GE Wind Turbine

Modeling of GE Wind
Turbine-Generators for
Grid Studies
Prepared by:
Kara Clark
Nicholas W. Miller
Juan J. Sanchez-Gasca
Version 4.5
April 16, 2010
General Electric International, Inc.

One River Road
Schenectady, NY 12345
USA
Legal Notice
This report was prepared by General Electric International, Inc. (GEII) as an account
of work sponsored by GE’s Wind Energy business. Neither Wind Energy nor GEII, nor
any person acting on behalf of either:
1. Makes any warranty or representation, expressed or implied, with respect to the
use of any information contained in this report, or that the use of any information,
apparatus, method, or process disclosed in the report may not infringe privately
owned rights.
2. Assumes any liabilities with respect to the use of or for damage resulting from the
use of any information, apparatus, method, or process disclosed in this report.
ii
Foreword
This document was prepared by GE Energy in Schenectady, NY. Technical and

commercial questions and any correspondence concerning this document should be
referred to:
Kara Clark
GE Energy
Building 53-302A, One River Road
Schenectady, New York 12345
Phone: (518) 385-5395
E-mail: kara.clark@ge.com
iii
Summary of Changes in Version 4.5
• Updated current limits in LVPL and both electrical control models
• Updated DFAG and full converter reactive power control figures, added new data to table
• Updated DFAG and full converter electrical control figures, added text describing input test signal
• Updated WindFREE description in DFAG and full converter sections
• Updated APC frequency response figure and high frequency tripping discussion
• Updated all benchmark simulations
• Addition of text and data recommendations for modeling 1.6 MW DFAG machine
• Addition of Q droop function to full converter control model
• Addition of Xc (compensating impedance) to full converter control model
• Clarification on reactive capability; voltage regulation recommended settings
• Clarification of turbine control modeling recommendations
• Clarification of fault ride-through modeling recommendations
• Clarification on 2.5 MW full converter mechanical model
• Addition of Q droop function to electrical control model for DFAG
• Addition of WindINERTIATM function to turbine control model
• Revised generator/converter model source current calculation, interface with network solution
• Moved low voltage power logic from electrical control to generator/converter model
• Updated text, figures, and data as appropriate to reflect the above changes
• Reran all DFAG and full converter test simulations for Sections 6, 7 and 8.
• Simplified PLL in generator model, added low voltage active current regulation and reactive current
limits to prevent excessive voltage to generator model
• Modified converter current limiter in control model, and XIQmax/XIQmin limits
• Removed LVRT voltage support function from control model
• Updated official names – i.e., WindVAR became WindCONTROL
• Updated example simulations in Sections 6 and 7. Added comparison to Windtrap.
• Added full converter model block diagrams, discussion, simulation results, etc
• Moved generator voltage protection discussion to generator model section
• Folded “Other Technical Issues” section into other parts of the report
• Modified PLL in generator model.
• Modified turbine control to include Active Power Control.
• Added figure showing details of improved pitch compensation in turbine model
Summary of Changes in Versions 3.4, 3.4a, 3.4b
• Corrections regarding Kqi and Kvi in Section 4.2.2 and Table 4-5 (version 3.4b)
• Adjustment of values of per unit Qmax and unit transformer MVA and impedance for 1.5 MW WTG
(version 3.4a)
• Phase-locked loop added to converter/generator model (version 3.4)
• WindVar emulator model changed to be closer to real control logic (version 3.4)
• Phase angle regulation option added to electrical control (version 3.4)
• Generator protection tripping model description updated and moved to Section 4.4 (version 3.4)
• Recommended parameter values updated based on latest validation testing (version 3.4)
iv
Revision History
Version 1.0 – December 4, 2002 Version 4.0 – September 22, 2006
Version 2.0 – March 14, 2003 Version 4.1 – March 2008
Version 3.0 – October 27, 2003 Version 4.2 – June 24, 2008
Version 3.1 – December 22, 2003 Version 4.3 – April 8, 2009
Version 3.2 – May 4, 2004 Version 4.4 – September 9, 2009
Version 3.3 – June 7, 2004 Version 4.5 – April 16, 2010
Version 3.4 – December 21, 2004
v
Table of Contents
1 INTRODUCTION.....................................................................................................................1.1
2 MODEL OVERVIEW AND PHILOSOPHY .............................................................................2.1

2.1 DOUBLY-FED ASYNCHRONOUS GENERATOR (DFAG) FUNDAMENTALS ................................2.1
2.2 FULL CONVERTER WTG FUNDAMENTALS ..........................................................................2.3
2.3 OVERALL MODEL STRUCTURE ...........................................................................................2.4
3 ANALYTICAL APPROACH....................................................................................................3.1
3.1 LOAD FLOW MODEL ..........................................................................................................3.1
3.2 INITIAL CONDITIONS FOR DYNAMIC SIMULATION ..................................................................3.3
4 DOUBLY FED ASYNCHRONOUS GENERATOR (DFAG) DYNAMIC MODELS ................4.1
4.1 GENERATOR/CONVERTER MODEL .....................................................................................4.1
4.1.1 Fault Ride Through ....................................................................................................4.2
4.2 CONVERTER CONTROL MODEL ..........................................................................................4.6
4.2.1 Reactive Power Control .............................................................................................4.6
4.2.2 Electrical Control ........................................................................................................4.9
4.2.3 Control Strategies ....................................................................................................4.10
4.2.4 Open Loop Control Logic .........................................................................................4.11
4.3 WIND TURBINE AND TURBINE CONTROL MODEL ...............................................................4.13
4.3.1 Turbine Control Model..............................................................................................4.15
4.3.2 Rotor Mechanical Model ..........................................................................................4.18
4.3.3 Wind Power Model ...................................................................................................4.20
4.3.4 Active Power Control Model & Rate Limit Function .................................................4.22
4.3.5 WindINERTIA Model ................................................................................................4.25
5 FULL CONVERTER WTG DYNAMIC MODELS....................................................................5.1
5.1 GENERATOR/CONVERTER MODEL .....................................................................................5.1
5.2 ELECTRICAL (CONVERTER) CONTROL MODEL ....................................................................5.2
5.3 WIND TURBINE AND TURBINE CONTROL MODEL .................................................................5.6
6 DFAG WTG BENCHMARK SIMULATIONS ..........................................................................6.1
6.1 TEST SYSTEM ...................................................................................................................6.1
6.2 DFAG WTG BENCHMARK SIMULATION RESULTS ...............................................................6.2
6.2.1 Fault Response with WindCONTROL Emulator ........................................................6.3
6.2.2 Q Droop Function Performance .................................................................................6.6
6.2.3 Active Power Control and Power Response Rate Limit Performance .......................6.8
6.2.4 Wind Speed Profile ..................................................................................................6.10
6.2.5 Zero Power (WindFREE) Operation ........................................................................6.10
6.2.6 WindINERTIA Control Performance ........................................................................6.13
6.3 DFAG WTG BENCHMARK SIMULATION DYNAMIC DATA....................................................6.15
7 FULL CONVERTER WTG BENCHMARK SIMULATIONS ...................................................7.1
7.1 TEST SYSTEM ...................................................................................................................7.1
7.2 FULL CONVERTER WTG BENCHMARK SIMULATION RESULTS ..............................................7.1
7.2.1 Fault Response with WindCONTROL Emulator ........................................................7.2
7.2.2 Zero-Power Operation................................................................................................7.5
7.2.3 Converter Current Limit Performance ........................................................................7.9
7.2.4 Low Voltage Power Logic Performance...................................................................7.11
7.2.5 Dynamic Braking Resistor Performance ..................................................................7.11
7.3 FULL CONVERTER WTG BENCHMARK SIMULATION DYNAMIC DATA ...................................7.13
8 PSLF MODEL VALIDATION ..................................................................................................8.1
vi
8.1 1.5 MW DFAG COMPARISON TO WINDTRAP .....................................................................8.1
8.2 2.5 MW FULL CONVERTER COMPARISON TO WINDTRAP ....................................................8.4
8.3 WINDINERTIA MODEL COMPARISON TO FIELD MEASUREMENTS ........................................8.6
9 CONCLUSIONS......................................................................................................................9.1
vii
Table of Figures
Figure 2-1. GE Doubly Fed Asynchronous WTG Major Components. ................................2.1
Figure 2-2. GE Full Converter WTG Major Components. ..................................................2.3
Figure 2-3. GE WTG Dynamic Model Connectivity. ...........................................................2.5
Figure 3-1. Simplified Wind Plant Power Flow Model. .......................................................3.2
Figure 4-1. DFAG Generator/Converter Model. .................................................................4.2
Figure 4-2 GE 1.5 and 1.6 MW ZVRT Model Settings and Equipment Specification. ........4.4
Figure 4-3 GE 1.5 and 1.6 MW LVRT II Model Settings and Equipment Specification. .....4.5
Figure 4-4. Overall DFAG Reactive Power and Electrical Control Model. ..........................4.6
Figure 4-5. Reactive Power Control Model. ........................................................................4.7
Figure 4-6. Q Droop Function Model. .................................................................................4.9
Figure 4-7. DFAG Electrical Control Model. ....................................................................4.10
Figure 4-8. Wind Turbine Model Block Diagram. .............................................................4.15
Figure 4-9. Pitch Control and Pitch Compensation Block Diagram...................................4.16
Figure 4-10. Two-Mass Rotor Model. ................................................................................4.19
Figure 4-11. Wind Power Cp Curves. ................................................................................4.21
Figure 4-12. Active Power Control Emulator. ...................................................................4.23
Figure 4-13. Example Frequency Response Curve. ...........................................................4.24
Figure 4-14. Simplified WindINERTIA Control Model.....................................................4.27
Figure 5-1. Full Converter WTG Generator/Converter Model. ..........................................5.1
Figure 5-2. Full Converter WTG Electrical Control Model. ................................................5.4
Figure 5-3. Converter Current Limit Model. ......................................................................5.5
Figure 6-1. Primary DFAG Test System..............................................................................6.1
Figure 6-2. Series of Bus Faults with Various Fault Impedances. ........................................6.4
Figure 6-3. 3-phase Fault to Ground, Cleared by Tripping 230 kV Line. ............................6.5
Figure 6-4. Q Droop Response to Capacitor Switching........................................................6.7
Figure 6-5. Active Power Control Response to Loss of Load. ..............................................6.9
Figure 6-6. Response to Wind Speed Profile without Zero Power Operation. ...................6.11
Figure 6-7. Response to Wind Speed Profile with Zero Power Operation..........................6.12
Figure 6-8. Response to Frequency Signal with WindINERTIA. .......................................6.14
Figure 7-1. Series of Bus Faults with Various Fault Impedances. ........................................7.3
Figure 7-2. 3-phase Fault to Ground, Cleared by Tripping 230 kV Line. ............................7.4
Figure 7-3. Increasing Wind Speed Results in Zero-Power Operation. ...............................7.6
Figure 7-4. Response to Wind Speed Profile with Zero Power Operation............................7.7
Figure 7-5. Voltage Regulation in Continuous Zero-Power Operation. ...............................7.8
Figure 7-6. Step Reduction in Converter Current Limit with P Priority. ..........................7.10
Figure 7-7. Low Voltage Power Logic Response to Fault on POI Bus................................7.12
Figure 8-1. Test System for PSLF and Windtrap Comparison. ...........................................8.1
Figure 8-2. 1.5 MW DFAG PSLF and Windtrap Terminal Voltage Response. ....................8.2
Figure 8-3. 1.5 MW DFAG PSLF and Windtrap Real Power Response. .............................8.3
Figure 8-4. 1.5 MW DFAG PSLF and Windtrap Reactive Power Response. .......................8.3
Figure 8-5. 2.5 MW Full Converter PSLF and Windtrap Terminal Voltage Response. .......8.4
Figure 8-6. 2.5 MW Full Converter PSLF and Windtrap Real Power Response. ................8.5
Figure 8-7. 2.5 MW Full Converter PSLF and Windtrap Reactive Power Response. ..........8.5
Figure 8-8. Frequency Test Signal. ......................................................................................8.6
Figure 8-9. WindINERTIA Model Performance and Field Measurements. .........................8.7
viii
Table of Tables
Table 3-1. Individual WTG Power Flow Data. ....................................................................3.2
Table 4-1. Typical Fault Ride Through Voltage Thresholds and Durations. .......................4.4
Table 4-2. DFAG Generator/Converter Parameters............................................................4.5
Table 4-3. Reactive Power Control Parameters (on Generator MVA Base). .......................4.8
Table 4-4. Q Droop Function Parameters............................................................................4.9
Table 4-5. DFAG Electrical Control Parameters. ..............................................................4.10
Table 4-6. Open Loop Reactive Power Control Logic........................................................4.12
Table 4-7. Open Loop Reactive Power Control Parameters. .............................................4.12
Table 4-8. DFAG WTG Turbine Control Parameters (on Turbine MW Base). .................4.18
Table 4-9. DFAG WTG Rotor Mechanical Model Parameters (on Turbine MW Base). ....4.19
Table 4-10. DFAG WTG Wind Power Coefficients. ..........................................................4.20
Table 4-11. Cp Coefficients αi,j ..........................................................................................4.22
Table 4-12. Active Power Control and Rate Limit Function Parameters. ..........................4.25
Table 4-13. WindINERTIA Control Parameters. ..............................................................4.27
Table 5-1. Full Converter Generator/Converter Parameters. ..............................................5.1
Table 5-2. Full Converter WTG Electrical Model. ..............................................................5.5
Table 5-3. Full Converter WTG WindCONTROL Emulator Parameters. ..........................5.5
Table 5-4. Full Converter WTG Turbine Control Parameters (on Turbine MW Base). ......5.6
Table 5-5. Full Converter WTG Rotor Model Parameters (on Turbine MW Base). ............5.7
Table 5-6. Full Converter WTG Wind Power Coefficients. .................................................5.7
Table 6-1. DFAG Generator Model (gewtg) Data for Simulations. ....................................6.15
Table 6-2. DFAG Electrical Control Model (exwtge) Data for Simulations. ......................6.16
Table 6-3. DFAG Turbine Control Model (wndtge) Data for Simulations. ........................6.17
Table 7-1. Full Converter WTG Generator Model Data for Simulations. .........................7.13
Table 7-2. Full Converter WTG Electrical Control Model Data for Simulations...............7.14
Table 7-3. Full Converter WTG Turbine Control Model Data for Simulations. ................7.15
ix
1 Introduction
GE Energy has an ongoing effort dedicated to the development of models of GE
wind turbine generators (WTG) suitable for use in system impact studies. This report
documents the present recommendations for dynamic modeling of wind plants with
either doubly fed asynchronous WTGs (GE 1.5, 1.6 and 3.6 MW) or WTGs with a full
converter (GE Multi-Megawatt 2.5 MW). This report includes recommended model
structure and data, as well as the assumptions, capabilities and limitations of the
resulting model.
The model provided is as detailed as is appropriate for bulk power system studies.
It is valuable to put the model limitations in the context of what analysis is required.
Most important, this model is for positive sequence phasor time-domain simulations –
e.g. PSLF or PSS/e. Second, this assumes that the analysis is mainly focused on how
the WTGs react to grid disturbances, e.g. faults, on the transmission system. Third, the
model provides for calculation of the effect of wind speed fluctuation on the electrical
output of the WTG. Details of the device dynamics have been substantially simplified.
Specifically, the very fast dynamics associated with the control of the generator
converter have been modeled as algebraic (i.e. instantaneous) approximations of their
response. Representation of the turbine mechanical controls has been simplified as
well. The model is not intended for use in short circuit studies or electromagnetic
transient studies.
The models, as implemented in GE’s PSLF dynamic simulation program, have been
validated against more detailed design models. Selected validation comparisons are
documented in Section 8. Additional PSLF simulation examples showing doubly-fed
asynchronous generator (DFAG) and full converter WTG performance are included in
Sections 6 and 7.
These models were developed specifically for the latest GE WTGs. The model is
applicable, with care, to other recent vintage GE WTGs and other WTGs, as long as the
basic principles of power conversion and control are the same. However, this model is
not designed for, nor intended to be used as, a general purpose WTG model. There are
substantial variations between models and manufacturers.
Please contact GE for advice on modeling any GE WTGs not currently included in
this document. In addition, updated versions of this document may be available on the
GE Wind and PSLF software websites.
GE Energy 1.1 GE WTG Modeling-v4.5.doc, 4/16/10

2 Model Overview and Philosophy
2.1 Doubly-Fed Asynchronous Generator (DFAG) Fundamentals
A simple schematic of an individual GE 1.5, 1.6 or 3.6 MW wind turbine-generator
(WTG) is shown in Figure 2-1. The GE WTG generators are unusual from a system
simulation perspective. Physically, the machine is a relatively conventional wound
rotor induction (WRI) machine. However, the key distinction is that this machine is
equipped with a solid-state voltage-source converter AC excitation system. The AC
excitation is supplied through an ac-dc-ac converter. For the GE 3.6 MW machine, the
converter will be connected as shown or to a third winding on the main unit step-up
transformer. For the GE 1.5 and 1.6 MW machines, it is connected directly at the stator
winding voltage. Machines of this structure are termed ‘doubly-fed asynchronous
generators’ (DFAG), and have significantly different dynamic behavior than either
conventional synchronous or induction machines.
Modeling of any of the GE machines with conventional dynamic models for either
synchronous or induction machines will not give correct results.
P net
Q net
3 φ AC Windings
fnet Collector
System
P stator (e.g.
34.5kV
bus)
frotor
P rotor
P rotor P conv
F rotor F network
Wind Turbine
Wound-Rotor Converter
Induction Generator
Figure 2-1. GE Doubly Fed Asynchronous WTG Major Components.
The fundamental frequency electrical dynamic performance of the DFAG is
completely dominated by the converter. Conventional aspects of generator
performance related to internal angle, excitation voltage, and synchronism are largely
irrelevant. In practice, the electrical behavior of the generator and converter is that of
a current-regulated voltage-source inverter. Like other voltage-source inverters (e.g., a
BESS or a STATCOM), the WTG converter synthesizes an internal voltage behind a

transformer reactance, which results in the desired active and reactive current being
delivered to the device terminals.
In the case of the doubly-fed machines, the machine rotor and stator windings are
primary and secondary windings of the transformer. The rotation of the machine
means that the ac frequency on the rotor winding corresponds to the difference
between the stator frequency (50 or 60 Hz) and the rotor speed. This is the slip
frequency of the machine. In the vicinity of rated power, the DFAG machines will
normally operate at 120% speed, or -20% slip. Control of the excitation frequency
allows the rotor speed to be controlled over a wide range, ±30%. The rotation also
means that the active power is divided between the stator and rotor circuits, roughly in
proportion to the slip frequency. For rotor speeds above synchronous, the rotor active
power is injected into the network through the converter. The active power on the
rotor is converted to terminal frequency (50 or 60 Hz), as shown in Figure 2-1.
In addition to controlling the rotor speed, the reactive power output of the
generator can be controlled by varying the magnitude of the rotor currents. This gives
the doubly-fed machine the voltage regulation capability of a synchronous generator
but with greater speed of response.
For all GE machines, the control of active and reactive power is handled by fast,
high bandwidth regulators within the converter controls. The time responses of the
converter regulators are sub-cycle, and as such can be greatly simplified for simulation
of bulk power system dynamic performance.
Broadly stated, the objectives of the turbine control are to maximize power
production while maintaining the desired rotor speed and avoiding equipment
overloads. There are two controls (actuators) available to achieve these objectives:
blade pitch control and torque order to the electrical controls (the converter). The
turbine model includes all of the relevant mechanical states and the speed controls.
The implementation of the turbine model, while relatively complex, is still considerably
simpler than the actual equipment. Losses are not considered throughout the model,
since “fuel” efficiency is not presently a consideration. These simplifications are
examined in the detailed model discussion in Section 4.
The model presented in Section 4 describes the relevant dynamics of a single
doubly fed GE WTG. However, the primary objective of this model is to allow for
analysis of the performance of groups of WTGs and how they interact with the bulk
power system. Wind plants with GE WTGs normally include a wind farm management
system (WindCONTROL). Two components of this system are currently incorporated -
the Volt-Ampere-Reactive or Var control system and the Active Power Control (APC).
The Var control interacts with the individual WTGs through the electrical controls, the
APC is incorporated in the turbine model. Representation of all the individual machines
in a large wind plant is inappropriate for most grid stability studies. Hence, there is
provision within the model structure to allow a single equivalent WTG machine model
to represent multiple WTGs. The model implementation allows the user access to
parameters that might reasonably be customized to meet the particular requirements
of a system application. These parameters are discussed in more detail in Section 4.

2.2 Full Converter WTG Fundamentals
The GE Multi-Megawatt WTG product line uses full conversion technology. Unlike
the generators for the doubly fed asynchronous WTGs, the full converter machine is a
relatively conventional permanent magnet synchronous generator. The generator is
connected to the power grid through a full converter. This configuration decouples the
generator speed from the power system frequency and allows for a wide range of
variable speed operation. Figure 2-2 shows the configuration of the full converter
WTG.
Converter
Pnet= Pstator
f stator Q net
P stator
3φ AC Winding Q stator fnet
Wind Turbine f stator

net
Collector System
P stator
(e.g. 34.5kV bus)
Permanent
frotor
Magnet
P rotor
Rotor
Figure 2-2. GE Full Converter WTG Major Components.
Like the GE DFAG machines, the fundamental frequency electrical dynamic

performance of the GE Multi-Megawatt WTG is completely dominated by the converter.
For the full converter machines, the line-side of the converter corresponds to the
WTG terminals. The electrical behavior on the variable frequency machine side of the
converter is of no interest to the AC system. Further, operation (i.e., rotation) of the
turbine is not required for the converter to continue reactive operation on the line-side.
In the vicinity of rated power, the GE full converter machines will normally operate at a
speed selected to give optimum turbine performance. Control of the frequency
converter allows the rotor speed to be completely decoupled from the grid frequency,
and to be controlled over a wide range.
Similar to the DFAG WTGs, the control of active and reactive power is handled by
fast, high bandwidth regulators within the converter controls, and can be greatly
simplified for simulation of bulk power system dynamic performance.
The turbine control is also similar to that used for a DFAG WTG.
The model presented in Section 5 describes the relevant dynamics of a single full
converter WTG. As noted above in the DFAG discussion, the primary objective of this

model is to allow for analysis of the performance of groups of WTGs and how they
interact with the bulk power system. Hence, the Var control system and the APC are
both incorporated in the models. Representation of all the individual machines in a
large wind plant is inappropriate for most grid stability studies. Hence, there is
provision within the model structure to allow a single equivalent WTG machine model
to represent multiple WTGS. The model implementation allows the user access to
parameters that might reasonably be customized to meet the particular requirements
of a system application. These parameters are discussed in more detail in Section 5.
2.3 Overall Model Structure
From a load flow perspective, standard generator and transformer models are
required for initialization of the dynamic simulation program. These two devices are
represented by conventional load flow models. Details are presented in Section 3.
The dynamic models presented here are specific to GE WTGs. The implementation
is structured in a fashion that is similar to conventional generators. To construct a
complete WTG model, three device models are used, as shown in Figure 2-3:
• Generator/converter model
• Electrical control model
• Turbine and turbine control model
The generator/converter model injects real and reactive current into the network in
response to control commands, and represents low and high voltage protective
functions (e.g., low voltage ride through capability). The same generator/converter
model, with different data, is used to represent both DFAG WTGs (e.g., GE’s 1.5, 1.6 and
3.6 MW) and full converter WTGs (e.g., GE’s Multi-Megawatt 2.5 MW).
The electrical control model includes both closed and open loop reactive power
controls, and voltage regulation with either a simplified emulator of GE’s
WindCONTROL system or a separate, detailed model. This model sends real and
reactive commands to the generator/converter model. Different electrical control
models are used to represent DFAG WTGs and full converter WTGs
The turbine and turbine control model represents the mechanical controls,
including blade pitch control and power order (torque order in the actual equipment) to
the converter; under speed trip; rotor inertia equation; wind power as a function of
wind speed, blade pitch, rotor speed; and active power control. One model is used to
represent both DFAG and full converter WTGs. However, more functions (e.g., dynamic
braking resistor) are enabled for a full converter WTG than for a DFAG machine.
In addition, user-written models can be developed to represent wind gusts or other
profiles by varying input wind speed to the turbine model, or to represent additional
protective functions (e.g., over/under frequency).

Vreg bus
Vterm
Trip Signal
Ip (P)
Command
Electrical Generator/
Control Converter
Model Model Pgen , Qgen
E" or IQ (Q)
Command
Power
Order
Pelec
Wind
Wind Profile Speed Turbine &
Model Turbine Control
(User-written) Model Fterm
Figure 2-3. GE WTG Dynamic Model Connectivity.

3 Analytical Approach
In practice, a wind plant has a local grid collecting the output from the machines
into a single point of interconnection to the grid. Since the wind plant is made up of
many identical machines, it is a reasonable approximation to parallel all the machines
into a single equivalent large machine behind a single equivalent reactance. This
approach is consistent with the models presented in this report. However, there are
limitations. Disturbances within the local collector grid cannot be analyzed, and there
is some potentially significant variation in the equivalent impedance for the connection
to each machine. A single machine equivalent requires the approximation that the
power output of all the machines will be the same at a given instant of time. For grid
system impact studies, simulations are typically performed with the initial wind of
sufficient speed to produce rated output on all machines. Under this condition, the
assumption that all machines are initially at the same (rated) output is not an
approximation. Otherwise, this assumption presumes that the geographic dispersion
is small enough that the wind over the plant is uniform. Simulations of bulk system
dynamics using a single machine equivalent is adequate for most planning studies.
Detailed modeling of the WTG collector system is possible. The inclusion of the
WindCONTROL in each WTG’s electrical control model provides an emulation of the
action of a single centralized control. An intermediate level of modeling detail can also
be used in which groups of WTGs, e.g. those on a single collector feeder, are
represented by a single equivalent model.
3.1 Load Flow Model
The modeling of a GE WTG or wind plant for load flow analysis is generally simple.
As noted above, wind plants normally consist of a large number of individual WTGs.
While the wind plant model may consist of a detailed representation of each WTG and
the collector system, a simpler model is appropriate for most bulk system studies.
Such a model is shown in Figure 3-1. This model consists of a single WTG and unit
transformer with MVA ratings equal to N times the individual device ratings, where N is
the number of WTGs in the wind plant (or those considered on-line for study purposes).
An equivalent impedance to reflect the aggregate impact of the collector system can
be included together with the substation step-up transformer(s). The total charging
capacitance of the collector system should also be included. The charging
capacitance can be significant since underground cables are often used for the
collector system.
The aggregate WTG is modeled as a conventional generator connected to a (PV)
bus. The generator real power output (Pgen), maximum reactive power output (Qmax),
and minimum reactive power output (Qmin) are input as N times the unit capabilities
shown in Table 3-1. The nominal voltage at the generator terminals depends on the
WTG size and system frequency. Typical unit transformer ratings and impedances are
also shown. Typical collector system voltages are at distribution levels - 12.5 kV or
34.5 kV are common in 60 Hz applications, 33 kV in 50 Hz applications. The substation

transformer would be suitably rated for the number of WTGs, with an impedance
typically around 10%.
Project Substation
Point of High Side Bus

Interconnection (collector, e.g. 34.5kV) Terminal Bus
(POI) Bus
P gen
Collector
Equivalent
Impedance Q gen
Substation and Charging Unit
Vreg bus Transformer Capacitance Transformer Vterm
Figure 3-1. Simplified Wind Plant Power Flow Model.
Table 3-1. Individual WTG Power Flow Data.

GE 1.5 MW GE 1.6 MW GE 3.6 MW GE 2.5 MW
Generator Rating 1.67 MVA 1.78 MVA 4 MVA 3 MVA
Pmax 1.5 MW 1.6 MW 3.6 MW 2.5 MW
Pmin 0.07 MW 0.075 MW 0.16 MW 0 MW
Qmax 0.726 MVAr* 0.775 MVAr* 2.08 MVAr 1.20 MVAr*
Qmin -0.726 MVAr* -0.775 MVAr* -1.55 MVAr -1.20 MVAr*
Terminal Voltage (60 Hz) 690 V** 690 V 4160 V 690 V
Terminal Voltage (50 Hz) 690 V 690 V 3300 V 690 V
Unit Transformer Rating 1.75 MVA 1.75 MVA 4 MVA 2.8 MVA
Unit Transformer Z 5.75% 5.75% 7% 6.0%
Unit Transformer X/R 7.5 7.5 7.5 7.5
*These values are for +/- 0.90 power factor machines. GE also offers +/-0.95 power factor
machines, which would be modeled with a reactive range of +/-0.493 MVAr, +/-0.526 MVAr and
+/-0.822 MVAr for the 1.5 MW, 1.6 MW and 2.5 MW machines, respectively. Project specific
values must be confirmed.
**Older 60 Hz machines have 575 V terminal voltages.
The WindCONTROL system is structured to measure the voltage at a particular bus,

often the point of interconnection (POI) with the transmission system, and regulate this
voltage by sending a reactive power command to all of the WTGs. Line drop
compensation may be used to regulate the voltage at a point some distance from the
voltage measurement bus. For load flow modeling of the WindCONTROL, the
aggregate WTG (or each WTG) should be set to regulate the remote bus at the desired
voltage regulation point. Depending upon the applicable grid requirements for voltage
and reactive power range, the substation transformer may be an automatic load-tap-

change (LTC) transformer. Operation of the LTC controls may be autonomous, or
coordinated with the WindCONTROL.
3.2 Initial Conditions for Dynamic Simulation
The load flow provides initial conditions for dynamic simulations. The conditions
outlined above are generally applicable to the dynamic model presented below. The
maximum and minimum active and reactive power limits must be respected in order
to achieve a successful initialization.
If the WTG electrical control or additional substation controls are customized to
meet a particular set of desired performance objectives, then the load flow must be
initialized in accordance with those customized rules. For example, if the active power
controls are set to curtail power to 95% of that available in the wind, then the real
power at the load flow generator must be set accordingly. Similarly, it is possible to
inject or absorb reactive power (e.g., regulate voltage) at zero real power with a full
converter WTG. Therefore, the real power at the generator in the power flow must be
zero for this type of simulation.
Inconsistencies between the power flow and the dynamic model will result in an
unacceptable initialization.

4 Doubly Fed Asynchronous Generator (DFAG) Dynamic
Models
This section presents the engineering assumptions, detailed structure, and data for
each of the component models necessary to represent a GE 1.5, 1.6 or 3.6 MW WTG.
4.1 Generator/Converter Model

This model (gewtg in PSLF) is the equivalent of the generator and field converter,
and provides the interface between the WTG and the network. Unlike a conventional
generator model, it contains no mechanical state variables for the machine rotor –
these are included in the turbine model. Further, unlike conventional generator
models, all of the flux dynamics have been eliminated to reflect the rapid response of
the converter to the higher level commands from the electrical controls. The net result
is an algebraic, controlled-current source that computes the required injected current
into the network in response to the flux and active current commands from the
electrical control model. This controlled-current source also incorporates the low
voltage power logic and the fast-acting converter controls that mitigate over-voltages
by reducing reactive current output.
The model is shown in Figure 4-1. It holds constant both the active power (X-axis)
component of current and the X-axis voltage (Y-axis flux) behind the generator
effective reactance, X”. The real and reactive command signals are developed in the
electrical control model described in Section 4.2. The low-pass filters on the incoming
command signals are simple approximations to the complex, fast electronic control
system. This small lag (0.02 seconds) provides a reasonable representation in the time
frame of interest. As with all positive sequence fundamental frequency analysis, sub-
cycle behavior is not meaningful.
The Low Voltage Power Logic (LVPL) reduces system stress during and immediately
following sustained faults by limiting the real current command with both a cap (upper
limit) and a ramp rate limit. Under normal operating conditions, the filtered terminal
voltage is above a user-specified breakpoint (brkpt) and there is no cap. When the
voltage falls below the breakpoint during a fault, a cap is calculated and applied.
When the voltage is below a user-specified zero-crossing point (zerox), the cap
becomes zero. The user-specified ramp rate limit (rrpwr) is key to the post-fault power
recovery. During this recovery period, the voltage will exceed the breakpoint and the
cap is removed. However, the real current command rate of increase will be restricted
by the ramp rate limit.
Comparison with more detailed models of the generator and controls has shown
that this is an accurate model of the combined behavior of the doubly-fed generator
and its rotor converter when the value of X” is set to 0.80 pu (on the generator MVA
rating) for all the DFAG WTGs. X” represents an effective equivalent reactance and is
not the actual subtransient reactance of the doubly-fed induction generator.
The actual converter controls include a phase-locked loop (PLL) to synchronize the
generator rotor currents with the stator currents. However, the PLL dynamics are

extremely fast relative to the PSLF time frame, and under normal grid operating
conditions result in effectively perfect tracking. Under transient conditions of severe
voltage depression and relatively high system impedance, delivery of active current
becomes limited. The control actions of the PLL and current regulator effectively result
in reduced active current delivery. This fast regulator and PLL action is captured in the
model by a low voltage active current management function. This is a linear reduction
of active current injection for terminal voltages below 0.8 pu. This effect is modeled
within the network solution (i.e. without state variables), which is consistent with the
overall algebraic modeling of current injection by the generator/converter model. The
reactive current delivery remains high under these transient conditions, providing
voltage support and short circuit strength.
The fast regulator and PLL dynamics of the converter will also act to limit excess
voltage on the terminals of the generator by suppressing reactive current injection
when the terminal voltage rises excessively. This effect is modeled by a high voltage
reactive current management function in the network solution, which drives reactive
current injection down to limit terminal voltage to 120%. Reactive current injection is
limited to the machine rating.
Isorc
Eq"cmd 1 -1 High Voltage
(efd) 1+ 0.02s X" Reactive Current
From s0 Management
exwtge
LVPL & rrpwr
Low Voltage
IPcmd 1 IPlv Active Current
(ladifd) 1+ 0.02s Management
From s1
exwtge
LVPL
Vterm
1.22
LVPL V
1
jX"
1+ 0.02s
V s2
zerox brkpt
(0.50) (0.90)
Low Voltage Power Logic
Figure 4-1. DFAG Generator/Converter Model.
4.1.1 Fault Ride Through

The generator model also includes over/under voltage protective functions. In
particular, the low voltage tripping can be set to meet so-called “low-voltage ride
through” (LVRT) or “zero-voltage ride through” (ZVRT) requirements. These
requirements are explicitly defined such that wind plants must not trip for events that
are less severe than the defined thresholds and time durations. Wind plants may

tolerate more severe events without tripping. Use of the model therefore does not
ensure that the plant will trip, only that it is allowed to do so. The thresholds and time
durations for this protection will vary significantly from one project to another as
equipment designs are modified to meet specific grid codes or interconnection
agreements. Recommendations for modeling the generator protection functions are
as follows:
• For feasibility and reliability impact studies of future wind projects: Do not
include the generator protection model or else set the trip levels consistent with
applicable grid codes for the project. An objective of the study should be to
establish the voltage and frequency excursions that may occur. These results
should then be reflected in the equipment specifications. The mechanism for
communicating this is the interconnect agreement. Prior to establishing the
interconnect agreement, the product capability should be understood via
communication with the GE representative.
• For facility studies for projects in the design phase: Use trip settings
consistent with performance commitments. The results of the study should
indicate acceptable settings for the actual protective devices to satisfy system
requirements while providing adequate protection for the WTG equipment.
• For studies involving in-service projects: Use the actual trip settings of the
protective equipment.
Table 4-1 gives seven trip levels and durations based on specifications for a 60 Hz,
1.5 or 1.6 MW unit with the two fault ride through options offered by GE at the time this
document was updated. The PSLF model provides six levels of voltage tripping. The
short term high voltage threshold can be ignored as sub-cycle behavior in stability
simulations is not meaningful. It is important to note that the low voltage thresholds
are a stepwise fit to a curve that defines the equipment minimum performance
specifications. Figure 4-2 shows this graphically for the ZVRT option; with the step-
wise curve representing the trip points in the table. Figure 4-3 shows the same
information for the LVRT II option. The step-wise curve is conservative, in that it is
always inside the specification. As noted above, low voltage ride through requirements
vary from application to application. The tripping thresholds and durations should be
chosen to appropriately represent the application under study.
Any other desired protective functions (e.g., over/under frequency) would need to
be implemented with additional protective device models.
Table 4-2 includes recommended settings for the DFAG generator/converter model.
The maximum allowed ramp rate limit, rrpwr, is 10. The LVPL breakpoint, brkpt, must
be greater than 0.5, less than 1.0, and greater than the zero-crossing, zerox.

Table 4-1. Typical Fault Ride Through Voltage Thresholds and Durations.
Time (sec)
V (%) ΔV (pu) ZVRT LVRT II
75 -0.25 1.9 1.7
50 -0.50 1.2 1.1
30 -0.70 0.7 0.7
15 -0.85 0.2 0.02
110 0.10 1.0 1.0
115 0.15 0.1 0.1
130
0.30 0.02 0.02
(DFAG)
120
0.20 0.02 0.02
(full converter)
140
Voltage at Point of Interconnection
120
100
(Percent)
80
PSLF
60
40
GE ZVRT
20
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0
200 ms
Time (seconds)
Figure 4-2 GE 1.5 and 1.6 MW ZVRT Model Settings and Equipment Specification.

140
Voltage at Point of Interconnection
120
100
PSLF LVRT II
(Percent)
80
60
40
GE LVRT II
20
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Time (seconds)
Figure 4-3 GE 1.5 and 1.6 MW LVRT II Model Settings and Equipment Specification.
Table 4-2. DFAG Generator/Converter Parameters.

Recommended Values
Parameter Name ZVRT LVRT II
lpp 0.8 0.8
dvtrp1 -0.25 -0.25
dvtrp2 -0.50 -0.50
dvtrp3 -0.70 -0.70
dvtrp4 -0.85 -0.85
dvtrp5 0.10 0.10
dvtrp6 0.15 0.15
dttrp1 1.9 1.7
dttrp2 1.2 1.1
dttrp3 0.7 0.7
dttrp4 0.2 0.02
dttrp5 1.0 1.0
dttrp6 0.1 0.1
fcflg 0 0
rrpwr 10. 10.
brkpt 0.9 0.9
zerox 0.5 0.5

4.2 Converter Control Model
This model (exwtge in PSLF) dictates the active and reactive power to be delivered
to the system based on inputs from the turbine model (Pord) and from the supervisory
VAr controller (Qord). Qord can either come from a separate model or from the
WindCONTROL voltage and reactive control emulator function included in the electrical
control model. Qord can also be held constant or determined by a power factor
regulator. The model consists of the following control functions:
WindCONTROL Emulator
Power Factor Regulator
Electrical (Volt/VAr) Control
Open Loop Control Logic (used only on some older systems)
The overall block diagram for the Reactive Power Control and the Electrical Control
is shown in Figure 4-4. These controls are described in more detail in the following
sections.
Vrfq
WindCONTROL
Vreg
Emulator
From Qord Open Qcmd

separate Loop
WindCONTROL Control
model Logic Eq"cmd
Qref
PFAref Power
Pelec Electrical To Generator
Factor
Control Model
Regulator
Reactive Power Control
Qgen IPcmd
Vterm
From
Wind Turbine Pord
Model
Figure 4-4. Overall DFAG Reactive Power and Electrical Control Model.
4.2.1 Reactive Power Control

A more detailed representation of the Reactive Power Control is shown in Figure
4-5.
The WindCONTROL emulator function represents a simplified equivalent of the
supervisory VAr controller portion of the entire wind farm management system

(WindCONTROL). The function monitors a specified bus voltage and compares it
against the reference voltage. Three regulated bus options are available: the WTG
terminal bus, a user-specified remote bus (e.g. point of interconnection), or a
synthesized point in the power system. The latter bus is synthesized from local voltage
and current measurements, and the compensating reactance, Xc. The regulator itself
is a PI controller. The time constant, Tc, reflects the delays associated with cycle time,
communication delay to the individual WTGs, and additional filtering in the WTG
controls. The voltage measurement lag is represented by the time constant Tr. Table
4-3 gives suggested settings for the WindCONTROL emulator model.
The parameter, fN, is the fraction of wind turbines in the wind plant that are on-line.
For example, if a case represents a condition with half of the wind turbines on-line, fN
should be set to 0.5. In this case the MVA base of the generator should also be set to
one-half of its full value, and the MW capability of the turbine should be set to one-half
of its full value. If a wind plant is represented by more than one WTG model, the fN
values of each should be set to the same value.
The other reactive power control method available is power factor control. It is
enabled by setting pfaflg to 1. The data associated with this mode are also shown in
Table 4-3. The appropriate flag and gain settings to represent various control
strategies are described in the next section.
WindCONTROL Vrfq
Emulator (vref)
Vermx Qmax
Kiv/s
Vreg + +
1 s4 1
1/fN Qord
1+ sTr - + Qwv 1+ sTc
Kpv
s3 - Vermn 1+ sTv Qmin s5
Vqd s2
From Q Droop
Function
If yes, freeze
2 integrators
< Vfrz?
PFAref Qref
(vref) tan (vref) Qord from separate model
0
Pelec 1 1 (vref)
1+ sTpwr
x
s6 0 -1 Qmax
pfaflg Open
1
Qord Loop Qcmd
Control
varflg Qmin
Figure 4-5. Reactive Power Control Model.

Table 4-3. Reactive Power Control Parameters (on Generator MVA Base).
Recommended Value
Variable Name PSLF Parameter 1.5 MW 1.6 MW 3.6 MW
Tr (sec) tr 0.02 0.02 0.02
Tv(sec) tv 0.05 0.05 0.05
fN fn 1.0 1.0 1.0
Tc(sec) tc 0.15 0.15 0.15
Kpv** kpv 18. 18. 18.
Kiv** kiv 5. 5. 5.
Qmax (pu) qmax 0.436* 0.436* 0.52
Qmin (pu) qmin -0.436* -0.436* -0.39
Tpwr (sec) tpwr 0.05 0.05 0.05
Xc (pu) xc 0 0 0
Vermn (pu) vermn -0.1 -0.1 -0.1
Vermx (pu) vermx 0.1 0.1 0.1
Vfrz (pu) vfrz 0.7 0.7 0.7
*+/-0.296 pu on GE 1.5 and 1.6 MW machines with +/- 0.95 power
factor rating, and on some older GE WTGs.
**Subject to field tuning to meet system performance objectives, as
discussed below.
The PI gains, Kpv and Kiv, are field adjustable to meet performance objectives and
may be adjusted in the model, if necessary. When GE wind plants are commissioned,
the value of these parameters are field tuned to provide good voltage performance for
systemic response to variations in wind power due to wind fluctuations. The field
tuned values are typically on the order of 2.0 and 1.0, respectively. This gives a roughly
10 second response time. Experience has been very good with these values. The
values given in the table above are rough upper limits, based on GE simulation and
experience. They should be suitable for systems with a short circuit capacity of 5 or
more times the wind plant MW rating. These higher gains will give better voltage
response to grid voltages disturbances. However, higher gains result in increased risk
of instability – much the way AVR gains can destabilize conventional synchronous
machines. As a system weakens, the effective close-loop response gets faster. Thus,
selection of higher gains for system performance must be accompanied by analysis
that assures stable operation under all credible operating conditions – especially the
minimum short circuit strength condition.
Delivery of reactive power is given priority over active power by the controls. For
the 1.6 MW machine only, sustained high reactive power output at low terminal
voltages may result in a few percent drop in active power output. This is not modeled.
The Q Droop function, shown in Figure 4-6, is a relatively slow-acting function that
reduces the effective voltage reference (Vrfq-Vqd) as reactive power changes. This
improves coordination between multiple integral controllers regulating the same point

in the system. By default, the Q Droop function is disabled. It may be enabled by
setting the gain parameter, Kqd, to a non-zero value. Typical data is shown in Table
4-4. There are three options for the reactive power input to this function: reactive
power generated by the WTG, reactive power flow in a user-specified branch, or a
synthesized reactive power. The latter is the reactive power flow in the user-specified
branch plus a secondary term, Xqd*Im2, where Im is the magnitude of the current
flowing in that branch.
Q Input 1 Vqd
Kqd
1+ sTlpqd
s7
Figure 4-6. Q Droop Function Model.
Table 4-4. Q Droop Function Parameters.

Variable Name PSLF Parameter Recommended Value
Tlpqd (sec) tlpqd 5.0
Kqd kqd 0.04
Xqd (pu) xqd 0.0
4.2.2 Electrical Control

A more detailed representation of the Electrical Control is shown in Figure 4-7. The
electrical controller model is a simplified representation of the converter control
system.
The volt/var control monitors the generator reactive power, Qgen, and terminal
voltage, Vterm, to compute the voltage command Eq”cmd required to meet the Qcmd from
the Reactive Power Control. The Qcmd signal is compared to the reactive power
generated by the converter, and the resulting error is integrated with a gain of Kqi, to
generate a voltage reference, Vref. Thus, the reactive power command is implemented
via a slowly changing voltage reference. The subsequent voltage control block is
significantly faster. The voltage reference is compared to the actual terminal voltage,
and the resulting voltage error is multiplied by a gain and integrated to compute the
voltage command Eq”cmd. Thus, a drop in terminal voltage, e.g., in response to a system
fault, results in an immediate large voltage error and an increased reactive command.
The magnitude of the gain determines the effective time constant associated with the
voltage control loop. The voltage command, Eq”cmd, is limited to reflect hardware
constraints.
An auxiliary test signal can be injected into the terminal bus voltage regulator via
model[@index].sigval[0], as shown in Figure 4-7. A user-written dynamic model
(epcmod) is needed to generate the desired signal. The index of the wind generator
model (@index) can be obtained using the model_index function.

The real current command signal, IPcmd, is developed from the wind turbine model
power order and the terminal voltage.
Table 4-5 includes recommended settings for the electrical control model. All
settings are given in terms of rated MVA.
Qgen
Vterm
Vmax XIQmax
- -
Qcmd Vref
KQi / s KVi / s Eq"cmd
+ + s1
(efd)
Vmin
s0 + XIQmin
To Generator
Auxiliary Model
Test Signal
( model[@index].sigval[0] )
Pord IPmax
(vsig)
. IPcmd
.
(ladifd)
From
Wind Turbine
Model Vterm
Figure 4-7. DFAG Electrical Control Model.
Table 4-5. DFAG Electrical Control Parameters.

KQi kqi 0.1*
KVi kvi 40**
XIQmax xiqmax 1.45
XIQmin xiqmin 0.50
Vmax vmax 1.10
Vmin vmin 0.90
Ipmax ipmax 1.1
* North American WindCONTROL, see text for other configurations.
** Subject to field tuning
4.2.3 Control Strategies

A variety of control strategies can be represented, including voltage regulation of a
remote bus or constant power factor control, both with and without the supervisory
WindCONTROL. The various strategies can be implemented by setting the varflg,
pfaflg, and Kqi parameters as follows:

• Operation with WindCONTROL and with North American “volt/VAr” control
(varflg = 1, pfaflg = 0, Kqi = 0.1) This represents the normal configuration for
recent and future North American wind plants, using the WindCONTROL
emulator in the model to represent the wind farm management system.
• Operation without WindCONTROL and with North American “volt/VAr”
control (varflg = 0, pfaflg = 0, Kqi = 0.001) With the WindCONTROL turned off,
Kqi is reduced so there is a slow reset to desired reactive power and WTG
terminal voltage control is rapid. (This combination of flags and Kqi = 0.1 can be
used to emulate WindCONTROL at a fixed plant reactive power control.)
• Operation without WindCONTROL and with European fast power factor
control (varflg = 0, pfaflg = 1, Kqi = 0.5) This represents the common
configuration for European wind parks, where a set power factor angle is
rapidly regulated by the converter control. Closed loop voltage control is not
used on these systems, but is left in the model to approximately represent other
means that are used to limit voltage excursions that would otherwise cause
unit tripping.
• Operation with WindCONTROL and with European fast power factor control
(varflg = 1, pfaflg = 0, Kqi = 0.5) This represents the another configuration for
European wind parks when WindCONTROL is employed. Similar to the North
American model except the regulator gain is at the higher value. The power
factor control flag, pfaflg, is set to zero because the signal from the
WindCONTROL is a reactive power order, rather than power factor angle.
4.2.4 Open Loop Control Logic

This feature was used in some wind plants with GE WTGs before the
implementation of the local closed-loop terminal voltage control described above. The
open loop control logic is responsive to large variations in system voltage, and is
inactive whenever the terminal voltage is within its normal range. It is described by
Table 4-6. The functions in this table represent an optional open loop control that was
implemented to improve system performance for large voltage deviations resulting
from system events. The open loop control logic forces the reactive power to pre-
specified levels as voltage deviations persist. This feature has not been offered for
several years, and will not be offered for future projects with GE WTGs. However,
representative values from earlier projects for the open loop control parameters are
given in Table 4-7.

Table 4-6. Open Loop Reactive Power Control Logic.
Voltage Condition Time Duration Open Loop Reactive
Power Command
Vterm < VL1 t < TL1 QL1
TL1 < t < TL2 QL2
t > TL2 QL3
Vterm > VH1 t < TH1 QH1
TH1< t < TH2 QH2
t > TH2 QH3
Table 4-7. Open Loop Reactive Power Control Parameters.

Recommended Value
No Open Loop
Variable Name PSLF Parameter Open Loop Control**
Control
VL1 (pu) vl1 0.9 -9999.
VH1 (pu) vh1 1.1 9999.
TL1 (sec.) tl1 0.1 0.
TL2 (sec.) tl2 0.5 0.
TH1 (sec.) th1 0.1 0.
TH2 (sec.) th2 1.0 0.
QL1 (pu) ql1 0* 0.
QL2 (pu) ql2 0.45 0.
QL3 (pu) ql3 0* 0.
QH1 (pu) qh1 0* 0.
QH2 (pu) qh2 -0.245 0.
QH3 (pu) qh3 0* 0.
* The closed-loop Q command, Qord, is passed without modification by setting this to 0.
** Only for some projects before mid-2003, check with owner.

4.3 Wind Turbine and Turbine Control Model
The wind turbine model (wndtge in PSLF) provides a simplified representation of a
very complex electro-mechanical system. The block diagram for the model is shown in
Figure 4-8. In simple terms, the function of the wind turbine is to extract as much
power from the available wind as possible without exceeding the rating of the
equipment. The wind turbine model represents the relevant controls and mechanical
dynamics of the wind turbine. This model is used for both the DFAG and full converter
WTG models. However, some of the features (e.g., dynamic braking resistor) are not
applicable to the DFAG machines. The differences will be discussed in detail in the
following subsections, as well as in Section 5.3.
Details of the turbine control model (torque control, pitch control, pitch
compensation) are described in Section 4.3.1. The rotor mechanical model is described
in Section 4.3.2. The Wind Power Model is a moderately complex algebraic relationship
governing the mechanical shaft power that is dependent on wind velocity, rotor speed
and blade pitch. This model is described in Section 4.3.3. The details of the optional
controls are described in Section 4.3.4, Active Power Control emulator, and Section
4.3.5, WindINERTIA model.
The converters used in both the DFAG and full converter machines make it possible
to inject or absorb reactive power (e.g., regulate voltage) at zero real power. However,
the reactive capability of the DFAG model is reduced, from +/- 0.436pu to +/- 0.12pu,
while operating with the WindFREE control. Zero power may be the result of no wind,
excessive wind, or an operator directive to curtail output. All three of these zero power
scenarios may be simulated with this model by setting the wfflg parameter to 1.
For the first scenario (no wind), a user-written wind profile model is required to drive
a specific WTG’s wind speed below the low wind speed threshold. As the wind speed
drops, so does the machine speed, electrical power and mechanical power. Below the
low wind speed threshold, the decrease in mechanical power is implemented with a
pseudo drag term. Without the WindFREE control, a trip signal is generated when the
rotor speed falls below 0.10 pu. With this control feature, the generator model will still
respond to reactive power commands from the electrical control model (e.g.,
WindCONTROL emulator). The low wind speed threshold is set at 3 m/sec, and cannot
be changed by the user.
The second scenario (excessive wind) also requires a user-written wind profile
model. Once the wind speed exceeds the high wind speed threshold, the difference
between the wind speed and the threshold is integrated. If that value exceeds another
threshold and the WindFREE control is enabled, the machine goes into zero-power
operation (i.e., electrical power and rotor speed are zero). The integration represents
an inverse time function. The more excessive the wind speed, the earlier the unit
enters zero-power operation. All of the wind speed thresholds and timers are internal
to the model and cannot be changed by the user. The high wind speed threshold is
currently set at 25 m/sec and the inverse-time threshold at 8 m-sec/sec.

For the third scenario (curtailed output), the real power at the generator in the
power flow is zero. The WTG will initialize at zero wind speed, machine speed, electrical
and mechanical power. Again, the generator model will inject or absorb reactive
power in response to reactive power commands from the electrical control model.
The key limitation on zero-power operation is that once in that mode, the WTG
stays there. The model is neither applicable nor appropriate for simulating start-up
scenarios.

From
Pdbr
+ Σ
+ From
Pelec getwg
ewtgfc
(elimt) (pelec)
Wind
Speed
Trip
(glimv) Wind Pmech ω Under Signal
Rotor
Power Speed
Model
Model Trip To
s6 s9 getwg
Blade ωrotor (glimt )
Pitch
θ ω
θ d θ /dt max Anti-windup on
max & Pitch Limits
+
1 θ cmd + 1
Σ Kpp+ Kip/s Σ - 0.75P2elec+ 1.59Pelec+ 0.63
1+ sT p ω err ω ref 1 + 60s
s1
s0 + s5
θ θ Pitch Control
min & d /dt min Torque Control
Anti-windup on ω
Power Limits P & d P /dt
wmax max
K ptrq + Kitrq / s 1
X
1+ sTpc
s2 s4
Anti-windup on Pwmin& d P /dtmin

Pitch Pitch Limits
Compensation
+ pinp
K pc+ K ic / s Σ
s3
1.
pstl Power Response
Pmax Rate Limit
WTG Terminal Active Power 0 PsetAPC
Bus Frequency Control 1
(optional) apcflg
+
plim
Σ
fbus P
min +
Σ
+ perr
Auxiliary sTw
Signal 1 + sTw
(psig) + s10
WindINERTIA
Control
dpwi + Σ
(optional) + wsho
Pord
To extwge
or ewtgfc
(vsig)
Figure 4-8. Wind Turbine Model Block Diagram.
4.3.1 Turbine Control Model

The central part of Figure 4-8 is the model of the turbine controls. The practical
implication of the turbine control is that when the available wind power is above the

equipment rating, the blades are pitched to limit the mechanical power (Pmech)
delivered to the shaft to the equipment rating (1.0 pu). When the available wind power
is less than rated, the blades are set at minimum pitch to maximize the mechanical
power. The dynamics of the pitch control are moderately fast, and can have
significant impact on dynamic simulation results. A detailed diagram of that portion of
the model is shown in Figure 4-9.
rate limit PImax
Kpp (PIrate) Blade
ωerr + Pitch
θ cmd 1
ω Σ Σ Σ θ
+ * + 1 + sTPI
Kip + +
s PImin
ωref Non-
windup limit
Pitch Control
Kpc
+
pinp Σ Σ
+ * +
Kic
s
pstl Non-
windup limit
Pitch
Compensation
* The Pitch Control and Pitch Compensation integrators are non-windup integrators as a function of the pitch, i.e.,
the inputs of these integrators are set to zero when the pitch is in limits (Pimax or Pimin) and the integrator input
tends to force the pitch command further against its limit. The outputs of these integrators are not limited.
Figure 4-9. Pitch Control and Pitch Compensation Block Diagram.
The turbine control model sends a power order to the electrical control, requesting
that the converter deliver this power to the grid. The electrical control, as described in
Section 4.2, may or may not be successful in implementing this power order. The
control of turbine speed is quite complex. For modeling purposes, this is approximated
by closed loop control with a speed reference that is proportional to electric power
output. For power levels above rated, the rotor speed will be controlled primarily by
the pitch control, with the speed being allowed to rise above the reference transiently.
The actual control does not use a speed reference or a feedback of power.
In this model, the blade position actuators are rate limited and there is a time
constant associated with the translation of blade angle to mechanical output. The
pitch control does not differentiate between shaft acceleration due to increase in wind
speed or due to system faults. In either case, the response is appropriate and relatively
slow compared to the electrical control.

The model reference speed is normally 1.2 pu, but is reduced for power levels
below 46%. This behavior is represented in the model by using the following equation
for speed reference when the power is below 0.46 pu:
ωref = -0.75P2 + 1.59P + 0.63
The speed reference slowly tracks changes in power with a low pass filter time
constant of 60 seconds. Note: In the actual controller, the speed reference is not
directly a function of power, but the overall effect on the speed/power relationship
is similar.
The turbine control acts to smooth out electrical power fluctuations due to
variations in shaft power. By allowing the machine speed to vary around reference
speed, the inertia of the machine functions as a buffer to mechanical power variations.
The model includes a low rotor speed tripping function. If the DFAG machine speed
falls below 0.10 pu, the WTG is tripped instantaneously.
The model also includes high and low wind speed cut-out for the turbine. For the
DFAG machine, this results in a generator trip. All of the wind speed thresholds and
timers are internal to the model and can not be changed by the user.
The high wind speed threshold is currently set at 25 m/sec. The difference between
the wind speed and the high wind speed threshold is integrated and if that value
exceeds another threshold (8 m-sec/sec), the unit is tripped. The integration represents
an inverse time function. The more excessive the wind speed, the faster the unit is
tripped.
The low wind speed threshold is set at 3 m/sec. For this function, the decrease in
rotor speed and power is approximated with a pseudo drag term. The unit is tripped,
via the low rotor speed tripping function described above, when the rotor speed falls
below 0.10 pu.
Once a WTG has tripped, it can not be started again. The model is neither
applicable nor appropriate for simulating start-up scenarios.
Parameter values for the DFAG wind turbine control model are shown in Table 4-8.
None of these values should be modified by the user unless advised to so by the
manufacturer.

Table 4-8. DFAG WTG Turbine Control Parameters (on Turbine MW Base).
Kpp kpp 150.
Kip kip 25.
Tp (sec) tp 0.30
θmax (deg) pimax 27.
θmin (deg) pimin 0.0
dθ/dt max (deg/sec) pirat 10.0
dθ/dt min (deg/sec) -pirat -10.0
Pwmax (pu) pwmax 1.12
Pwmin (pu) pwmin 0.04
dP/dt max (pu/sec)* pwrat 0.45
dP/dt min (pu/sec)* -pwrat -0.45
Kpc kpc 3.0
Kic kic 30.0
Kptrq kptrq 3.0
Kitrq kitrq 0.6
Tpc tpc 0.05
wfflg wfflg 0
*This pair of values needs to be set to +/- 2.0 when the 2-mass
mechanical model is used, as described in the next section.
4.3.2 Rotor Mechanical Model

The block labeled “Rotor Model” in the upper part of Figure 4-8 includes the rotor
inertia equation for the WTG rotor. This equation uses the mechanical power from the
wind power model and the electrical power from the generator/converter model to
compute the rotor speed. GE recommends using the simple single-mass equivalent.
However, a two-mass rotor model, with separate masses for the turbine and
generator, is available, as shown in Figure 4-10. Note that the power from the
dynamic braking resistor, Pdbr (also shown in Figure 4-8) is only used with full
converter WTGs as discussed in Section 5.3.
The data for the rotor mechanical model are given in Table 4-9. These parameters
result in torsional oscillation frequencies of approximately 1.8 Hz for the 1.5 and
1.6 MW machines, and 2.6 Hz for the 3.6 MW machine. The torsional damping
coefficient, Dtg, is set to approximate the damping provided by a damping function in
the controller. Increasing the power rate limits, dP/dt max (pu/sec), as noted above is
also required to approximate this damping function.

ωο
spd0
+
Σ ωrotor
Turbine Speed
+
Tmech = Pmech
+
Tmech 1 1 ωbase 1 s6 + ω0
Σ 2H s s
+ s6 s7
+ - -
Dtg Σ Σ Ktg Tshaft
- + +
Telec 1 1 Telec = Pelec + Pdbr
Σ
1 ωbase
2Hg s s
- s8 s9
s8 + ω0
-
ωο
Δω spd0
+ ω
Σ Generator Speed
+
Figure 4-10. Two-Mass Rotor Model.
Table 4-9. DFAG WTG Rotor Mechanical Model Parameters (on Turbine MW Base).
GE 1.5 MW GE 1.6 MW GE 3.6 MW
Variable Name PSLF Parameter 60 Hz 50 Hz 60 Hz 50 Hz 60 Hz 50 Hz
One-Mass Model
H h 4.94 5.29 4.63 4.96 5.23 5.74
Two-Mass Model (Not Recommended)
H h 4.33 4.33 4.06 4.06 4.32 4.32
Hg hg 0.62 0.96 0.58 0.90 0.91 1.42
Ktg ktg 1.11 1.39 1.04 1.30 3.16 3.95
Dtg dtg 1.5 2.3 1.5 2.3 1.5 2.3
ωbase* wbase 125.66 157.08 125.66 157.08 125.66 157.08
* nominal generator speed

4.3.3 Wind Power Model
For power system simulations involving grid disturbances, it is a reasonable
approximation to assume that wind speed remains uniform for the 5 to 30 seconds
typical of such cases. However, the mechanical power delivered to the shaft is a
complex function of wind speed, blade pitch angle and shaft speed. Further, with wind
generation, the impact of wind power fluctuations on the output of the machines is of
interest. The turbine model uses the wind power model to provide this mapping.
The function of the wind power model is to compute the wind turbine mechanical
power (shaft power) from the energy contained in the wind, using the following
formula:
ρ
P= A r v 3w Cp (λ , θ)
2
P is the mechanical power extracted from the wind, ρ is the air density in kg/m3, Ar is
the area swept by the rotor blades in m2, vw is the wind speed in m/sec, and Cp is the
power coefficient, which is a function of λ and θ. λ is the ratio of the rotor blade tip
speed and the wind speed (vtip/vw), and θ is the blade pitch angle in degrees. The
relationship between blade tip speed and turbine rotor speed, ω, is a fixed constant, Kb.
Thus, the calculation of λ becomes:
λ = Kb (ω/vw)
For GE WTGs, the parameters given in Table 4-10 will result in Pmech in pu on the
unit’s MW base.
Table 4-10. DFAG WTG Wind Power Coefficients.
GE 1.5 MW GE 1.6 MW GE 3.6 MW
½ρ Ar 0.00159 0.00159 0.00145
Kb 56.6 56.6 69.5
Cp is a characteristic of the wind turbine and is usually provided as a set of curves
relating Cp to λ, with θ as a parameter. Representative Cp curves for the GE wind
turbines are shown in Figure 4-11. Curve fitting was performed to obtain the
mathematical representation of the Cp curves used in the model:
4 4
Cp (θ, λ) = ∑ ∑ αi, j θi λj
i=0 j=0
The coefficients αi,j are given in Table 4-11. The curve fit is a good approximation
for values of 3 < λ < 15, which are suitable for stability simulations for all blade
configurations and models. These curves should not be used for energy production or
other economic evaluation. Values of λ outside this range represent very high and low
wind speeds, respectively, that are outside the continuous rating of the machine.

0.5
θ=1o
0.4
θ=3o
θ=5o
0.3
θ=7o
Cp
θ=9o
0.2
θ=11o
o
θ=13
o
θ=15
0.1
0
0 2 4 6 8 10 12 14 16 18 20
λ
Figure 4-11. Wind Power Cp Curves.
Initialization of the wind power model recognizes two distinct states: 1) initial
electrical power (from the load flow) is less than rated, or 2) initial electrical power
equal to rated. In either case, Pmech = Pelec is known from the load flow and ω = ωref is
set at the corresponding value (1.2 pu if P > 0.46 pu). Then, using the Cp curve fit
equation, the wind speed vw required to produce Pmech with θ = θmin is determined.
(Notice from Figure 4-11, that two values of λ will generally satisfy the required Cp for a
given θ. The wind speed vw, corresponding to the higher λ is used.) If Pmech is less than
rated, this value of wind speed is used as the initial value. If Pmech is equal to rated and
the user-input value of wind speed, spdw1, is greater than the θ = θmin value, then θ is
increased to produce rated P at the specified value of wind speed. If Pmech is equal to
rated and the user-input value of wind speed is zero, the initial wind speed, spdw1, is
14 m/s.
Large negative values of Cp are not allowed. The minimum is set to –0.05.

Table 4-11. Cp Coefficients αi,j
i j αij
4 4 4.9686e-010
4 3 -7.1535e-008
4 2 1.6167e-006
4 1 -9.4839e-006
4 0 1.4787e-005
3 4 -8.9194e-008
3 3 5.9924e-006
3 2 -1.0479e-004
3 1 5.7051e-004
3 0 -8.6018e-004
2 4 2.7937e-006
2 3 -1.4855e-004
2 2 2.1495e-003
2 1 -1.0996e-002
2 0 1.5727e-002
1 4 -2.3895e-005
1 3 1.0683e-003
1 2 -1.3934e-002
1 1 6.0405e-002
1 0 -6.7606e-002
0 4 1.1524e-005
0 3 -1.3365e-004
0 2 -1.2406e-002
0 1 2.1808e-001
0 0 -4.1909e-001
4.3.4 Active Power Control Model & Rate Limit Function

The Active Power Control (APC) model and rate limiting function are shown in detail
in Figure 4-12. The APC model is a simple representation of the active power control
required by many European grid codes. This function is a portion of the wind farm
management system (WindCONTROL). The primary objectives of the APC are to:
• Enforce a maximum wind plant power output
• Provide a specified margin by generating less power than is available
• Enforce a plant power ramp rate limit
• Respond to system frequency excursions

By default, the APC model is disabled. It may be enabled by setting the data
parameter, apcflg, to 1.
Under normal operating conditions with near nominal system frequency, the
control is either enforcing a maximum plant output (i.e., Pmax) or providing a specified
margin by generating less power than is available from the wind (e.g., actual power
generated is 95% of the available power, or Pbc = 0.95).
In response to frequency excursions, the control switches into another mode and
calculates a plant power order as a function of system frequency. This path requests a
higher than usual power order for low frequency events, and lower than usual power
order for high frequency events. Thus, the wind plant will generate additional power in
response to the loss of other generating facilities or less power in response to the loss
of load.
An auxiliary frequency signal, normally zero, may be set by a user-written model to
test APC performance in response to other types of frequency deviations. Such
deviations could include frequency steps, ramps, or other functions defined by the host
utility’s interconnection requirements.
An example frequency response curve is shown in Figure 4-13. Points A through D
on this response curve may be set to meet specific performance objectives or
requirements of the host grid. The value of Pd should be greater than or equal to the
minimum power, which is discussed below. The value of Fb must be less than 1, and
that of Fc must be greater than 1. The value of Tpav may be increased to simulate fixed
power reference.
Wind Wind 1. Active Power Control

Speed Power (optional)
(glimv) Model pavl
1
1+sTpav
WTG Terminal 1.
Bus Frequency pavf s11
Pmax
+ 0
fbu Frequency
pset 1
Σ s Response pstl
+ Curve
Pmin
Auxiliary To gewtg apcflg
Signal Trip Signal
(glimt) Release
(psig) Pmax
If (fbus < fb) fflg 1 if fflg set
OR (fbus > fc)
Figure 4-12. Active Power Control Emulator.

1.2
Point Point B
A (Fb,Pbc) Point C
1 (F P ) (Fc,Pbc)
0.8
Active Power Output (pu)
0.6
Point D
0.4
(Fd,Pd)
0.2
0
0.95 0.96 0.97 0.98 0.99 1 1.01 1.02 1.03 1.04 1.05 1.06 1.07
Frequency (pu)
Figure 4-13. Example Frequency Response Curve.
The two primary inputs to the frequency response curve are available power
(determined from wind speed, a Cp curve and the constant Kb) and WTG terminal bus
frequency. At nominal frequency, the filtered version of the available power is
multiplied by the factor, Pbc, to generate a power set point, pset. This set point is
compared to specified limits, Pmin and Pmax. The minimum power is nominally 0.20 pu
of maximum plant output. The maximum power represents an operator specified
plant output limit. For example, this may represent a limit that would be imposed on a
given wind plant after the loss of a local transmission line, or under light load
conditions.
In response to frequency excursions, the filtered available power is multiplied by
the appropriate interpolated factor to generate a power set point. No operator limit is
imposed for frequency excursions. The plant is still limited to the maximum power
rating of the WTGs and to the available power from a given wind speed. If the
frequency exceeds Fd for 1 second, the wind model is tripped.
Under all frequency conditions, the maximum power set point, PsetAPC, is an input to
both the pitch compensation (described in Section 4.3.1) and the power response rate
limit function. This rate limit is implemented by applying the maximum power set point
(PsetAPC) to the power order (pinp) from the turbine control, calculating the difference
between the original power order and the limited power order, processing that error
through a washout filter, and adding the output of the washout to the limited power
order to generate the final power order (Pord) for the converter control (extwge) model.

The time constant of the washout filter determines the ramp rate limit imposed on
changes to the power order signal. This function is always in service and is not
disabled by setting apcflg to 0. Example data for both the APC and power response
rate limit functions are shown in Table 4-12.
Table 4-12. Active Power Control and Rate Limit Function Parameters.
Tw (sec) tw 1.0
apcflg apcflg 0
Tpav (sec) tpav 0.15
Pa (pu) pa 1.0
Pbc (pu) pbc 0.95
Pd (pu) pd 0.40
Fa (pu) fa 0.96
Fb (pu) fb 0.996
Fc (pu) fc 1.004
Fd (pu) fd 1.04
Pmax (pu) pmax 1.0
Pmin (pu) pmin 0.2
4.3.5 WindINERTIA Model

System disturbances that include the loss of generation normally result in transient
depressions of system frequency. The rate of frequency decline, the depth of the
frequency excursion, and time required for system frequency to return to normal are
affected by the dynamic characteristics of generation connected to the grid. In the
first few seconds following the loss of a large generating plant, the frequency
dynamics of the system are dominated by the inertial response of the generation.
Conventional synchronous generation inherently contributes some stored inertial
energy to the grid, reducing the initial rate of frequency decline and allowing slower
governor actions to stabilize grid frequency.
Most modern MW-class wind generation does not exhibit this inertial response.
However, GE’s WindINERTIA™ feature provides an inertial response capability for wind
turbines, similar to that of conventional synchronous generators, for large under-
frequency grid events. Note that this control is asymmetric: it only responds to low
frequencies. High frequency controls are handled separately by the APC described
above. Fast supplemental controls, added to the fast power electronics and
mechanical controls of the WTG, take advantage of the inertia in the rotor. For these
large underfrequency events, this feature temporarily increases the power output of
the wind turbine in the range of 5% to 10% of the rated turbine power. The duration of
the power increase is on the order of several seconds. This benefits the grid by

allowing other non-wind power generation assets time to increase their power
production during under-frequency events.
The power delivery of the wind turbine-generator is limited not only by the
available wind, but by the physical limitations of the components of the WTG. Most
critical are aero-mechanical ratings and speed limits. A key point is that slowing the
turbine tends to reduce the aerodynamic lift, thereby reducing the delivered
mechanical shaft torque and exacerbating the speed decline caused by increased
generator electrical torque. This positive feedback tends to push the blade towards
aerodynamic stall, which must be avoided. The inertial control must provide margin
above stall, and is consequently limited when the initial rotor speed is low. The power
and energy of the inertial response declines below about 50% rated power, dropping
to zero at below about 20%. Inertial energy extracted by slowing the rotation of the
turbine must ultimately be recovered. After the initial increase in electrical power, it
must temporarily drop below the mechanical power to allow the energy to recover,
reaccelerating the rotor.
The control philosophy is to sense significant grid frequency depressions, as
observed at the terminals of the individual wind turbine generators, and to temporarily
increase power output. The simplified model of this function is shown in Figure 4-14.
Frequency error is simply the deviation from nominal, with an auxiliary frequency
signal, normally zero, available to implement specific test functions. A positive
frequency error means the frequency is low and extra power is needed. The deadband
suppresses response of the controller until the error exceeds a threshold. This limits
the WindINERTIA response to large events - those for which inertial response is
important to maintain grid stability, and for which seriously disruptive consequences,
like UFLS, may result. The continuous small perturbations in frequency that
characterize normal grid operation are not passed through to the controller. The
deadband output signal is further filtered, coordinated with the other turbine controls,
and limited. The coordination, as implemented in the turbine control model, consists of
modifications to several gains and time constants. Specifically, the proportional gain in
the torque control, Kptrq, is reduced to 0.5, the integral gain, Kitrq, is reduced to 0.05, and
the subsequent filter time constant, Tpc, is increased to 4.0 seconds, when the
WindINERTIA function is active (i.e., an under-frequency event is simulated). There is no
reset, so these parameter changes are permanent for any given simulation. The final
WindINERTIA command is added to the power order and implemented by the WTG
converter controls, ultimately resulting in additional power delivery.
By default, the WindINERTIA model is disabled. It may be enabled by setting the
gain parameter, Kwi, to a non-zero value. Example data for this function are shown in
Table 4-13.

WindINERTIA Control
WTG Terminal (optional)
Pmxwi
Bus Frequency
+
ferrwi dfdbwi fltdfwi dpwi
fbus 1 sTwowi
Σ 1 + sTlpwi
Kwi
1 + sTwowi
+ - dbwi
+
s12 s13
Auxiliary
Signal Frequency Pmnwi
(psig) Reference
Rate limits:
urlwi = up rate limit
drlwi = down rate limit
Figure 4-14. Simplified WindINERTIA Control Model.
Table 4-13. WindINERTIA Control Parameters.

Kwi kwi 10.
dbwi dbwi 0.0025
Tlpwi tlpwi 1.
Twowi twowi 5.5
urlwi urlwi 0.1
drlwi drlwi -1.0
Pmxwi pmxwi 0.1
Pmnwi pmnwi 0.0

5 Full Converter WTG Dynamic Models
This section presents the engineering assumptions, detailed structure, and data for
each of the component models necessary to represent a GE full converter WTG. Since
the grid performance of full converter WTGs, like the DFAG WTGs, is dominated by the
converter controls, the dynamic models described in Section 4 are the basis for the full
converter models. The modifications made to represent the full converter WTGs are
the focus of this section. Therefore, the common model features are not, in general,
discussed in this section.
5.1 Generator/Converter Model
This model (gewtg in PSLF with fcflg = 1) is the equivalent of the generator and the
full converter and provides the interface between the WTG and the network. The GE
DFAG WTG is represented by a flux and active current command. The GE full converter
WTG model differs in that it is represented by both reactive and active current
commands. The model is shown in Figure 5-1. Modified data parameters appropriate
for the full converter model are shown in Table 5-1.
Isorc
IQcmd 1
-1 High Voltage
(efd) 1+ 0.02s
Reactive Current
From s0 Management
ewtgfc
LVPL & rrpwr
Low Voltage
IPcmd 1 IPlv Active Current
(ladifd) 1+ 0.02s Management
From s1
ewtgfc
LVPL
Vterm
1.22
LVPL V
1
1+ 0.02s
V s2
zerox brkpt
(0.40) (0.90)
Low Voltage Power Logic
Figure 5-1. Full Converter WTG Generator/Converter Model.
Table 5-1. Full Converter Generator/Converter Parameters.

PSLF Parameter Recommended Value
fcflg 1
brkpt 0.9
zerox 0.4

5.2 Electrical (Converter) Control Model
This model (ewtgfc in PSLF) dictates the active and reactive power to be delivered
to the system based on inputs from the turbine model (Pord) and from the supervisory
VAr controller (Qord). Figure 5-2 shows the electrical control as modified to represent
the full converter WTG. The reactive power control, shown in Figure 4-5, remains
unchanged. Therefore, the WindCONTROL emulator (Section 4.2.1) can be used.
The primary philosophical change to the model was to generate a reactive current
command rather than a flux command. Additional functions include a dynamic
braking resistor and converter current limit. The open loop control logic available in
the DFAG electrical control model for representing old 1.5 MW WTGs is not included in
the control model for the full converter WTGs.
The objective of the dynamic braking resistor (DBR) is to minimize the WTG
response to large system disturbances, such as extended periods of low voltage. This
is accomplished by absorbing energy in the braking resistor when the power order is
significantly greater than the electrical power delivered to the grid. In this model, the
power order is compared to the actual electrical power to determine the power
absorbed by the braking resistor, Pdbr. This power is integrated to determine the
resulting energy absorbed by the braking resistor, Edbr. As long as that energy level is
less than the threshold, EBST, no other action occurs. When the energy level exceeds
the threshold, the resulting error signal is greater than zero and the amount of power
absorbed by the dynamic braking resistor is reduced. This ensures that the energy
capability of the resistor is respected. The model does not include thermal reset, so
simulations with multiple events may result in limited DBR response.
An auxiliary test signal can be injected into the terminal bus voltage regulator via
model[@index].sigval[0], as shown in Figure 5-2. A user-written dynamic model
(epcmod) is needed to generate the desired signal. The index of the wind generator
model (@index) can be obtained using the model_index function.
The details of the converter current limit are shown in Figure 5-3. The objective of
this function is to prevent the combination of the real and reactive currents from
exceeding converter capability. Depending upon the value of a user-specified flag,
pqflag, either real or reactive power has priority. This flag is dependent upon the
equipment features selected, and is normally dictated by the host system grid code.
When real power has priority, the real current order, IPcmd, is limited to the
minimum of the maximum temperature dependent converter current, ImaxTD, and a
hard active current limit, Iphl. The calculation of the limit on the reactive current begins
by determining the minimum of a hard reactive current limit, Iqhl, and a voltage
dependent limit, Iqmxv. The voltage dependent limit will be equal to the steady-state
rating of the wind plant (as defined by the input parameter Qmax) at 1.0 pu voltage and
will linearly increase as voltage drops. The maximum voltage dependent reactive
current limit is 1.6 pu at zero voltage. The minimum of Iqhl and Iqmxv is compared to the
remaining converter current capability, SQRT(ImaxTD2 - IPcmd2). That minimum is the

maximum (capacitive) limit, Iqmx, applied to the reactive current order, IQcmd. The
minimum (inductive) reactive current, Iqmn, is the negative of the maximum. No
minimum is applied to the real current order.
When reactive power has priority, the calculation of the limit on the reactive
current begins by determining the minimum of a hard reactive current limit, Iqhl, and
the voltage dependent limit, Iqmxv, as described above. The minimum of Iqhl and Iqmxv is
compared to a maximum temperature dependent converter current, ImaxTD. That
minimum is the maximum limit, Iqmx, applied to the reactive current order, IQcmd. The
minimum reactive current, Iqmn, is the negative of this maximum limit. The remaining
converter current capability, SQRT(ImaxTD2 – IQcmd2), becomes the maximum, Ipmx,
applied to the real current order, IPcmd. No minimum is applied to the real current
order. Reactive power priority is recommended, which is equivalent to the default
value of 0 for pqflag.
The preliminary values for the additional data associated with the converter
current limit and dynamic braking resistor are shown in Table 5-2. The converter
current limit, ImaxTD, is a function of time and operation. However, it is constant in this
model (1.7 pu) and not user-specified.
The bulk of the remaining data is unchanged from that used for a DFAG WTG. Data
associated with the reactive power control model that should be modified to correctly
represent a full converter WTG are shown in Table 5-3.

Auxiliary Test Signal
Qgen ( model[@index].sigval[0] )
Vmax Iqmx
- + IQcmd
Qcmd + Vref
KQi / s KVi / s
+ (efd)
s0 - s1
to Wind
Vterm Vmin Iqmn
Generator Model
P,Q Priority Flag

Converter Current Limit
Ipmx
Pord Porx . IPcmd
.
(vsig) (ladifd)
from Wind to Wind
Turbine Model Generator Model
Vterm
+ 1
Pelec Pdbr
-
from Wind + (elimt)
Generator
Pdlt - 0 to Wind
Model Turbine Model
Eerr
+
Kdbr 1/s
0 - Edbr s7
EBST
Dynamic Braking Resistor
Figure 5-2. Full Converter WTG Electrical Control Model.

P,Q Priority Flag
(pqflag)
0 1
Iqmn Iqmx Vt Iqmx Iqmn
Q Priority Iqmxv P Priority

1.6
qmax
Vt
-1 1.0 -1
Iqmxv
Iqhl Minimum
Minimum Minimum
IPcmd
ImaxTD ImaxTD2 - IPcmd2
IQcmd
ImaxTD2 - IQcmd2
Iphl
Minimum Minimum
Ipmx Ipmx
Figure 5-3. Converter Current Limit Model.
Table 5-2. Full Converter WTG Electrical Model.

Iphl (pu) iphl 1.24
Iqhl (pu) iqhl 1.25
pqflag pqflag 0=Q priority
EBST (pu) ebst 0.2
Kdbr kdbr 10.
Table 5-3. Full Converter WTG WindCONTROL Emulator Parameters.
KVi kvi 120.
Qmax (pu) qmax 0.40
Qmin (pu) qmin -0.40

5.3 Wind Turbine and Turbine Control Model
Only minor modifications were made to represent the additional functions of the
full converter wind turbine. Therefore, the same model is used in PSLF (wndtge) for
both DFAG and full converter WTGs. The modifications included the following:
• Incorporation of the dynamic braking resistor power, and
• Implementation of the no-wind (i.e., zero power) var capability.
The block diagram for the turbine model is unchanged, and is shown in Figure 4-8.
The signal representing the dynamic braking resistor power, Pdbr, is now added to
the electrical power as an input to both the rotor model and the creation of the speed
reference. When this model is used to represent the DFAG WTG, the dynamic braking
resistor power is automatically set to zero.
The converters used in both the DFAG and full converter machines make it possible
to inject or absorb reactive power (e.g., regulate voltage) at zero real power. However,
the reactive capability of the DFAG model is reduced while operating with the
WindFREE control. With a full converter WTG, the full reactive capability is retained
with this feature. Zero power may be the result of no wind, excessive wind, or an
operator directive to curtail output. All three of these zero power scenarios may be
simulated with this model by setting the wfflg parameter to 1.
The key limitation on zero-power operation is that once in that mode, the WTG
stays there. The model is neither applicable nor appropriate for simulating start-up
scenarios.
Three turbine control parameters are different from those used to represent a
DFAG WTG. Those parameters and their recommended values are shown in Table 5-4.
The remaining turbine control parameters can be used for the full converter without
change. Note that they are on the MW rating base of the turbine. The turbine-
generator mechanical model parameters for the full converter WTG are given in Table
5-5. While two-mass parameters are provided for the 2.5 MW full-converter machine,
GE does not recommend the two-mass model. Such a two mass representation in
simulations is not technically valid, because the converter isolates the machine from
the power system. Therefore, a single mass model is recommended. The same Cp
curves are used for the full converter WTG together with the wind power coefficients
shown in Table 5-6.
Table 5-4. Full Converter WTG Turbine Control Parameters (on Turbine MW Base).
Pwmin (pu) pwmin 0.
Kptrq kptrq 0.3
Kitrq kitrq 0.1

Table 5-5. Full Converter WTG Rotor Model Parameters (on Turbine MW Base).
One-mass Model
H h 4.18
Two-mass Model (Not Recommended)
H h 3.36
Hg hg 0.82
Ktg ktg 3.86
Dtg dtg 1.5
ωbase wbase 144
Table 5-6. Full Converter WTG Wind Power Coefficients.

Variable Name Recommended Value
½ρ Ar 0.00159
Kb 56.6

6 DFAG WTG Benchmark Simulations
The models described in this report have been implemented in GE’s PSLF load flow
and dynamic simulation software. The PSLF models have been validated by
comparison with more detailed simulation models and field measurements, as
described in Section 8. Representative results using the PSLF models are presented in
this section. Note that these simulations are not necessarily updated with each
version of this document. Therefore, the simulations may not always use the latest
model or data recommendations. They do, however, illustrate the general
performance characteristics of the DFAG wind plant model. The data used for these
simulations is shown in Section 6.3.
Upon request, GE will provide the PSLF benchmark simulation results to those who
wish to implement the models in other simulation programs. The results can be
supplied in ASCII format for cross-plotting in order to validate the model
implementation.
6.1 Test System
One line diagrams of the primary test system are shown in Figure 6-1. The top
diagram shows real and reactive power flow (MW, MVAr), and the bottom diagram
shows impedances (pu on 100 MVA). The test system represents an aggregate model
of a wind plant, suitable for analyzing the response of the wind plant to grid
disturbances. The many WTGs of the actual wind plant are combined into a single
WTG model, a single unit transformer, a single 34.5 kV feeder representing the collector
system, and a substation transformer to the system point of interconnection. For this
test system, the wind plant is rated 100.5 MW, representing the aggregation of 67
1.5 MW WTGs.
Figure 6-1. Primary DFAG Test System.

The dynamic data used for the benchmark simulations are listed in Section 6.3.
Note that the WindCONTROL emulator was on (varflg = 1) for all cases. The Q Droop
function was only on for the example showing its performance (Section 6.2.2).
Similarly, the Active Power Control and WindINERTIA functions were only on to illustrate
their performance in Sections 6.2.3 and 6.2.6, respectively.
6.2 DFAG WTG Benchmark Simulation Results
Plotted simulation results are provided for each example. Traces plotted together
generally share the same scaling, as shown on the y-axis. However, there are
exceptions. Scaling for a particular trace can be confirmed by checking the legend
below each plot. The same plot format was used for all simulations, and a brief
description follows.
From top to bottom, the left column shows the infinite bus voltage (dark blue line,
pu), WTG terminal voltage (red line, pu), and point of interconnection voltage (green
line, pu) in the first plot.
The second plot shows WTG real power generated (dark blue line, MW), reactive
power generated (red line, MVAr), and mechanical power (green line, MW).
The third plot shows variables associated with the active power control.
Specifically, it shows the total available power in the wind (Pavf, dark blue line, pu),
turbine power order from the torque control (Pinp, red line, pu), the output of the active
power control (Pstl, green line, pu), the power order after the active power control limit
is applied (Plim, light blue line, pu), the output of the washout in the power response
rate limit (Wsho, pink line, pu), and the final power order from the turbine to the
electrical control model (Pord, black line, pu).
The fourth and final plot in the left column shows the WTG terminal bus frequency
(dark blue line, Hz), and the flag from the active power control indicating a frequency
excursion (red line).
From top to bottom, the right column shows the WindCONTROL emulator voltage
reference (dark blue line, pu) and regulated voltage (red line, pu). In all cases, the
regulated bus was the point of interconnection.
The second plot shows the rotor speed (dark blue line, pu).
The third plot shows the blade pitch (dark blue line, deg) and wind speed (red line,
m/sec).
The fourth and final plot in the right column shows the real power current
command (Ipcmd, dark blue line, pu) and the reactive power voltage behind a
reactance command (E”qcmd, red line, pu).
Note that per unit values of real and reactive power are on the MVA base for
variables associated with the generator and electrical control models. Per unit values
of real and reactive power are on the MW base for variables associated with the
turbine model. Transfer variables between models (e.g., Pord from the turbine to the
electrical control) are on the MVA base of the receiving model.

6.2.1 Fault Response with WindCONTROL Emulator
Two simulations illustrate WTG response to grid disturbances. One disturbance is a
series of 3-phase bus faults with various fault impedances, the other is a 3-phase fault
to ground cleared by tripping one of the 230 kV lines from the POI bus to the Infinite
bus. The results are shown in Figure 6-2 and Figure 6-3, respectively.
In both examples, the reactive power output of the WTG is at or near its maximum
during the fault in an effort to regulate voltage. After each fault is removed, the
WindCONTROL emulator quickly restores the voltage at the POI bus to its initial value.

Figure 6-2. Series of Bus Faults with Various Fault Impedances.

Figure 6-3. 3-phase Fault to Ground, Cleared by Tripping 230 kV Line.

6.2.2 Q Droop Function Performance
The Q Droop function is presented in this section. This is the only case for which
that function was active, i.e., the gain parameter Kqd was non-zero. Figure 6-4 shows
the Q Droop response to a capacitor bank switching event.
Initially, the wind plant is regulating the POI voltage to approximately 1.03 pu. The
Q Droop function is using reactive power at the generator terminals as its input. At 1
second, a 30 MVAr capacitor is switched on. The reactive power output from the
generator decreases to maintain the regulated bus voltage. The output of the Q Droop
function, shown in the upper right plot as the green trace (Vqd), decreases with the
reactive power output. Therefore, the effective reference voltage decreases. This
results in a higher final POI bus voltage, shown in the upper right plot as the red trace
(Vreg), of about 1.035 pu.

Figure 6-4. Q Droop Response to Capacitor Switching.

6.2.3 Active Power Control and Power Response Rate Limit Performance
The Active Power Control (APC) and power response rate limit performance are
presented in this section. This is the only case for which the APC was active, i.e., the
input parameter apcflg was set to 1.
Figure 6-5 shows the APC’s response to a loss of load. Initially, the wind plant is
constrained to 0.95 pu of the power available in the wind. Therefore, the blade pitch is
non-zero.
At 1 second, 100 MW of load is tripped. As a result, system frequency increases.
Shortly thereafter, the frequency exceeds the first high frequency threshold (Fc =
1.004 pu or 60.24 Hz) in the APC’s frequency response curve. This defines a frequency
excursion event, and the frequency excursion flag is set to 1. The output of the APC
function, which is the upper limit imposed on the power order from the turbine, begins
to decrease (Pstl, left column, third plot, green line). The washout in the power
response rate limit (Wsho, pink line) transiently allows the higher power order from the
turbine control (Pinp, red line) through to the final power order (Pord, black line).
At about 8 seconds, the washout settles to near zero, and the final power order is
equal to the output of the APC. Between about 12 and 14 seconds, the frequency is
below the high frequency threshold for APC operation. However, this is only
temporary. The frequency rises a bit and the wind plant APC continues to operate.
Throughout the simulation, the large machine on the Infinite bus is attempting to
control frequency by reducing its output. By the end of the simulation, the frequency
has settled to about 60.3 Hz and the APC output to about 0.92 pu in accordance with
the frequency response curve characteristic.

Figure 6-5. Active Power Control Response to Loss of Load.

6.2.4 Wind Speed Profile
Wind plant response to a wind speed profile is shown in Figure 6-6. The initial wind
speed of about 8.5 m/sec results in a power output of about 50 MW, and a blade pitch
of 0 degrees.
Starting at 5 seconds, the wind speed increases over the next 40 seconds to a
maximum of 15 m/sec. The power output follows the wind speed up to the maximum
output of 100 MW. The blade pitch increases to shed the excess power available in the
wind. The reactive power output increases to maintain the POI voltage with the higher
power output.
At about 65 seconds, the wind speed starts to fall, with the power output following.
The reactive power also decreases to regulate POI voltage with the lower real power
output.
At about 105 seconds, the wind speed falls below the minimum wind speed
threshold, which is a non-user-specified value of 3 m/sec. The rotor speed continues
to fall until it reaches 0.1 pu, which is another non-user-specified threshold. At this
point, about 130 seconds, the WTG models are tripped. Real and reactive power
output go to zero, and the POI voltage is no longer regulated.
6.2.5 Zero Power (WindFREE) Operation

The above wind profile is used to test zero power operation, as shown in Figure 6-7.
This is the only case for which that function was active, i.e., the flag wfflg was non-zero.
Again, the initial wind speed of about 8.5 m/sec results in a power output of about
50 MW, and a blade pitch of 0 degrees.
The simulation is identical to that shown in Figure 6-6 until the WTGs trip at about
130 seconds. At this point, the real power output goes to zero, but the reactive power
output does not. The wind plant is still regulating the POI voltage. However, the
reactive power capability of the plant has been reduced, since the DFAG machine does
not use a full-sized converter. Thus, the POI voltage is not perfectly regulated. The
voltage performance is, however, better than that achieved without the WindFREE
capability.

Figure 6-6. Response to Wind Speed Profile without Zero Power Operation.

Figure 6-7. Response to Wind Speed Profile with Zero Power Operation.

6.2.6 WindINERTIA Control Performance
WindINERTIA control performance is presented in this section. This is the only case
for which that function was active, i.e., the gain parameter Kwi was non-zero.
The control was tested using the auxiliary signal shown in Figure 4-14. This test
signal is added to the constant bus frequency, and results in the frequency input to the
WindINERTIA control as shown in the bottom, left plot of Figure 6-8. This frequency
input signal falls below 59 Hz. The output of the WindINERTIA control is shown by the
pink trace in the third plot on the left side. The control output reaches 0.10 pu, and
lasts about 15 seconds. This results in an increase in wind plant power output of up to
10%, as shown in the second plot on the left side. There is little impact on rotor speed
because the WTGs are able to provide additional power by decreasing blade pitch. The
incremental energy delivered to the system is drawn from the energy available in the
wind rather than from the rotating inertia of the machine.
At lower wind speeds, the incremental energy delivered by WindINERTIA would
result in a drop in rotor speed, and a power back swing as the energy needed to
reaccelerate the rotor is recovered from the system.

Figure 6-8. Response to Frequency Signal with WindINERTIA.

6.3 DFAG WTG Benchmark Simulation Dynamic Data
The generator, electrical control and turbine control dynamic data used in the
benchmark simulations are shown in Table 6-1, Table 6-2, and Table 6-3, respectively.
Table 6-1. DFAG Generator Model (gewtg) Data for Simulations.
Model Parameter 1.5 MW WTG
MVA 111.
lpp 0.80
dvtrp1 -0.25
dvtrp2 -0.50
dvtrp3 -0.70
dvtrp4 -0.85
dvtrp5 0.10
dvtrp6 0.15
dttrp1 1.90
dttrp2 1.20
dttrp3 0.70
dttrp4 0.20
dttrp5 1.00
dttrp6 0.10
fcflg 0
rrpwr 10.0
brkpt 0.90
zerox 0.50

Table 6-2. DFAG Electrical Control Model (exwtge) Data for Simulations.
Model
Model Parameter 1.5 MW WTG 1.5 MW WTG
Parameter
varflg 1 ql1 0
kqi 0.10 ql2 0
kvi 40. ql3 0
vmax 1.10 qh1 0
vmin 0.90 qh2 0
qmax 0.436 qh3 0
qmin -0.436 pfaflg 0
xiqmax 1.45 fn 1.0
xiqmin 0.50 tv 0.05
tr 0.02 tpwr 0.05
tc 0.15 ipmax 1.22
kpv 18. xc 0
kiv 5. kqd 0 or 0.04*
vl1 -9999 tlpqd 5.0
vh1 9999 xqd 0.0
tl1 0 vermn -0.1
tl2 0 vermx 0.1
th1 0 vfrz 0.7
th2 0
* Different values for different simulations

Table 6-3. DFAG Turbine Control Model (wndtge) Data for Simulations.
Model Model
1.5 MW WTG 1.5 MW WTG
Parameter Parameter
mwcap 100.5 tw 1.0
usize 1.5 apcflg 0 or 1*
spdw1 14.0 tpav 0.15
tp 0.30 pa 1.0
tpc 0.05 pbc 0.95
kpp 150. pd 0.40
kip 25. fa 0.96
kptrq 3.0 fb 0.996
kitrq 0.60 fc 1.004
kpc 3.0 fd 1.04
kic 30. pmax 1.0
pimax 27. pmin 0.20
pimin 0. kwi 0 or 10.*
pirat 10. dbwi 0.0025
pwmax 1.12 tlpwi 1.
pwmin 0.04 twowi 5.5
pwrat 0.45 urlwi 0.1
h 4.94 drlwi -1.0
nmass 1 pmxwi 0.1
hg - pmnwi 0.0
ktg - wfflg 0 or 1*
dtg -
wbase -

7 Full Converter WTG Benchmark Simulations
The models described in this report have been implemented in GE’s PSLF load flow
and dynamic simulation software. The PSLF models have been validated by
comparison with more detailed simulation models, as described in Section 8.
Representative results using the PSLF models are presented in this section. These
simulations focus on the capabilities of this type of machine, rather than repeating the
tests performed for the DFAG benchmark simulations. Note that these simulations are
not necessarily updated with each version of this document. Therefore, the
simulations may not always use the latest model or data recommendations. They do,
however, illustrate the general performance characteristics of the full converter wind
plant model. The data used for these simulations is shown in Section 7.3.
Upon request, GE will provide the PSLF benchmark simulation results to those who
wish to implement the models in other simulation programs. The results can be
supplied in ASCII format for cross-plotting in order to validate the model
implementation.
7.1 Test System
The test system was similar to that used for the 1.5 MW DFAG simulations, which is
described in Section 6.1. The wind plant was rated at 100 MW, and consisted of 40
2.5 MW full converter WTGs. Minor modifications were made to the wind plant power
output and system voltage profile as necessary to test the various functions.
7.2 Full Converter WTG Benchmark Simulation Results
As noted above, the test simulations described in this section focus on the
capabilities of the full converter machine, rather than repeating the tests performed for
the DFAG benchmark simulations. Specifically, the following subsections will describe
simulation results illustrating the performance of the zero-power operation capability,
converter current limit, low voltage power logic, and dynamic braking resistor. A
variety of disturbances are represented – fault events, wind profiles, and control input
changes.
Plotted results are provided for each test simulation. Traces plotted together
generally share the same scaling, as shown on the y-axis. However, there are
exceptions. Scaling for a particular trace can be confirmed by checking the legend
below each plot. The same plot format was used for most simulations, and a brief
description follows.
From top to bottom, the left column shows the infinite bus voltage (dark blue line,
pu), WTG terminal voltage (red line, pu), and point of interconnection voltage (green
line, pu) in the first plot.
The second plot shows WTG real power generated (dark blue line, MW), reactive
power generated (red line, MVAr), and mechanical power (green line, MW).
The third plot shows the final power order from the turbine to the electrical control
model (Pord, dark blue line, pu).

The fourth and final plot in the left column shows variables associated with the
dynamic braking resistor. Specifically, this plot shows the difference between electrical
power and the power order (Pdlt, dark blue line, pu), the power absorbed by the
dynamic braking resistor (Pdbr, red line, pu), the energy absorbed by the dynamic
braking resistor (Edbr, green line, pu), and the difference between the energy absorbed
by the dynamic brake and the energy threshold (Eerr, light blue line, pu).
From top to bottom, the right column shows the WindCONTROL emulator voltage
reference (dark blue line, pu) and regulated voltage (red line, pu). In all cases, the
regulated bus was the point of interconnection.
The second plot shows the rotor speed (dark blue line, pu).
The third plot shows the blade pitch (dark blue line, deg) and wind speed (red line,
m/sec).
The fourth and final plot in the right column shows the real power (Ipcmd, dark blue
line, pu) and reactive power (Iqcmd, red line, pu) current commands from the electrical
control to the generator. Three converter current limits are also plotted: the maximum
real current (Ipmx, green line, pu), the maximum reactive current (Iqmx, light blue, pu),
the voltage dependent reactive current limit (Iqxv, pink line, pu). Finally, the real power
current command after the low voltage power logic (Iplv, black line, pu) is shown.
Note that per unit values of real and reactive power are on the MVA base for
variables associated with the generator and electrical control models. Per unit values
of real and reactive power are on the MW base for variables associated with the
turbine model. Transfer variables between models (e.g., Pord from the turbine to the
electrical control) are on the MVA base of the receiving model.
7.2.1 Fault Response with WindCONTROL Emulator
Two simulations illustrate WTG response to grid disturbances. One disturbance is a
series of 3-phase bus faults with various fault impedances, the other is a 3-phase fault
to ground cleared by tripping one of the 230 kV lines from the POI bus to the Infinite
bus. The results are shown in Figure 7-1 and Figure 7-2, respectively.
In both examples, the reactive power output of the WTG is at or near its maximum
during the fault in an effort to regulate voltage. After each fault is removed, the
WindCONTROL emulator quickly restores the voltage at the POI bus to its initial value.

Figure 7-1. Series of Bus Faults with Various Fault Impedances.

Figure 7-2. 3-phase Fault to Ground, Cleared by Tripping 230 kV Line.

7.2.2 Zero-Power Operation
Three scenarios involving zero-power operation of the full converter WTG are
illustrated in this section.
Figure 7-3 shows the WTG’s response to increasing wind speed. As wind speed
increases, the blade pitch increases in an effort to maintain the desired power output.
At about 19 seconds, the wind speed exceeds the high speed threshold, 25 m/sec, and
the inverse time tripping characteristic starts incrementing. At about 24 seconds, the
tripping threshold is reached and the turbine is tripped. The rotor speed, electrical and
mechanical power are all zero at this point. With WindFREE, however, the full
converter WTG continues to generate reactive power in order to regulate voltage at
the POI bus.
Figure 7-4 shows the WTG’s response to the same wind profile use in Section 6.2.5.
Starting at 5 seconds, the wind speed increases over the next 40 seconds to a
maximum of 15 m/sec. The power output follows the wind speed up to the maximum
output of 100 MW. The blade pitch increases to shed the excess power available in the
wind. The reactive power output increases to maintain the POI voltage with the higher
power output.
At about 65 seconds, the wind speed starts to fall, with the power output following.
The reactive power also decreases to regulate POI voltage with the lower real power
output.
At about 105 seconds, the wind speed falls below the minimum wind speed
threshold, which is a non-user-specified value of 3 m/sec. At about 120 seconds, the
electrical and mechanical power go to zero, but the reactive power output does not.
The wind plant is still regulating the POI voltage.
The final example of zero-power operation is shown in Figure 7-5. This simulation
used a modified test system with an initial wind plant power output of 0 MW. At 1
second, the POI bus voltage schedule is increased by 0.03 pu. While real power output
remains zero, reactive power is increased to regulate the POI bus voltage to the new
reference. At 5 seconds, the voltage schedule is decreased by 0.03 pu. Again, the WTG
regulates the voltage with zero real power output.

Figure 7-3. Increasing Wind Speed Results in Zero-Power Operation.

Figure 7-4. Response to Wind Speed Profile with Zero Power Operation.

Figure 7-5. Voltage Regulation in Continuous Zero-Power Operation.

7.2.3 Converter Current Limit Performance
The performance of the converter current limit function is illustrated in this section.
In general, the converter current capability will not be limiting except for severe system
disturbances. Therefore, an arbitrary step was applied to one of the hardwired current
limits, ImaxTD, as a test. This test can not be replicated with the public version of PSLF.
Realistic limits and the converter current limit function’s response to a system fault
were shown in Section 7.2.1.
Figure 7-6 shows WTG response to a reduction in the converter current limit, ImaxTD,
with real power or P priority. The WTG was initially operating at about 0.83 pu of
generator MVA base, equivalent to 1.0 pu on the turbine MW base. At 1 second, the
current limit was reduced from 1.7 pu to 0.65 pu. As a result the real power command
is reduced to 0.65 pu. Due to the P priority, the reactive power command is reduced to
zero. The converter current limit was returned to its initial value of 1.7 pu at 5 seconds.

Figure 7-6. Step Reduction in Converter Current Limit with P Priority.

7.2.4 Low Voltage Power Logic Performance
The performance of the low voltage power logic (LVPL) is illustrated in this section.
The rate of post-fault power recovery in the LVPL is governed by the value of the
generator input data parameter rrpwr. The recommended value for this parameter is
10. However, a value of 3 was used in this simulation for a clearer illustration of the
LVPL’s performance.
Figure 7-7 shows WTG response to a long, depressed voltage due to a fault on the
POI bus. The fault is applied at 1 second, reducing POI bus voltage to near 0 pu for the
next 0.5 seconds. The LVPL becomes active shortly after fault application when the
terminal bus voltage falls below 0.90 pu. As a result, the real power of the generator is
reduced the voltage dependent cap calculated by the LVPL. After the fault is removed,
the terminal voltage returns to its pre-fault level. However, the recovery of the real
power is ramp rate limited. The chosen rrpwr value results in a 0.2 second recovery
time. Thus the generated power reaches its pre-fault level at about 1.4 seconds.
7.2.5 Dynamic Braking Resistor Performance
The performance of the dynamic braking resistor is also illustrated by the test
simulation shown in Figure 7-7.
As described above, the fault induced voltage depression causes the low voltage
power logic to operate. This, in turn, results in a difference between the power order
and the electrical power delivered to the system which is dissipated in the dynamic
braking resistor. At about 1.3 seconds, the energy absorbed by the resistor exceeds
the specified threshold, EBST, of 0.2 pu. From that point forward, the rate at which
energy is absorbed into the braking resistor is reduced to minimize energy absorption
in excess of the capability.

Figure 7-7. Low Voltage Power Logic Response to Fault on POI Bus.

7.3 Full Converter WTG Benchmark Simulation Dynamic Data
The generator, electrical control and turbine control dynamic data used in the
benchmark simulations are shown in Table 7-1, Table 7-2, and Table 7-3, respectively.
Any data changes required to test specific functions were noted in the simulation
descriptions provided in Section 7.2.
Table 7-1. Full Converter WTG Generator Model Data for Simulations.
MVA 120.
lpp 0.80
dvtrp1 -0.25
dvtrp2 -0.50
dvtrp3 -0.70
dvtrp4 -0.85
dvtrp5 0.10
dvtrp6 0.15
dttrp1 1.90
dttrp2 1.20
dttrp3 0.70
dttrp4 0.20
dttrp5 1.00
dttrp6 0.10
fcflg 1
rrpwr 10
brkpt 0.9
zerox 0.4

Table 7-2. Full Converter WTG Electrical Control Model Data for Simulations.
varflg 1
kqi 0.10
kvi 120.
vmax 1.10
vmin 0.90
qmax 0.40
qmin -0.40
tr 0.02
tc 0.15
kpv 18.
kiv 5.
pfaflg 0
fn 1.0
tv 0.05
tpwr 0.05
iphl 1.24
iqhl 1.25
pqflag 0 or 1*
kdbr 10.
ebst 0.2
xc 0.0
kqd 0.0
tlpqd 5.0
xqd 0.0
vermn -0.1
vermx 0.1
vfrz 0.7

Table 7-3. Full Converter WTG Turbine Control Model Data for Simulations.
Model
Model Parameter 2.5 MW WTG 2.5 MW WTG
Parameter
mwcap 100. tw 1.0
usize 2.5 apcflg 0
spdw1 14.0 tpav 0.15
tp 0.30 pa 1.0
tpc 0.05 pbc 0.95
kpp 150. pd 0.40
kip 25. fa 0.96
kptrq 0.3 fb 0.996
kitrq 0.1 fc 1.004
kpc 3.0 fd 1.04
kic 30. pmax 1.0
pimax 27. pmin 0.2
pimin 0. kwi 0
pirat 10. dbwi 0.0025
pwmax 1.12 tlpwi 1.02
pwmin 0.0 twowi 5.5
pwrat 0.45 urlwi 0.1
h 4.18 drlwi -1.0
nmass 1 pmxwi 0.1
hg - pmnwi 0.
ktg - wfflg 1
dtg -
wbase -

8 PSLF Model Validation
Performance of both the PSLF fundamental frequency 1.5 MW DFAG and 2.5 MW
full converter WTG models is compared to that of the detailed cycle-by-cycle EMTP
type models incorporated in GE’s Windtrap program, and to field measurements.
8.1 1.5 MW DFAG Comparison to Windtrap
The performance of GE 1.5 MW WTGs as modeled in PSLF and Windtrap is
compared to provide validation of the PSLF model.
The study system, shown in Figure 8-1, was based on one developed for the WECC
generic WTG modeling project. A single machine equivalent of a 100 MW wind plant is
connected via appropriate transformers and a collector system to a 230 kV equivalent
system. Figure 8-1 shows the weak system with MW and MVAr flows in the upper
diagram and pu impedances in the lower diagram. The point of interconnection (POI) is
defined as bus 2.
Figure 8-1. Test System for PSLF and Windtrap Comparison.
Several test disturbances were applied to this system. However, the results of only
one are described here. This disturbance conted of a 150 msec, 3-phase bus fault
through an impedance at the POI (bus 2).
The PSLF 1.5 MW DFAG model, as described in this document, and the Windtrap
1.5 MW WTG model with the most recent GE controller (WCNTRLFLAG = 33) were used
for this comparison.
Cross plots of PSLF (blue line) and Windtrap (pink line) simulation results are shown
in Figure 8-2 through Figure 8-4. The first figure shows terminal voltage (pu), the
second shows real power output from the WTG (MW), and the third shows reactive
power output (MVAr).

In all figures, the PSLF response matches the Windtrap response in the frequency
range of interest. Thus the PSLF model is an appropriate representation of the GE
1.5 MW WTG’s behavior for fundamental frequency analysis.
1.50
Windtrap
PSLF
1.00
Terminal Voltage (pu)
0.50
0.00
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-2. 1.5 MW DFAG PSLF and Windtrap Terminal Voltage Response.

3.0
Windtrap
PSLF
2.0
Single WTG Real Power (MW)
1.0
0.0
-1.0
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-3. 1.5 MW DFAG PSLF and Windtrap Real Power Response.
3.0
Windtrap
PSLF
2.0
Single WTG Reactive Power (MVAr)
1.0
0.0
-1.0
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-4. 1.5 MW DFAG PSLF and Windtrap Reactive Power Response.

8.2 2.5 MW Full Converter Comparison to Windtrap
The performance of GE 2.5 MW full converter WTGs as modeled in PSLF and
Windtrap is compared to provide validation of the PSLF model. The test system was
similar to that used for the 1.5 MW DFAG comparison, which is described in Section 8.1.
The wind plant was rated at 100 MW, and consisted of 40 2.5 MW full converter WTGs.
Several test disturbances were applied to this system. However, the results of only
one are described here. This disturbance consisted of a 150 msec, 3-phase fault to
ground at the POI (bus 2).
The PSLF 2.5 MW model, as described in this document, and the Windtrap 2.5 MW
WTG model with the most recent GE controller (WCNTRLFLAG = 4) were used for this
comparison.
Cross plots of PSLF (blue line) and Windtrap (pink line) simulation results are shown
in Figure 8-5 through Figure 8-7. The first figure shows terminal voltage (pu), the
second shows real power output from the WTG (MW), and the third shows reactive
power output (MVAr).
In all figures, the PSLF response matches the Windtrap response in the frequency
range of interest. Thus the PSLF model is an appropriate representation of the GE
2.5 MW WTG’s behavior for fundamental frequency analysis.
The low frequency oscillations in the Windtrap results are merely the result of
sampling aliasing.
1.50
Windtrap
PSLF
1.00
Terminal Voltage (pu)
0.50
0.00
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-5. 2.5 MW Full Converter PSLF and Windtrap Terminal Voltage Response.

4.0
Windtrap
PSLF
3.0
Single WTG Real Power (MW)
2.0
1.0
0.0
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-6. 2.5 MW Full Converter PSLF and Windtrap Real Power Response.
3.0
Windtrap
PSLF
2.0
Single WTG Reactive Power (MVAr)
1.0
0.0
-1.0
0 0.5 1 1.5 2
Time (Seconds)
Figure 8-7. 2.5 MW Full Converter PSLF and Windtrap Reactive Power Response.

8.3 WindINERTIA Model Comparison to Field Measurements
As noted in Section 4.3.5, an auxiliary signal can be used to inject any desired test
function into the WindINERTIA model. This is also true of the actual control. The test
frequency signal shown in Figure 8-8 was used to test both the WindINERTIA model
and the actual control in an operating WTG. At the wind turbine, this signal was
injected into the WindINERTIA function periodically under a wide variety of wind
conditions. Hence, a large amount of field results was acquired. These results were
aggregated by wind speed and averaged. Comparisons of model and field test results
for three different wind speeds (8 m/s, 10 m/s, 14 m/s) are shown in Figure 8-9. These
results indicate that the simple WindINERTIA model is an appropriate representation of
the actual, and very complicated, turbine and WindINERTIA control response.
Figure 8-8. Frequency Test Signal.

1800
1500
1200
Power (kW)
900
600
8 m/s Avg Meas 10 m/s Avg Meas 14 m/s Avg Meas

300
8 m/s PSLF 10 m/s PSLF 14 m/s PSLF
0
0 10 20 30 40 50 60 70 80 90
Time (Seconds)
Figure 8-9. WindINERTIA Model Performance and Field Measurements.

9 Conclusions
The wind turbine model presented in this report is based on presently available
design information, test data and extensive engineering judgment. The modeling of
wind turbine generators for bulk power system performance studies is still in a state of
rapid evolution, and is the focus of intense activity in many parts of the industry. More
important, the GE equipment is being continuously improved, to provide better
dynamic performance. These ongoing improvements necessitate continuing changes
and improvements to these models. This model is expected to give realistic and
correct results when used for bulk system performance studies. It is expected that as
experience and additional hard test data is obtained, these models will continue to
evolve, in terms of parameter values and structure.
This document is continuously being updated to reflect these changes. Those
using this document for modeling purposes are encouraged to verify that they are
using the most up-to-date version. This document is available through the GE Wind
and PSLF software websites.

GE Wind Turbine

Uploaded by

Copyright:

Available Formats

GE Wind Turbine

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GE Wind Turbine

Uploaded by

Copyright:

Available Formats

Modeling of GE Wind

April 16, 2010

General Electric International, Inc.

This document was prepared by GE Energy in Schenectady, NY. Technical and

2 MODEL OVERVIEW AND PHILOSOPHY .............................................................................2.1

GE Energy 1.1 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 2.1 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 2.2 GE WTG Modeling-v4.5.doc, 4/16/10

3φ AC Winding Q stator fnet

Wind Turbine f stator

Figure 2-2. GE Full Converter WTG Major Components.

Like the GE DFAG machines, the fundamental frequency electrical dynamic

GE Energy 2.3 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 2.4 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 2-3. GE WTG Dynamic Model Connectivity.

GE Energy 2.5 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 3.1 GE WTG Modeling-v4.5.doc, 4/16/10

Point of High Side Bus

Figure 3-1. Simplified Wind Plant Power Flow Model.

Table 3-1. Individual WTG Power Flow Data.

The WindCONTROL system is structured to measure the voltage at a particular bus,

GE Energy 3.2 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 3.3 GE WTG Modeling-v4.5.doc, 4/16/10

4.1 Generator/Converter Model

GE Energy 4.1 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 4-1. DFAG Generator/Converter Model.

4.1.1 Fault Ride Through

GE Energy 4.2 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.3 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.4 GE WTG Modeling-v4.5.doc, 4/16/10

Table 4-2. DFAG Generator/Converter Parameters.

GE Energy 4.5 GE WTG Modeling-v4.5.doc, 4/16/10

From Qord Open Qcmd

4.2.1 Reactive Power Control

GE Energy 4.6 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 4-5. Reactive Power Control Model.

GE Energy 4.7 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.8 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 4-6. Q Droop Function Model.

Table 4-4. Q Droop Function Parameters.

4.2.2 Electrical Control

GE Energy 4.9 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 4-7. DFAG Electrical Control Model.

Table 4-5. DFAG Electrical Control Parameters.

4.2.3 Control Strategies

GE Energy 4.10 GE WTG Modeling-v4.5.doc, 4/16/10

4.2.4 Open Loop Control Logic

GE Energy 4.11 GE WTG Modeling-v4.5.doc, 4/16/10

Table 4-7. Open Loop Reactive Power Control Parameters.

GE Energy 4.12 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.13 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.14 GE WTG Modeling-v4.5.doc, 4/16/10

Anti-windup on Pwmin& d P /dtmin

Figure 4-8. Wind Turbine Model Block Diagram.

4.3.1 Turbine Control Model

GE Energy 4.15 GE WTG Modeling-v4.5.doc, 4/16/10

Figure 4-9. Pitch Control and Pitch Compensation Block Diagram.

GE Energy 4.16 GE WTG Modeling-v4.5.doc, 4/16/10

GE Energy 4.17 GE WTG Modeling-v4.5.doc, 4/16/10