

dtd ωωωθ

−⋅= 0 (12)

M

MM

PTω

= (13)

( )MGSSSG DKT ωωθ −⋅−⋅= (14)

Where: HWTR and HGEN – are inertia constants of turbine and generator, respectively,

ωM and ωG – are the angular speed of turbine and generator in pu, respectively, θS – is the

deflections angle of the soft shaft in elect. rad,

TM, TG, TE – are per unit torque generated by turbine, per unit torque applied to the

generator and per unit electromechanical torque of generator, respectively; DM, DG and DS

are the damping coefficients and KS is stiffness of the shaft in pu/rad.

The parameters used for the simulation of the two-mass mechanical system are

summarized in Table 12. Parameter Value

KS [pu/rad] 0.47748 DM [pu] 0 DG [pu] 0 DS [pu] 0 HWTR 3.4999 HGEN 0.5

Table 12: Parameters of Two-Mass Mechanical System

102

The value of the shaft stiffness – KS was calculated with Eq. (15) using the kms - parameter

given in Table 11.

47748.0280502

103.210949.8

226

7

220 =

⋅⋅

⋅⋅⋅

=⋅

⋅=πω

zGBn

msS pfS

kK (15)

Where:

pz – is the number of pole-pairs and

ω0 – is the angular speed corresponding to the nominal grid frequency.

In order to evaluate the influence of the mechanical model on the system behaviour the

same simulation cases as in section 5.4 were performed. The resulting curves for the case

of residual voltage equal to 0.2 pu and with a reactive power compensation unit are shown

in Figure 5-7. The other results are summarized in the appendix 5.7.2.

The results show that there are significant differences in the behaviour of the wind turbine

with two-mass model comparing to the lumped-mass model during and after the fault. The

slip in the two-mass model oscillates considerably after fault occurred which leads to the

fluctuation of active and reactive power of the machine. These power oscillations have a

large influence on the voltage profile on the 20 kV level.

Furthermore, in case of a fault with residual voltage equal to 0.2 pu and without reactive

power compensation the system becomes unstable due to increasing magnitude of slip

oscillations, Figure 5-10.

General practice at Energinet.dk is to use two-mass models.

103

0

0,2

0,4

0,6

0,8

1

0 0,5 1 1,5 2 2,5Time [s]

Volta

ge [p

u]

3

U_pcc

U_20kV

-150

-100

-50

0

50

100

150

0 0,5 1 1,5 2 2,5Time [s]

Act

.- &

Rea

ct. P

ower

3

P_pcc [MW]

Q_pcc [Mvar]

0

0,2

0,4

0,6

0,8

0 0,5 1 1,5 2 2,5 3Time [s]

Cur

rent

PC

C [k

A]

-8

-6

-4

-2

0

2

0 0,5 1 1,5 2 2,5 3Time [s]

Slip

[%]


104

5.6. Conclusions and Recommendations The analysis described in this paper shows that the detailed definition of all parameters of

the test system is crucial for the simulation results using different software and gives

respective recommendations.

Moreover, there can be some differences in the simulation results provided by various

software tools for dynamic calculations, even using the same simulation parameters, since

there can be differences in the internal representation of the system components – e.g.

machines – in each software package, as for example different transformation formula into

the per unit system or different definition of some machine parameters. Additionally,

depending on the software used, the test system parameters are required in different form.

For example PowerFactory requires the definition of inertia of the generator exclusively as

J in [kg m2]. Out of this input data the inertia time constant TM=2H is then calculated

internally.

105

5.6.1. Recommendations for Simulation Settings In order to get comparable results to [15], which are stable, the following settings are

recommended:

- use of two-mass model with the following system parameters:

Parameter Value

Sn 50.6 MVA Un 0.69 kV n0 1500 rpm Rs 0.00998 p.u. Xs 0.12478 p.u. Xm 3.9931 p.u. Rr’ 0.00998 p.u. Xr’ 0.12478 p.u.

PSET-POINT 44 MW

Table 13: Electrical Parameter of the Equivalent Farm Model

Parameter Value Inertia of the Turbine – JWTR 91.872·106 kg m2 Inertia Constant of the Turbine – HWTR 3.4999 Inertia of the Generator– JGEN 2050.84 kg m2 Inertia Constant of the Generator– HGEN 0.5 Stiffness Constant – kms 0.47746 pu/rad Gear Box Ratio – fGB=nGEN/nWTR 80

Table 14: Mechanical Parameter of the Equivalent Farm Model

The damping constants of the mechanical system were assumed to 0.

The size of the compensating capacitor at the terminals of the generator (0.69 kV level)

was equal to 39236 µF, which correspond to 17.65 MVar (Delta-circuit).

Fault reactance in case of URES=0.2pu is equal to 1Ω and in case of URES=0.8pu is equal to

15.5Ω.

106

5.7. Appendix - SCIG, Denmark

5.7.1. Simulation Results with PowerFactory – Load Flow

Figure 5-8 Load Flow Results in System without Compensation

107

Figure 5-9 Flow Results in System with reactive power compensation (Qrat=17.65 MVar)

108

5.7.2. Simulation Results with Two-Mass Model in PowerFactory

0

0,2

0,4

0,6

0,8

1

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5Time [s]

Volta

ge [p

u]

U_pcc

U_20kV

-150

-100

-50

0

50

100

150

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,2

0,4

0,6

0,8

1

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5Time [s]

Cur

rent

PC

C [k

A]

-100

-80

-60

-40

-20

0

20

0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5Time [s]

Slip

[%]

Figure 5-10 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu) – without reactive power compensation (Two-Mass Model)

109

0

0,2

0,4

0,6

0,8

1

0 0,5 1 1,5 2 2,5Time [s]

Volta

ge [p

u]

3

U_pcc

U_20kV

-75

-50

-25

0

25

50

75

0 0,5 1 1,5 2 2,5 3Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0 0,5 1 1,5 2 2,5 3Time [s]

Cur

rent

PC

C [k

A]

-2

-1,6

-1,2

-0,8

-0,4

0

0 0,5 1 1,5 2 2,5Time [s]

Slip

[%]

3

Figure 5-11 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu) – with reactive power compensation (Two-Mass Model)

110

0

0,2

0,4

0,6

0,8

1

0 0,5 1 1,5 2 2,5Time [s]

Volta

ge [p

u]

3

U_pcc

U_20kV

-75

-50

-25

0

25

50

75

0 0,5 1 1,5 2 2,5 3Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,1

0,2

0,3

0,4

0,5

0 0,5 1 1,5 2 2,5 3Time [s]

Cur

rent

PC

C [k

A]

-2,4

-2

-1,6

-1,2

-0,8

-0,4

0

0 0,5 1 1,5 2 2,5Time [s]

Slip

[%]

3

Figure 5-12 Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.8pu) – without reactive power compensation (Two-Mass Model)

111

6. Validation of Wind Turbine Models - DFIG - in Denmark

(Wind Turbines with Doubly Fed Induction Generator)

6.1. Introduction This chapter describes the validation process of the wind turbine (WT) with doubly fed

induction generator (DFIG) in the PowerFactory simulation environment. The validation

process was performed according to the procedure included in [15]. Also the available

information about the parameters of the test system as well as about the machine

parameters has been taken from [15]. Some additional assumptions had to be made due

to lack of information in [15], which are described in this report giving some respective

recommendations.

6.2. Test System The test system defined in [15] has been implemented into PowerFactory as shown in

Figure 6-1

.

Figure 6-1 Test System Implementation into PowerFactory The parameters given for the system components are summarized in Table 15.

112

System Component Parameter Value

SK’’ 3 GVA Equivalent Grid RN/XN 0.1 U1rat 110 kV U2rat 20 kV UK 11%


SN 50 MVA U1rat 20 kV U2rat 0.69 kV UK 6 %


SN 50 MVA

Table 15: Parameters of the Test System The description of the test system given in [15] does not define all parameters required in

the simulation software. Thus, the values of missing parameters have been assumed as

given in Table 16 . System Component Parameter Assumed Value

c – Factor 1.

Voltage Magnitude Setpoint 1 pu

Equivalent Grid

Voltage Angle Setpoint 0°







Phase Shift 0°

Cooper Losses – PCu 0 kW






Phase Shift 0°

Table 16: Assumed Values of Additional System Parameters

113

6.3. Wind Turbine Model with DFIG

6.3.1. General Structure of the Model To simulate the behavior of the wind turbine with DFIG the available PowerFactory model

described in [22] has been used. Some adaptation works of the control system has been

performed as will be discussed in section 6.3.4.

6.3.2. Electric Parameters The model of the wind turbine with DFIG has been partially parameterized in [15] as

shown in Table 17. Parameter Value Unit

Pn 1500 [kW] Un 690 [V] xn Un

2/Pn [Ω] ln xn/ω1 [H] p 2 ω1 2πfn [rad/s] fn 50 [Hz] r1 0.008 xn [Ω] r2’ 0.008 xn [Ω] l1s 0.080 ln [H] l2s’ 0.080 ln [H] lh 3.0 ln [H] l1 lh +l1s [H] l2 lh + l2s’ [H]

kP and kQ 3.56 e-5 [V/W] TiP and TiQ 0.0128 [s]

w21 2.7 (=winding ration rotor/ stator)

Table 17: Parameter of DFIG given in [15] The parameters of the generator equivalent circuit has been transformed into the per unit

system using a basis impedance as defined by Eq. (1).

Ω==== 3174.05.1

)69.0( 22

MVAkV

PUxZ

n

nnBASE (16)

The resulting per unit parameters used in PowerFactory are given in Table 3. Parameter Value

114


Table 18: Parameter of DFIG given in [15] transformed into p.u.-system as implemented into PowerFactory

The per unit parameters defined for the single wind turbine model has also been used for

the representation of the equivalent farm model. Moreover, in [15] the value of the wind

turbine´s active power Pn_FARM=30*Pn=45 MW is defined and this value has also been

assumed as rated power of the whole farm. For the given rated active power and

parameters of the machine equivalent circuit (as defined in Table 3) the apparent power of

the machine has been calculated by PowerFactory and is equal to 51.379 MW.

Since no other information about the test wind turbine was given in [15] some additional

assumptions had to be made. First, it has been assumed that the wind turbine can be

operated in the angular speed range corresponding to ± 20% of the generator slip.

Moreover, the parameters of the mechanical system as well as the structure of the

machine's control system have been assumed as will be discussed in section 6.3.3 and

section 6.3.4 of this report, respectively. It was also necessary to define in PowerFactory

the value of the rated slip ring voltage (with opened rotor circuit and standstill) which was

equal to UROTrat=1939 V.

In case of a wind turbine with DFIG the operational point, that is given by active and

reactive power as well as by the angular speed of the machine, can be varied in an

allowable range. In general, the active power output as well as the angular speed of the

turbine are determined by the control algorithms, which optimize the active power

production in partial load operation and limit the output power in full load operation. The

reactive power can be changed in order to support the grid voltage, however very often the

set point for reactive power is chosen to be equal to 0 Mvar.

In the simulations performed for this report the set point for active power has been

adapted to obtain about 45 MW on the 110 kV side of the test system, which corresponds

to the rated power of the farm. Moreover, it has been assumed that the generator operates

with -20% slip which corresponds to super-synchronous operation with an angular speed

115

equal to 1.2 pu. The reactive power of the generator has been set to 7 Mvar in order to

have the power factor equal to 1 in the point of common coupling on the 110 kV level while

the reactive power of the grid-side converter was equal 0 Mvar on the 0.69 kV node - so

that the reactive power demand of the connecting reactor (grid filter) has been

compensated.

6.3.3. Mechanical System Representation

The mechanical system of the wind turbine has been represented as two–mass model.

The block diagram of this system is shown in Figure 8 of [22]. The mechanical parameters

are summarized in Table 11. Parameter Value

Base Power – PBASE 5 MW Turbine Damping - DTUR 0 Nms/rad Inertia of the Turbine – JWTR 6.1·106 kg m2 Inertia of the Generator – JGEN 101.72 kg m2 Shaft Damping - DSHAF 1.4·106 Nms/rad Stiffness Constant – kms 8.3·107 Nm/rad Nominal Angular Speed of Turbine 18 rpm

Table 19: Mechanical Parameters of WT with DFIG (Rated Power 5MW) [23] Since the input and output signals of the mechanical system, as mechanical power of the

turbine and mechanical power transmitted through the shaft to the generator, are defined

in the model using the per unit system, the parameters defined in Table 19 have not been

modified, except the inertia of the equivalent generator, which was calculated according to

Eq. (17).

TURBGENEQV NJJ ⋅= (17)

Where NTURB=30 defines the number of individual wind turbines that are replaced by a

single equivalent farm model.

The resulting value of the equivalent generator inertia is equal to JEQV=101,72 kg m2 ⋅30

=3051.6 kg m2.

116

6.3.4. Control System of the Wind Turbine

6.3.4.1 Generator and Rotor Side Frequency Converter

The model of the wind turbine presented in [22] has been adapted in order to allow an

enhanced synchronization algorithm of the wind turbine generator with the grid after a fault

as discussed in [19] The goal of this control algorithm is to limit the excessive transient

current in the rotor of the machine during a fault and to protect the power electronic

converter connected to this rotor. For this purpose an additional resistance – crowbar – is

used to short-circuit the rotor while the commutation of the frequency converter is blocked.

If the crowbar is connected to the rotor circuit the doubly fed induction generator becomes

a regular induction generator with an additional resistance within the rotor circuit, which

influences the torque-slip characteristic of the machine. During this time period the

behavior of the machine – as reactive power exchange or electromechanical torque – can

not be controlled and therefore the machine will start to speed up and to absorb reactive

power from the grid, which causes decreasing of the voltage. Thus, the controllability of

the machine has to be restored as fast as possible, which happens by disconnecting the

crowbar and restarting the rotor side frequency converter. Depending on the wind turbine

manufacturer there are different schemes in which way the crowbar connection and

disconnection is proceeding. In this report the scheme described in [19] has been

followed.

The characteristic steps of this scheme are as follows:

1. insertion of the crowbar and blocking of the rotor side converter if the current limit in

the rotor circuit is exceeded,

2. disconnection of the crowbar and restart of the rotor side converter if the rotor current

is lower than the defined limit and the grid voltage is re-established,

3. synchronization of the wind turbine in order to reach the proper operational point

(regarding the active power – angular speed characteristic, as well as the reactive

power set-point).

117

The disconnection of the crowbar and restart of the converter can be a critical point, since

in this moment the rotor current might exceed the limit and the crowbar will be inserted

again. Therefore, at the moment of the crowbar disconnection the frequency converter has

to be synchronized to the respective point of operation. It means that the dq - components

of the voltage at the terminals of the rotor side converter must have a value which

corresponds to the voltage drop on the crowbar resistance. For this purpose the outputs of

the PI-controllers have to be adjusted to the desired value and at the same time a too fast

change of the controller outputs has to be avoided.

This has been realized in the extended wind turbine model as presented in Figure 6-5 to

Figure 6-7 in the appendix and will be discussed here. In the first step a non-windup PI

controller with reset possibility to a desired value has been implemented. The general

structure of this controller is shown in Figure 6-2. In normal state, when the trigger signal is

equal to zero, the switch is in position 1 and the controller operates on basis of the

measured value of the signal – XMEASURE and the reference value – XREF.

Figure 6-2 Block Scheme of a Non-Windup PI Controller with the Reset Function

In case of controller blocking the trigger signal is equal to 1 and the switch is in position 2.

In this operation mode the output of the PI - controller will be adjusted in order to have the

value equal to XRESET.

Moreover, in the control system of the rotor side converter the I – controller with reset

option has been used as well. The structure of this controller corresponds to that given in

Figure 6-2 with the proportional constant – K equal to zero.

118

The respective further gains and time constants for the used PI-controllers according to

Figure 6-6 and Figure 6-7 are defined in Table 20. In the used model the value of the KR-

parameter was set to 100.

Controller K T Rotor Active Power 0.1 0.01 Rotor Reactive Power

0.1 0.01

Rotor Current q-axis 0.44 0.01 Rotor Current d-axis 0.44 0.01 Grid Current d-axis 1. 0.015 Grid Current q-axis 3. 0.015 DC-Voltage 5 0.1 Both I-controllers in Rotor Power Controller

--- 1

Reset Gain KR 100 --- Table 20: Gains and time constants for used PI-controllers

Using the introduced PI and I – controllers the control system for active and reactive power

has been developed as shown in Figure 6-6. The inputs to this model are summarized in

Table 21 and outputs in Table 22, respectively. Input No.

Signal Description

0 bypass from protection module (equal to 1 if crowbar activated and immediately 0 if crowbar deactivated)

1 bypdelP from protection module (equal to 1 if crowbar activated and set to zero first if the active power controller synchronized)

2 Pref reference value of the active power from MPT (maximal power tracking) controller

3 P measured value of the active power in the PCC

4 Ifq measured value of the rotor current in the q-axis

5 Qref reference value of reactive power usually set constant to keep unitary power factor, but can be used to support the grid voltage

6 bypdelQ from protection module (equal to 1 if crowbar activated and set to zero first if the reactive power controller is synchronized)

7 Q measured value of the reactive power in the PCC

8 Ifd measured value of the rotor current in the d-axis

Table 21: Input Description of the Power Controller from Figure 6-6

119

Output

No. Signal Description

0 Pref_INT value of the output of the I-controller used for active power control

1 Ifq_ref reference value of the rotor current in the q-axis, which is used as input for the current controller

2 Ifd_ref reference value of the rotor current in the d-axis, which is used as input for the current controller

3 Qref feed-through of the Qref input for using in other blocks

4 Qref_INT value of the output of the I-controller used for reactive power control

Table 22: Output Description of the Power Controller from Figure 6-6

During normal operation, i.e. the crowbar is disconnected and therefore the signals

"bypass", "bypdelP" and "bypdelQ" are zero, the value of inputs “yi” and “yi1” (see Figure

6-6) for active and reactive power controllers “x1” and “x2” are calculated according to Eq.

(18) and Eq. (19), respectively.

P1Prefyi −= (18)

Q1Qrefyi1 −= (19)

In case the rotor current exceeds the defined limit the crowbar is activated and signals

"bypass", "bypdelP" and "bypdelQ" become 1. In this situation the states of all switches

change to the opposite state (i.e. to the internal switches of PI and I – controllers, as

shown in Figure 6-2 ) and the controller is in the emergency modus. In this modus all PI

and I – controllers are reset to an appropriate value in order to avoid the transients when

the crowbar will be deactivated. In the power controller the output signals – "Ifd_ref" and

"Ifq_ref", which are then used as input signals for the current controller, are set equal to

the measured rotor current components in the d and q axis, respectively. This guarantees,

that, after return to the normal operation mode from the emergency mode, the inputs for

the PI – controllers of the current controller will be equal to zero and no immediate step-

wise change of the modulation indexes "Pmd" and "Pmq", which correspond to the voltage

applied to the rotor, will be present. During the emergency operation the both modulation

indexes "Pmd" and "Pmq" are set to zero through the change of the switch position “S5”

and “S6”. At the same time the outputs of PI-controllers “x3” and “x4” are reset to the

values “Pmqref” and “Pmdref”, respectively. These values are calculated in the block

120

"current measurement" (see Figure 6-5 in appendix) using the measured values of the d

and q components of the rotor current as well as of the crowbar resistance and correspond

to the voltage drop on the crowbar resistance.

6.3.4.2 Grid-Side Frequency Converter

In general, the grid side frequency converter is responsible for controlling the voltage level

in the DC circuit as well as for the controlling of the reactive power exchange with the grid.

Therefore, it can be used as SVC unit for support of the grid voltage. The control system of

the grid side frequency converter used in the presented model is described in [22].

6.3.5. Protection System - Crowbar

6.3.4.3 General Remarks The protection system of wind turbines with doubly fed induction generators is more

sophisticated as in case of wind turbines with conventional asynchronous generators,

which usually are equipped with under- and over-voltage protection and trip immediately if

the lower or upper limit is exceeded.

The wind turbines with doubly fed induction generator, however, are obliged to stay

connected to the grid during a grid fault and to support the power system operation. Since

specific parts of these units, as frequency converters, are very sensitive to over-currents,

an appropriate protection scheme has to be applied. As already discussed in the previous

section the protection of the frequency converters is realized by the insertion of an

additional resistance (crowbar) into the rotor circuit and disconnection (blocking) of the

converter.

The challenge is the return to normal operation, because not proper disconnection of the

crowbar and activation of the converter can lead to excessive current transients and to a

re-activation of the crowbar. Therefore, the operation of the converter controllers as well

as the state of the crowbar is controlled by the protection system that generates the logic

control signals. The possible operational states of the wind turbine as well as of the

protection system are shown in Figure 6-3. Three states can be distinguished:

121

normal state,

- emergency state and

- resynchronization state.

Between these states there are transition conditions T1 – T3. In order to change from one

state to another the transition conditions between the considered states have to be

fulfilled.

The transition condition T1 are fulfilled if the measured value of the rotor current exceeds

the defined limit, the transition conditions T2 are fulfilled if the grid voltage is reestablished

and the transition conditions T3 are fulfilled if the controllers of the rotor side converter are

reset to the desired values and the unit is synchronized to the existing conditions (as wind

condition, grid voltage conditions). In this model the limit of the rotor current was set to 1,5

times the base rotor current and its value is given by Eq. (20).

kAII ROTbaseROT 45.325.1max =⋅= (20)

The rotor base current is defined as given by Eq. (21).

AV

MVAU

SIROTrat

ratROTbase 27.21635

31939379.512

32 =

⋅=

⋅=

(21)

In normal state the crowbar is disconnected and the controllers are following the defined

reference values of the active and reactive power. In this state the logic control signals –

"bypass", "bypdelP" and "bypdelQ" equal to zero. If the wind turbine is in emergency state

the crowbar is active and the controllers are reset to the appropriate value while the rotor-

side-converter is blocked. In this state all control signals of the protection system equal to

one.

If the voltage is reestablished the crowbar is disconnected and the rotor side converter

supplies the rotor with the voltage corresponding to the voltage drop at the crowbar

resistance. The unit is now in resynchronization state. In this state not all controllers are in

normal operation since only the "bypass" – signal is equal to zero, since the crowbar has

been disconnected. The other signals "bypdelP" and "bypdelQ" are meanwhile equal to

one until the outputs of both I – controllers are different than present reference values for

122

active and reactive power. If the reference value of active power "Pref" is equal to the

output from the I-controller the signal "bypdelP" is set to zero and similar in case of the

signal "bypdelQ" but here the "Qref" is observed.

Figure 6-3 Operational States of the Protection System

123

6.3.4.4 Influence of Crowbar Resistance Size on the Machine Behaviour

An important aspect is also the value of the crowbar resistance since it has influence on

damping of current transients in the rotor circuit. Since there was no information in [15]

about the size of the crowbar resistance its value was chosen for the validation at hand

according to the recommendations given in [19]. These recommendations say that an

optimal value for the crowbar resistance is equal to 20 times the resistance of the rotor

circuit. Therefore, for the given parameters of the generator (see Table 2 and Table 3) the

crowbar resistance used for the analysis at hand is equal to

RCROW=20 x 0.008pu = 0.16pu. In the appendix the influence of the crowbar resistance on

the wind turbine behavior during a 3-phase fault is shown (see Figure 5-10).

6.4. Analysis of the Test System with DFIG Wind Turbine In the test system a three-phase short circuit at the 110 kV bus has been simulated. The

duration of the fault was set to 150 ms. In the first case the reactance of the short circuit

was adjusted in order to obtain the desired level of the residual voltage at the 110 kV node

equal to 0.2 p.u., which was the requirement from [15]. For this purpose the reactance

XFAULT has been set to XFAULT = 1 Ω.

In the second case the short circuit reactance was equal to XFAULT = 15.5 Ω, which results

in residual voltage equal to 0.8pu in the point of common coupling. The simulation results

for both cases are shown in and Figure 6-9 of the appendix, respectively.

Both figures show significant differences in comparison to the results presented in [15].

The main reasons for these differences results from the fact, that the test wind turbine has

not been completely specified in [15], i.e. control system structure, mechanical system,

protection system and their parameters. Thus, different behaviour is possible.

The behavior of the wind turbine with DFIG depends mainly on the control algorithm. But

also such parameters like crowbar resistance have a significant influence on the dynamic

behavior, especially after fault clearing, see Figure 5-10.

124

In the next chapter the parameter settings developed in this report will be summarized.

6.5. Conclusions and Recommendations This report discusses an exemplary modeling approach of the wind turbine with DFIG that

has been developed on basis of [15] and [22].

The analysis described in this paper shows that the detailed definition of all parameters of

the test system is crucial for the simulation results using different software. As already

discussed in [24] the main reason for the differences between simulation results are not

completely defined parameters of the model components and its structure in [15] thus,

some recommendations have been made in this report.

Therefore, in order to obtain similar simulation results a more detailed description of the

model is required, or the model in [15] should be adapted to the recommendations of this

report.

This concerns, first of all, the structure and the parameters of the control system (rotor-

side converter, grid-side converter, crowbar initialization and disconnection conditions and

corresponding logic), but also the uniform representation and the parameters of the

mechanical system.

Moreover, the structure and the parameters of the DC-circuit (voltage level, size of

capacitor) as well as the size of the grid filter used for interconnection of the grid-side

converter have to adapted respectively.

125

6.5.1. Recommendations for Simulation Settings The following settings are recommended:

Parameter Value Parameter Value Sn 51.379 MVA CDC 144412.8 µF Un 0.69 kV SGridFilter 20MVA n0 1500 rpm UN_GridFilter 0.69 kV Rs 0.008 p.u. uK_GridFilter 31.8% Xs 0.080 p.u. Copper Losses Grid Filter 21 kW Xm 3.0 p.u. Rr’ 0.008 p.u. Xr’ 0.080 p.u.

Slip Range ± 20% UROTrat 1939V IROTmax 32.45kA *) RCROW 0.16pu

PSET-POINT_PCC 45 MW QSET-POINT_PCC 0 Mvar

PSET-POINT_GEN_STATOR 38 MW QSET-POINT_GEN_STATOR 7 Mvar

Table 23: Summary of electrical parameter settings developed in this chapter. *) Such a high value results from the assumption of the basis current for the rotor in PowerFactory

Parameter Value Base Power – PBASE 5 MW Turbine Damping - DTUR 0 Nms/rad Inertia of the Turbine – JWTR 6.1·106 kg m2 Inertia of the Generator – JGEN 101.72 kg m2 Shaft Damping - DSHAF 1.4·106 Nms/rad Stiffness Constant – kms 8.3·107 Nm/rad Nominal Angular Speed of Turbine 18 rpm Inertia of Equivalent Generator - JEQV=30xJGEN

3051.6 kg m2

Table 24: Summary of mechanical parameter settings developed in this chapter. Fault reactance in case of URES=0.2pu is equal to 1Ω and in case of URES=0.8pu is equal to

15.5Ω.

126

6.6. Appendix DFIG - Denmark

6.6.1. Simulation Results with PowerFactory – Load Flow

Figure 6-4 Load Flow Results in the Test System with DFIG (P=45MW; Q=0Mvar, slip=-

20%)

127

6.6.2. Block Diagrams of the Enhanced Wind Turbine Model

Figure 6-5 Block Diagram of the Model Representing the Generator and Rotor-Side

Converter

128

Figure 6-6 Block Diagram of the Model Representing the Power Controller

129

Figure 6-7 Block Diagram of the Model Representing the Current Controller

130

6.6.3. Simulation Results for Dynamic Analysis

0

0,5

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cro

wba

r

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Volta

ge [p

u]

U_pcc

U_20kV

-100

-50

0

50

100

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,2

0,4

0,6

0,8

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

131

-50

-25

0

25

50

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Rot

or C

urre

nt [k

A]

abc

-3

-2

-1

0

1

2

3

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [p

u]

RealImag

1

1,1

1,2

1,3

1,4

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Ang

ular

Spe

ed [p

u]

Figure 6-8 Simulation Results of 3-phase fault at 110 kV node

(tFAULT=150ms, UFAULT=0.2pu, Two-Mass Model, RCROWBAR=0.16pu)

132

0

0,5

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cro

wba

r

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Volta

ge [p

u]

U_pcc

U_20kV

-100

-50

0

50

100

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

.- &

Rea

ct. P

ower

P_pcc [MW]

Q_pcc [Mvar]

0

0,2

0,4

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [k

A]

-50

-25

0

25

50

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Rot

or C

urre

nt [k

A]

abc

133

-3

-2

-1

0

1

2

3

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [p

u]

RealImag

1

1,1

1,2

1,3

1,4

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Ang

ular

Spe

ed [p

u]

Figure 6-9 Simulation Results of 3-phase fault at 110 kV node

(tFAULT=150ms, UFAULT=0.8pu, Two-Mass Model, RCROWBAR=0.16pu)

134

6.6.4. Influence of Crowbar Resistance on Behavior of the Machine

0

0,5

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cro

wba

r

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Volta

ge P

CC

[pu

]

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Volta

ge 2

0kV

[pu]

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

-75

0

75

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Act

. Pow

er P

CC

[MW

]

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

135

-150

-75

0

75

150

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Rea

ct. P

ower

PC

C [M

var]

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

0

0,2

0,4

0,6

0,8

1

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Cur

rent

PC

C [

kA] Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

1,1

1,15

1,2

1,25

1,3

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1Time [s]

Ang

ular

Spe

ed [

pu]

Rcrow =0.16

Rcrow =0.016

Rcrow =0.36

Figure 6-10 Influence of Crowbar Resistance on Simulation Results of 3-phase fault at 110 kV node (tFAULT=150ms, UFAULT=0.2pu, Two-Mass Model, Different Crowbar Resistances)

136

7. References

[1] CIGRE Technical Brochure on Modeling and Dynamic Behavior of Wind Generation as it

relates to Power System Control and Dynamic Performance; Working Group 601 of Study

Committee C4, Final Report January 2007.

[2] Nicholas W. Miller, William W. Price, Juan J. Sanchez-Gasca: Stability Modeling of Vestas

V80 Wind Turbine-Generator (Version 2.0); GE Energy, March 2003.

[3] Rudion, K.; Ruhle, O., Styczynski: Simulation of Large Wind Farms using Coherency

Approach; Eurosim 2007, Sept. Ljubljana, Slovenia.

[4] Soens, J.: Impact of Wind Energy in A Future Power Grid; PhD Thesis, Katholieke

Universiteit Leuven, Belgium, 2005.

[5] Petersson, A.: Analysis, Modeling and Control of Doubly-Fed Induction Generators for Wind

Turbines. Ph Thesis, Chalmers University of Technology, Göteborg, Sweden 2005.

[6] G. Duschl, H.-D. Pannhorst, O. Ruhle - SIEMENS AG, Erlangen Dynamic simulation of

DFIGs for wind power plants using NETOMAC, Workshop for Wind power integration,

Glasgow, 2005

[7] Erlich, I.; Winter, W.; Dittrich, A.: Advanced Grid Requirements for the Integration of Wind

Turbines into the German Transmission System, IEEE PES General Meeting, Montreal 2006.

[8] “Procedimiento de Operación “P.O.12.3 – Requisitos de respuesta frente a huecos de

tensión de las instalaciones de producción de régimen especial”. Red Eléctrica de España.

B.O.E. . October 2006. http://www.ree.es/operacion/procedimientos_operacion.asp

[9] “Procedimiento de verificación, validación y certificación de los requisitos del Procedimiento

de Operación 12.3 sobre la respuesta de las instalaciones eólicas ante huecos de tensión”.

January 2007; Available on: http://www.aeeolica.org/

[10] “IEC 60909. Short-circuit currents in three-phase a.c. systems” July 2001

[11] Akhmatov V.: Analysis of Dynamic Behavior of Electric Power System with Large Amount of

Wind Power, PhD. Thesis available on http://www.oersted.dtu.dk/eltek/res/phd/00-05/va-thesis.pdf

[12] Price B.: Wind Turbine Generators Technology -Performance –Issues Modeling, 2004 PSLF

User’s meeting, San Diego California

[13] Rodríguez-Bobada, F., Ledesma Larrea P., Martínez, S., Coronado L. and Prieto E.

Simplified Wind Generator Model for Transmission System Operator Planning Studies; 7th

International Energynautics Workshop on Large Scale Wind Integration. Madrid May 2008

137

http://www.oersted.dtu.dk/eltek/res/phd/00-05/va-thesis.pdf

http://www.oersted.dtu.dk/eltek/res/phd/00-05/va-thesis.pdf

[14] Toal, J.; Urdal, H.; Horne, J.; Coughlan, Y.; Wiechowski, W. et al: Building Confidence in

Computer Models of Wind Turbine Generators from a Utility Perspective; Nordic Wind Power

Conference, 1-2 November 2007, Roskilde, Denmark.

[15] Winter, W.: Validation of Wind Turbines – Draft. 2007-11-12.

[16] Sorensen, P.; Hansen, A. D.; Christensen, P.: Simulation and Verification of Transient Events

in Large Wind Power Installations. RISO-R-1331(EN), October 2003.

[17] Kundur, P.: Power System Stability and Control. McGraw-Hill, Inc. 1994. ISBN 0-07-035958-

X.

[18] Miller, N. W.; Price, W. W.; Sanchez-Gasca, J. J.: Modeling of GE Wind Turbine-Generators

for Grid Studies (Version 3.4b), March 4, 2005.

[19] Akhmatov, V.: Induction Generators for Wind Power. Multi-Science Publishing Company, Ltd.

2005. ISBN 0 906522 40 4.

[20] Regulation TF 3.2.5: Wind Turbines Connected to Grids with Voltages above 100 kV,

Technical regulation for the properties and the regulation of wind turbines, December 2004,

available at www.energinet.dk

[21] Regulation TF 3.2.6: Wind Turbines Connected to Grids with Voltages below 100kV,

Technical regulations for the properties and the control of wind turbines, May 2004m

available at www.energinet.dk.

[22] DigSilent GmbH: Technical Documentation for Dynamic Modelling of Doubly-Fed Induction

Machine Wind-Generators. doc.TechRef, 14 August 2003.

available at www.digsilent.de/images/company/news/dfigref1.pdf.

[23] Model of Wind Turbine with DFIG available in PowerFactory and specified in reference [22].

[24] Energinet.dk Report for Validation of Wind Turbine with Squirel Cage Induction Generator.

December 2007. (= chapter 5 of this report)

138

http://www.energinet.dk/

http://www.energinet.dk/

ewis - european wind integration study - contract number:...

Documents