Document name WECC Second Generation Wind Turbine Models

Category ( ) Regional Reliability Standard

( ) Regional Criteria
( ) Policy
( ) Guideline
(X) Report or other
( ) Charter
Document date January 23, 2014
Adopted/approved TSS
by
Date January 23, 2014
adopted/approved
Custodian (entity M&VWG
responsible for
maintenance and
upkeep)
Stored/filed Physical location:
Web URL:
http://www.wecc.biz/library/WECC%20Documents/Documents
%20for%20Generators/WECC%20Second%20Generation%20
Wind%20Turbine%20Models%20012314.pdf
Previous (if any)
name/number
Status (X) in effect
( ) usable, minor formatting/editing required
( ) modification needed
( ) superseded by _____________________
( ) other _____________________________
( ) obsolete/archived)
Specification of the Second Generation Generic
Models for Wind Turbine Generators
Specification of the Second Generation Generic Models for
Wind Turbine Generators
Prepared under Subcontract No. NFT-1-11342-01 with NREL

Issued: 9/20/13 (revised 9/2713, 10/1/13; 10/7/13; 10/8/13, 11/11/13)

Prepared by:
Pouyan Pourbeik
942 Corridor Park Boulevard
Knoxville, TN 37932
USA
Ph: (919) 794 7204
[email protected]

ELECTRIC POWER RESEARCH INSTITUTE

3420 Hillview Avenue, Palo Alto, California 94304-1395 ▪ PO Box 10412, Palo Alto, California 94303-0813 ▪ USA
800.313.3774 ▪ 650.855.2121 ▪ [email protected] ▪ www.epri.com
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI) FOR
THE NATIONAL RENEWABLE ENERGY LABORATORY (NREL). NEITHER EPRI, NREL, ANY MEMBER OF
EPRI, ANY COSPONSOR, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH
RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM
DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED
RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS
SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING
ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE OR NREL OR ANY
NREL REPRENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING
FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.

Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.

Copyright © 2013 Electric Power Research Institute, Inc. All rights reserved.

iii
 AKNOWLEDGMENTS
EPRI wishes to acknowledge the Western Electricity Coordinating Council (WECC)
Renewable Energy Modeling Task Force and the International Electrotechnical
Commission Technical Committee 88, Working Group 27, and all the members of these
groups, as forums through which the author was able to participate as an active member
to contribute to the development of the models described herein. The comments,
feedback, support and encouragement of these groups and their respective members are
gratefully acknowledged.

EPRI expresses its sincere gratitude also to ABB, Siemens and Vestas for sharing, under
non-disclosure agreements, data from their field measurements of their equipment which
significantly helped in this research effort to improve the generic wind turbine models.
In particular, the author is grateful to the following individuals for fruitful and insightful
discussions:
Babak Badrzadeh, Vestas Technology R&D (presently no longer with Vestas)
Nikolaus Moeller Goldenbaum, Siemens Wind Power
Slavomir Seman, ABB (presently no longer with ABB)

iv
 CONTENTS
1 INTRODUCTION .................................................................................................................... 1-1
1.1 Background .................................................................................................................... 1-1
2 TYPE 1 AND 2 WTG .............................................................................................................. 2-1
2.1 The Type 1 WTG............................................................................................................ 2-1
2.2 The Type 2 WTG............................................................................................................ 2-4
3 TYPE 3 AND 4 WTG .............................................................................................................. 3-1
3.1 Overview ........................................................................................................................ 3-1
3.2 REGC_A ........................................................................................................................ 3-2
3.3 REEC_A ......................................................................................................................... 3-3
3.4 WTGT_A ........................................................................................................................ 3-7
3.5 WTGAR_A ..................................................................................................................... 3-8
3.6 WTGPT_A ...................................................................................................................... 3-8
3.7 WTGTRQ_A ................................................................................................................... 3-9
3.8 REPC_A ....................................................................................................................... 3-11
3.9 Creating Type 3 and 4 WTG Models............................................................................ 3-12
4 EXAMPLE SIMULATION CASES .......................................................................................... 4-1
5 CONCLUSION AND SUMMARY ........................................................................................... 5-1
6 REFERENCES ....................................................................................................................... 6-1
A CONVERTER MODEL GRID INTERFACE .......................................................................... A-1
B CURRENT LIMIT LOGIC ...................................................................................................... B-1

v
1
INTRODUCTION
1.1 Background
There are presently two major industry groups working towards the development of generic
models for use in power system simulations for wind turbine generators – the Western Electricity
Coordinating Council (WECC) Renewable Energy Modeling Task Force (REMTF) and the
International Electrotechnical Commission (IEC) Technical Committee (TC) 88, Working Group
(WG) 27. In 2010, the North American Electric Reliability Corporation (NERC), Integration of
Variable Generation Task Force (IVGTF) Task 1-1 published a report [1] that outlines the need
for such generic models for variable generation technologies such as wind and solar-
photovoltaics (PV). The NERC IVGTF Task 1-1 document explains that the term “generic”
refers to a model that is standard, public and not specific to any vendor, so that it can be
parameterized in order to reasonably emulate the dynamic behavior of a wide range of
equipment. Furthermore, the NERC document, as well as working drafts of the documents from
WECC REMTF and IEC TC88 WG27, explains that the intended usage of these models is
primarily for power system stability analysis. Those documents also discuss the range in which
these models are expected to be valid and the models’ limitations. It is outside the scope of this
report to discuss such details.
As an active participant in these various industry groups, EPRI has been working closely with
these industry groups and several of the wind turbine generator manufacturers, as well as with
the National Renewable Energy Laboratory, to help in the process of both the development and
validation of these generic models.
In North America, much of this collaborative work culminated in the issuing of several reports in
2012 and 2013 ([2], [3] and [4]) that were then reviewed and collectively tested and approved by
the WECC REMTF and Modeling Validation Working Group (MVWG) in early 2013.
Subsequently the process started whereby the models are presently being implemented by the
various commercial software vendors in North America.
At the last several WECC REMTF and MVWG meetings (i.e. March, 2013 and June, 2013)
these reports and thus specifications were all approved and finalized. A few minor changes were
made based on feedback during the process of implementation and testing of the models by GE
and Siemens PTI. Thus, this document constitutes the final version of the specification for all
these models as they related to wind turbine generators (WTG).
Although not explicitly covered in this report, it should be noted that the proposed building
blocks for the type 4 WTG also forms the basis for the first generation utility scale photovoltaic
(PV) models, with a few simplifications. This is covered in a separate report.
For those who may be unfamiliar with the four main wind turbine generator technologies, they
are shown pictorially in Figure 1-1.

1-1
In general, the most commonly sold and installed technologies in today’s market (both in the US
and overseas) tend to be the type 3 and 4 units. All the major equipment vendors supply one or
both of these technologies. There are, however, large numbers of the type 1 and 2 units in-
service around the world, and so modeling them is also of importance. Some vendors do still
supply the type 1 and 2 turbines as well.
The EPRI report [5] gives a brief outline of the history of these model developments as well as
the issues identified with the first generation generic models and the various proposals discussed
in the WECC REMTF and IEC TC88 WG27 groups, which was the initiation for the
development of these second generation models. Here we do not delve into those detail
discussions, but rather jump straight to the present the new model specification.

To Grid
To Grid
To Grid

To Grid

Figure 1-1: The four many wind turbine technologies.

Finally, with respect to the naming convention of the models, after the June 2012 WECC
REMTF meeting it was agreed to change the names of these 2nd generation model modules in
order to make them truly generic and usable for any appropriate renewable generation. For
example, the wtgg would be called the regc_a (renewable energy generator/converter model),
etc. These changes are reflected here. Also, these models proposed here will have a version

1-2
number designated by “_a” to allow for future revisions. Need-less-to-say, it is almost inevitable
that changes will be required as the technology advances.

1-3
2
TYPE 1 AND 2 WTG
2.1 The Type 1 WTG
The type 1 WTG is a conventional squirrel-cage induction generator. There are several possible
variations [6], namely:
1. Fixed speed stall controlled type 1 WTGs. With stall design the blades of the turbine are
bolted to the hub at fixed angle, and aerodynamically designed to stall (and stop the
turbine) once wind speeds reach a certain level. Thus, these WTGs have no pitch control.
These units should be modeled simply as an induction generator for the purpose of power
system stability studies.
2. Fixed speed active-stall controlled type 1 WTG. With this design the turbine has pitch
control. At low wind speeds by changing the blade pitch the overall turbine efficiency
can be improved. At high wind speeds pitch control is used to better control the turbine.
For sudden increases in wind speed the blades can be pitched in the opposite direction in
order to force stalling quickly and bring the turbine to a stop. In this case a pitch
controller should be modeled.
Thus, the generic model for a type 1 WTG consists of three components:
1. Generator Model – this is a conventional induction generator model. The preference is to
use a two-cage model representing both transiency and sub-transiency. The state
equations for a two-cage induction machine model may be found in many references.
This is presently available in most of the commercial software platforms. In GE PSLFTM
and Siemens PTI PSS®E, this is the wt1g model.
2. Drive Train Model – this is the standard two-mass drive train model, and already
available in the standard commercial software platforms. Presently, for the type 1
generic WTG, this model is called the wt1t model. There is also the option of modeling
the drive train by a single lumped mass, if desired.
3. The Pitch Controller – this model is new for the 2nd generation generic models and
described below in more detail.

It is a known fact that many type 1 WTG with active-stall employ a scheme whereby the
mechanical power is ramped down and then back up following a major voltage-dip (e.g. nearby
transmission fault) when at or near rated power [6]. As discussed in [6], this is done to prevent
the turbine from accelerating away and going unstable. The aim of the new generic pitch-
controller model is to emulate this behavior. As shown in [7], through simulations comparing
detailed vendor specific PSCAD models the actual control behavior is dependent on several
factors:

1. the amount of over-speed of the turbine during the event,

2-1
 2. the magnitude of the voltage dip, and

3. the initial turbine power.

In the actual controls there is a combination of monitoring of the shaft acceleration following an
event together with the level of voltage dip. Furthermore, there are some variations in the
control between various vendors as shown in [7]. Following discussions at the last WECC
MVWG meeting is was agreed that the simpler version of this generic-pitch controller proposed
in [4] is acceptable for the purposes of large interconnected studies.

The model is shown below in Figure 2-1.

A single integrator is used to ramp mechanical power down and back up. The rate limit
parameter rmin together with Pmin can then be used to effect the rate at which mechanical power
is reduced and to what value during the disturbance. The rate limit rmax determines how quickly
power is ramped back up after a given duration T. The time duration T, during which
mechanical power is ramped down is based solely on voltage and determined from a four-piece
curve (see table below). This is a very simple model and does not in any way represent actual
controls, but it allows for an emulation of the behavior of typical type 1 WTG active-stall pitch
control systems. The switch is automatically toggled by the model (Flag1) based on the
following principles:

1. If Vt (after filtering) < vt4 (last point on V/T curve) and Po ≥ Pset then Flag1 = 1, and
remains in this position for the duration T seconds.

2. Otherwise, Flag1 = 0.

Figure 2-1: Overall model structure for new proposed type 1 (and 2) wind turbine pitch controller –
WT1P_B.

The parameter list for this model is provided below. The user should realize that this model is a
simplified model for the purpose of emulating the general behavior of type 1 WTGs during
electrical disturbances. It does not claim to in anyway represent the actual control strategy
associated with the pitch control of such turbines.

2-2
Parameter Description Typical Range Units
of Values
Tr Voltage measurement time constant N/A s
rmax Rate limit for increasing power N/A MWs/MVA
rmin Rate limit for decreasing power N/A MWs/MVA
T1 Lag time constant N/A pu
Pmin Minimum power setting N/A pu
Po Initial turbine mechanical power (Initialized by N/A pu
the model from power flow – not user defined)
Flag = 1 if Vt < vt4 & Po ≥ Pset, else N/A N/A
=0
not user defined, set by model
Pset If Po <= Pset then ramp power N/A pu
F(Vt)= N/A [pu,s]
[vt1,t1;
vt2,t2;
vt3,t3;
vt4,t4]

This is the look-up table that determines the duration

in seconds of the power ramping after a dip based
on the size of the voltage dip.

2-3
2.2 The Type 2 WTG
The type 2 WTG is a wound-rotor induction generator, with a variable resistor in the rotor circuit
which is typically controlled using power electronics [6]. Typically, type 2 WTGs employed
pitch control and so a pitch controller should be modeled.
Thus, the generic model for a type 2 WTG consists of three components:
1. Generator Model – this is a conventional induction generator model, including access to
the external rotor resistance variable. This is presently available in most of the
commercial software platforms. In GE PSLFTM and Siemens PTI PSS®E, this is the
wt2g model.
2. External Resistance Controller – a simple model of the external resistance controller.
This model already exists in the GE PSLFTM and Siemens PTI PSS®E programs as wt2e.
3. Drive Train Model – this is the standard two-mass drive train, and already available in the
standard commercial software platforms. Presently, for the type 2 generic WTG, this
model is called the wt2t model. There is also the option of modeling the drive train by a
single lumped mass, if desired.
4. The Pitch Controller – this model is new for the 2nd generation generic models and is as
described above, i.e. wt1p_b.

2-4
3
TYPE 3 AND 4 WTG
3.1 Overview
The 2nd generation type 3 and 4 models are built up of several generic modules that are put
together to either constitute a type 3 or 4 WTG. There are seven (7) modules, or building block
models. These are:
1. The renewable energy generator/converter model (regc_a), which has inputs of real
(Ipcmd) and reactive (Iqcmd) current command and outputs of real (Ip) and reactive (Iq)
current injection into the grid model. This is also used in the PV models.
2. The renewable energy electrical controls model (reec_a), which has inputs of real power
reference (Pref) that can be externally controlled, reactive power reference (Qref) that
can be externally controlled and feedback of the reactive power generated (Qgen). The
outputs of this model are the real (Ipcmd) and reactive (Iqcmd) current command. A
simplified version of this model (reec_b) is used in the PV models..
3. The emulation of the wind turbine generator driven-train (wtgt_a) for, simulating drive-
train oscillations. The output of this model is speed (spd). In this case speed is assumed
to be a vector spd = [t g], where t is the turbine speed and g the generator speed.
4. A simple linear model of the wind turbine generator aero-dynamics (wtgar_a). This is
based on reference [8], and the same as the 1st generation generic models.
5. A simplified representation of the wind turbine generator pitch-controller (wtgpt_a). This
is similar to the 1st generation type 3 pitch-control model, with the addition of one
parameter Kcc. This parameter was added through consultation and discussions within
the IEC group.
6. A simple emulation of the wind turbine generator torque control (wtgtrq_a)1.
7. A simple renewable energy plant controller (repc_a), which has inputs of either voltage
reference (Vref) and measured/regulated voltage (Vreg) at the plant level, or reactive
power reference (Qref) and measured (Qgen) at the plant level. The output of the repc_a
model is a reactive power command that connects to Qref on the reec_a model. Note:
presently this plant controller can control ONLY one aggregated WTG model
representing a single plant with the same type of WTG. Future versions may need to be
considered for having a controller that controls multiple adjacent plants or multiple types
of WTGs in a single plant. This model can also be used for PV plants.
The repc_a model includes a simple droop control for emulating primary frequency control.
This is intended mainly for emulating down-regulation for over-frequency events, but an up-

1
The version shown in this final specification is based on an earlier version of the model discussed in March, 2012.
At the last WECC REMTF meeting the members agreed to go to this earlier and simpler version.

3-1
regulation feature has also been provided. This is a simple model and is not based on any
validation work and is based on recommendations among the various stakeholders and vendors
participating in the WECC REMTF. This may be refined in the future. Warning: Care must be
taken not to simulate up-regulation (i.e. increasing plant output with decreasing frequency)
where it is not physically meaningful – e.g. when the plant is converting the available incident
wind energy to electrical power, which is certainly the typical operating condition of a wind
power plant.
Warning: For completeness, and based on various comments from the WECC REMTF and
IEC group members, various options (voltage, Q or pf control, with and without deadband
etc.) have been provided for the control options at the plant level. Very preliminary tests have
been done with data just recently made available in the last month. This work is very
preliminary and so the plant level model is not yet necessarily fully validated. Plant level data
has been scarce up to this point. Thus, care must be taken with the selection of these options
and appropriately setting the controller parameters so as to not produce an undesired
response. Further work and research with plant level model validation may in the future
suggest changes to these model features.
In the next few sections each of these building block models is described. In section 3.9 a
description is given on how to build a type 4 WTG or type 3 WTG from these building block
models.

3.2 REGC_A
The regc_a model is shown in Figure 3-1. This model is similar to the existing 1st generation
wt4g model in GE PSLFTM and Siemens PTI PSS®E, with the following exceptions:
1. The time constants for the real and reactive current injection are a model parameter Tg,
instead of being hardcoded.
2. The time constant for the voltage filter is also a parameter Tfltr, instead of being
hardcoded.
3. A rate limit has been added to the reactive current block. It is important to understand
how this rate limit is effected:
a. If the model initializes with an initial reactive power output that is greater than
zero (i.e. reactive power being injected into the grid), then upon fault clearing the
recovery of reactive current is limited at the rate of Iqrmax. In this case the rate
limit (Iqrmin) on reducing reactive current is not effective, reactive current can be
reduced as quickly as desired.
b. If the model initializes with an initial reactive power output that is less than zero
(i.e. reactive power being absorbed from the grid), then upon fault clearing the
recovery of reactive current back down to its original value is limited at the rate of
Iqrmin. In this case the rate limit (Iqrmax) on increasing reactive current is not
effective, reactive current can be increased as quickly as desired.
The action of this reactive current limit is best illustrated by the simulations shown in [3] (see
Figure 4-4 and 4-5 in [3]).

3-2
The rest of the parameters and functionality of the regc_a model is as already described and
implemented in GE PSLFTM and Siemens PTI PSS®E. The logic behind the “high Voltage
Reactive Current Management” and the “Low Voltage Active Current Management” are
provided in Appendix A.
Rate limits on reactive current for recovery after fault.
Upward limit is active when Qgeno > 0
Downward limit is active Qgeno < 0
Qgeno

Iqrmax

Iqcmd
-1 Iq
1 + s Tg
High Voltage
s0 Vt
Iqrmin Reactive Current
Management

Interface to Network Model

LVPL & rrpwr

Ipcmd
1 Ip
1 + s Tg
s1

LVPL Low Voltage

Lvplsw Active Current
0 Lvpl1
1
Management
1 + s Tfltr
1
s2
Zerox Brkpt V

Figure 3-1: Renewable energy generator/converter model (regc_a).

3.3 REEC_A
The reec_a model is shown in Figure 3-2. The table below is a list of all the parameters of the
model. The user must take great care to consult with equipment vendors to identify what is
appropriate for an actual installation. The typical range of values are give only as guidance and
should not be interpreted as a strict range of values, numbers outside of these typical ranges may
be plausible. Where “N/A” is listed in the typical range of values column this indicates that there
is no typical range to be provided. This model is per unitized on its own MVA BASE. Note: the
details of the current limit logic are provided in Appendix B.

Parameter Description Typical Range of Units

Values
MBASE Model MVA base N/A MVA
Vdip The voltage below which the reactive current injection 0.85 – 0.9 pu
(Iqinj) logic is activated (i.e. voltage_dip = 1)
Vup The voltage above which the reactive current injection >1.1 pu
(Iqinj) logic is activated (i.e. voltage_dip = 1)
Trv Filter time constant for voltage measurement 0.01 – 0.02 s

3-3
Parameter Description Typical Range of Units
Values
dbd1 Deadband in voltage error when voltage dip logic is -0.1 – 0 pu
activated (for overvoltage – thus overvoltage response
can be disabled by setting this to a large number e.g.
999)
dbd2 Deadband in voltage error when voltage dip logic is 0 – 0.1 pu
activated (for undervoltage)
Kqv Gain for reactive current injection during voltage dip 0 – 10 pu/pu
(and overvoltage) conditions
Iqh1 Maximum limit of reactive current injection (Iqinj) 1 – 1.1 pu
Iql1 Minimum limit of reactive current injection (Iqinj) -1.1 – 1 pu
Vrefo The reference voltage from which the voltage error is 0.95 – 1.05 pu
calculated. This is set by the user. If the user does
not specify a value it is initialized by the model to
equal to the initial terminal voltage.
Iqfrz Value at which Iqinj is held for Thld seconds following a -0.1 – 0.1 pu
voltage dip if Thld > 0
Thld Time delay for which the state of the reactive current -1 – 1 s
injection is held after voltage_dip returns to zero:
1. If Thld > 0, then once voltage_dip goes back to
0 Iqinj is held at Iqfrz for Thld seconds.
2. If Thld < 0, then once voltage_dip goes back to
0 Iqinj remains in its current injection state (i.e.
Iqinj = (Vrefo – Vt) x Kqv) for Thld seconds.
3. If Thld = 0 then Iqinj goes back to zero
immediately after the voltage_dip is turned off.
Thld2 Time delay for which the active current limit (Ipmax) is 0 s
held after voltage_dip returns to zero for Thld2 seconds
at its value during the voltage dip.
pfaref Power factor angle. This parameter is initialized by N/A rad
the model based on the initial powerflow solution
(i.e. initial P and Q of the model).
Tp Filter time constant for electrical power measurement 0.01 – 0.1 s
Qmax Reactive power limit maximum 0.4 – 1.0 pu
Qmin Reactive power limit minimum -1.0 – -0.4 pu
Vmax Voltage control maximum 1.05 – 1.1 pu
Vmin Voltage control minimum 0.9 – 0.95 pu
Kqp Proportional gain N/A pu
Kqi Integral gain N/A pu
Kvp Proportional gain N/A pu
Kvi Integral gain N/A pu
Vref1 User-define reference/bias on the inner-loop voltage N/A pu
control (default value is zero)
Tiq Time constant on lag delay 0.01 – 0.02 s
dPmax Ramp rate on power reference N/A pu/s
dPmin Ramp rate on power reference N/A pu/s
Pmax Maximum power reference 1 pu

3-4
Parameter Description Typical Range of Units
Values
Pmin Minimum power reference 0 pu
Imax Maximum allowable total converter current limit 1.1 – 1.3 pu
PfFlag Power factor flag (1 – power factor control, 0 – Q N/A N/A
control, which can be commanded by an external
signal)
VFlag Voltage control flag (1 – Q control, 0 – voltage control) N/A N/A
QFlag Reactive power control flag ( 1 – voltage/Q control, 0 – N/A N/A
constant pf or Q control)
Pqflag P/Q priority selection on current limit flag N/A N/A

VDL1
vq1 N/A pu
Iq1 N/A pu
vq2 N/A pu
Iq2 N/A pu
vq3 N/A pu
Iq3 N/A pu
vq4 N/A pu
User-define pairs of points
Iq4 N/A pu
VDL2
vp1 N/A pu
Ip1 N/A pu
vp2 N/A pu
Ip2 N/A pu
vp3 N/A pu
Ip3 N/A pu
vp4 N/A pu
User-define pairs of points
Ip4 N/A pu

3-5
Figure 3-2: Renewable energy electrical control model (reec_a)2.

2
The non-windup integrators for s3 and s2 are linked as follows: if s3 hits its maximum limit and ds3 is positive,
then ds3 is set to 0; if ds2 is also positive, then it is also set to 0 to prevent windup, but, if ds2 is negative, then ds2 is

3-6
3.4 WTGT_A
The wtgt_a model is shown in Figure 3-3. The table below is a list of all the parameters of the
model. The user should realize that this model is a simplified model for the purpose of
emulating the behavior of torsional mode oscillations. The shaft damping coefficient (Dshaft) in
the drive-train model is fitted to capture the net damping of the torsional mode seen in the post
fault electrical power response. In the actual equipment, the drive train oscillations are damped
through filtered signals and active damping controllers, which obviously are significantly
different from the simple generic two mass drive train model used here. Therefore, the
parameters (and variables) of this simple drive-train model cannot necessarily be compared with
actual physical quantities directly. See reference [2] for a discussion of the active damping
controllers, as they pertain for example to the type 3 WTG.
Parameter Description Typical Range Units
of Values
MBASE Model MVA base N/A MVA
Ht Turbine inertia N/A MWs/MVA
Hg Generator inertia N/A MWs/MVA
Dshaft Damping coefficient N/A pu
Kshaft Spring constant N/A pu

Figure 3-3: Wind turbine generator drive-train model (wtgt_a).

not set to 0. A similar rule is applied for s3 hitting the lower limit, but the check is whether ds3 and ds2 are
negative.
Also, note that for the freezing of the states s2, s3, s4 and s5, only the states are frozen, thus in the case of s1 and s2
the proportional gain, if non-zero, still acts during the voltage dip.
Finally, for s5, if Tpord is zero then the time constant and freezing of the state are by-passed, however, the
Pmax/Pmin limits are still in effect.

3-7
3.5 WTGAR_A
The table below is a list of all the parameters of the wtgar_a model shown in Figure 3-4. The
user must define the initial pitch angle based on the current conditions being simulated.
Parameter Description Typical Units
Range of
Values
Ka Aero-dynamic gain factor 0.007 pu/degrees
o Initial pitch angle 0 degrees

Figure 3-4: Wind turbine generator aero-dynamic model (wtgar_a).

3.6 WTGPT_A
The table below is a list of all the parameters of the wtgpt_a model shown in Figure 3-5.
Parameter Description Typical Units
Range of
Values
Kiw Pitch-control integral gain N/A pu/pu
Kpw Pitch-control proportional gain N/A pu/pu
Kic Pitch-compensation integral gain N/A pu/pu
Kpc Pitch-compensation proportional gain N/A pu/pu
Kcc Proportional gain N/A pu/pu
T Pitch time constant 0.3 s
max Maximum pitch angle 27 – 30 degrees
min Minimum pitch angle 0 degrees
dmax Maximum pitch angle rate 5 to 10 degrees/s
dmin Minimum pitch angle rate -10 to -5 degrees/s

3-8
Figure 3-5: Wind turbine generator pitch-controller model (wtgpt_a).

3.7 WTGTRQ_A
The table below is a list of all the parameters of the wtgtrq_a model (Figure 3-6).
Parameter Description Typical Units
Range of
Values
Kip Integral gain N/A pu/pu
Kpp Proportional gain N/A pu/pu
Tp Power measurement lag time constant 0.05 to 0.1 s
Tref Speed reference time constant 30 to 60 s
Temax Maximum torque 1.1 to 1.2 pu
Temin Minimum torque 0 pu
Tflag 1 - for power error, and 0 – for speed error 0 N/A
p1 User-define pairs of points, function f(Pe) 0.2 pu
spd1 0.58 pu
p2 0.4 pu
spd2 0.72 pu
p3 0.6 pu
spd3 0.86 pu

3-9
Parameter Description Typical Units
Range of
Values
p4 0.8 pu
spd4 1.0 pu

Figure 3-6: Wind turbine generation torque model (wtgtrq_a).

3-10
3.8 REPC_A
The table below is a list of all the parameters of the repc_a model shown in Figure 3-7.
Parameter Description Typical Range Units
of Values
MBASE Model MVA base N/A MVA
Tfltr Voltage or reactive power measurement filter time 0.01 – 0.05 s
constant
Kp Proportional gain N/A pu/pu
Ki Integral gain N/A pu/pu
Tft Lead time constant N/A s
Tfv Lag time constant N/A s
RefFlag 1 – for voltage control or 0 – for reactive power N/A N/A
control
Vfrz Voltage below which plant control integrator state 0 – 0.7 pu
(s2) is frozen
Rc Line drop compensation resistance 0 Pu
Xc Current compensation constant (to emulate droop or -0.05 – 0.05 pu
line drop compensation)
Kc Gain on reactive current compensation N/A pu
VcompFlag Selection of droop (0) or line drop compensation (1) N/A N/A
emax Maximum error limit pu
emax Minimum error limit pu
dbd Deadband in control 0 pu
Qmax Maximum Q control output pu
Qmin Minimum Q control output pu
Kpg Proportional gain for power control pu/pu
Kig Integral gain for power control pu/pu
Tp Lag time constant on Pgen measurement s
fdbd1 Deadband downside pu
fdbd2 Deadband upside pu
femax Maximum error limit pu
femin Minimum error limit pu
Pmax Maximum Power pu
Pmin Minimum Power pu
Tlag Lag time constant on Pref feedback s
Ddn Downside droop 20 pu/pu
Dup Upside droop 0 pu/pu
Pgen_ref Initial power reference From powerflow pu
Freq_ref Frequency reference 1.0 pu

3-11
Parameter Description Typical Range Units
of Values
vbus The bus number in powerflow from which Vreg, Freq N/A N/A
is picked up (i.e. the voltage being regulated and
frequency being controlled; it can be the terminal of
the aggregated WTG model or the point of
interconnection)
branch The branch (actual definition depends on software N/A N/A
program) from which Ibranch, Qbranch and Pbranch is
being measured.
Freq_flag Flag to turn on (1) or off (0) the active power control 0 N/A
loop within the plant controller
Note: Vref and Qref are initialized by the model based on Vreg and Qgen in the initial
powerflow solution, and Qext is initialized based on the initialization of the initial Q reference
from the down-stream aggregated WTG model.

Figure 3-7: Renewable energy plant control model (repc_a).

3.9 Creating Type 3 and 4 WTG Models

The above set of building block models can be put together to create a type 3 or 4 WTG. As
shown in [3], there are actually two classes of type 4 WTG. The first class, which is called type
4 WTG A, is designed such that for nearby grid faults there are some visible torsional
oscillations that appear in electrical power output of the unit. The second class, is called type 4

3-12
WTG B, is designed such that for nearby grid faults there are no noticeable torsional oscillations
that appear in electrical power output of the unit. Thus, for the type 4 WTG A the drive-train
model is needed, and for type 4 WTG B it is not needed. Table 3-1 shows the make-up for a
type 4 WTG A, type 4 WTG B and a type 3 WTG model. Figure 3-8, Figure 3-9 and Figure 3-10
show how these blocks come together to form the type 3, type 4 A and type 4 B WTG models.
Table 3-1: Building the type 3 and 4 WTG form the building block models.

Device regc_a reec_a wtgt_a wtgar_a wtgpt_a wtgtrq_a repc_a

Type 4 WTG A X X X X
Type 4 WTG B X X X
Type 3 WTG X X X X X X X

Figure 3-8: A type 3 WTG model.

3-13
Figure 3-9: A type 4 A WTG model.

Figure 3-10: A type 4 B WTG model.

3-14
Table 3-2 and Table 3-3 provide a simply summary of the various control strategies that can be
emulated by these models – namely, the combination of the reec_a and repc_a model features.

Table 3-2: Reactive power control options

Functionality Models Needed PfFlag Vflag Qflag RefFlag

Constant pf control reec_a 1 N/A 0 N/A
Constant Q control reec_a 0 N/A 0 N/A
Local V control only reec_a 0 0 1 N/A
Local coordinated Q/V control only reec_a 0 1 1 N/A
Plant level Q control reec_a + repc_a 0 N/A 0 0
Plant level Vcontrol reec_a + repc_a 0 N/A 0 1
Plant level V Control + coordinated local Q/V control reec_a + repc_a 0 1 1 1
Plant level Q Control + coordinated local Q/V control reec_a + repc_a 0 1 1 0

Table 3-3: Real power control options

Functionality Models Needed PFlag

Do not Emulate torsioanl oscillation reec_a 0
Emulated torsional oscillations in power output reec_a + wtgt_a 1

The protection models associated with the wind turbine generator (i.e. low/high voltage and
low/high frequency tripping) has not been addressed in this document since the existing generic
protection models (lhvrt and lhfrt) that exist in GE PSLFTM (and similar models in Siemens PTI
PSS®E) are adequate for application with this generic model.

Important Note: The actual implementation of these models in software may require subtle
adjustment to accommodate the way the models need to be initialized in commercial tools.

3-15
4
EXAMPLE SIMULATION CASES
The previous documents [2] and [3] provide many examples of simulations performed by EPRI
using the EPRI WTGMV tool [9] using data for single wind turbine generators. The data used in
those cases were provide to EPRI under non-disclosure agreements (NDA) with the various
turbine manufacturers for the purpose of research and investigation of the suitability of the
various model structures being developed and proposed. These vendors graciously agreed to
allow the public dissemination of the research results, as presented here and in the other
references. The actual data, however, is covered under the NDA and cannot be disclosed. Those
examples show that the models presented here, for modeling a single WTG, appear to be
reasonable and adequate. Another vendor has performed their own internal work and reported at
WECC REMTF meetings that in general they found reasonable response from these new generic
models for emulating their equipment. These results can be found in the previous references and
are not presented again here.
More importantly, recently as the commercial vendors have started to implement these models
into the commercial software tools (i.e. PowerWorld, GE PSLFTM, Siemens PIT PSS®E,
PowerTech Labs TSAT, etc.) some extensive testing has been done to both:
1) Test the models from an implementation perspective in the commercial tools (i.e.
identifying bugs and fixing them), and
2) Tests the commercial models against the benchmark simulation cases in [2] and [3].
This work was reported at the June 2013 WECC REMTF meeting [10]. Below are a few
example simulations from that work that illustrates that the commercial models are also able to
reasonably capture the actual turbine behavior – for a single WTG.

4-1
Figure 4-1: Validation result of simulation versus measured real and reactive power for a type 3
WTG. The comparison shown is between EPRI implementation, the commercial tool and the
actual field measured response.

Figure 4-2: Validation result of simulation versus measured real and reactive power for a type 4
WTG A. The comparison shown is between EPRI implementation, the commercial tool and the
actual field measured response.

4-2
Figure 4-3: Validation result of simulation versus measured real and reactive power for a type 4
WTG B. The comparison shown is between EPRI implementation, the commercial tool and the
actual field measured response.

4-3
5
CONCLUSION AND SUMMARY
At this point, with the gracious input of the various equipment vendors, this specification for the
2nd generation generic models has been established. Much dialogue has occurred in the process
of coming to a collective agreement on the final specification. It is certain that further
refinements are likely to be identified through further discussions, particularly for the plant level
controller. However, it is believed that what is presented here makes enough of a significant
improvement to warrant implementing it as soon as possible in order to reap the benefits of being
able to model a variety of WTGs.
Finally, it should be kept in mind that the model under discussion here is a “generic” model for
interconnected power system stability simulations and so one must keep the models simple,
while catering to as wide a possible range of equipment. It would be an insurmountable task to
try to achieve a model that would cater to every possible equipment configuration. Therefore,
when doing detailed plant specific studies, vendor specific models (obtained directly from the
equipment vendors) will still always be the best option. The “generic” models are for bulk
system studies performed by TSOs, TOs, reliability entities, etc.

5-1
6
REFERENCES
[1] NERC IVGTF 1-1, Standard Models for Variable Generation, May 18, 2010,
http://www.nerc.com/docs/pc/ivgtf/IVGTF_Report_PhaseII_Task1-1_Final(5.24).pdf
[2] P. Pourbeik, “Proposed Changes to the WECC WT3 Generic Model for Type 3 Wind
Turbine Generators”, Prepared under Subcontract No. NFT-1-11342-01 with NREL, Issued:
03/26/12 (revised 6/11/12, 7/3/12, 8/16/12, 8/17/12, 8/29/12, 1/15/13, 1/23/13, 9/27/13)
[3] P. Pourbeik, “Proposed Changes to the WECC WT4 Generic Model for Type 4 Wind
Turbine Generators”, Prepared under Subcontract No. NFT-1-11342-01 with NREL, Issued:
12/16/11 (revised 3/21/12, 4/13/12, 6/19/12, 7/3/12, 8/16/12, 8/17/12, 8/29/12, 1/15/13,
1/23/13)
[4] P. Pourbeik, “Proposed Changes to the WECC WT1 Generic Model for Type 1 Wind
Turbine Generators”, Prepared under Subcontract No. NFT-1-11342-01 with NREL, Issued:
1/21/13
[5] Generic Models and Model Validation for Wind and Solar PV Generation: Technical Update.
EPRI, Palo Alto, CA: 2011, 1021763. (Available for free download at www.epri.com)
[6] CIGRE Technical Brochure 328, Modeling and Dynamic Behavior of Wind Generation as it
Relates to Power system Control and Dynamic Performance, August 2007, CIGRE WG
C4.601. (www.e-cigre.org)
[7] Pseudo Governor Model for Type 1 and 2 Generic Turbines, October 2012; prepared by
Enernex (B. Zavadil); Work supported by Sandia National Laboratories and the US
Department of Energy under Sandia Contract #1257843
[8] W. W. Price and J. J. Sanchez-Gasca, “Simplified Wind Turbine Generator Aerodynamic
Models for Transient Stability Studies”, Proceedings of the IEEE PSCE 2006.
[9] Wind Turbine Generator Model Validation (WTGMV) Software; EPRI Product ID 1024346;
http://my.epri.com/portal/server.pt?Abstract_id=000000000001024346
[10] P. Pourbeik, “Update of Model Testing in GE PSLF for Latest Generic Wind and PV
Models”, presentation at June 18, 2013 WECC REMTF meeting.

6-1
A
CONVERTER MODEL GRID INTERFACE
In the generator/converter model block diagram there are two blocks labeled, “high-voltage
reactive current management” and “low-voltage active current management”. These blocks
represent logic associated with the dynamic model and the ac network solution. The actual
implementation of this logic may be software dependant. In the past a simple block diagram was
provide in an effort to attempt to explain the logic, this however seemed to have caused more
confusion. Here we provide a flow chart, provided by GE3, for greater clarification.
High-Voltage Reactive Current Management:

3
N. Miller, “High and Low Voltage Algebraic Network solution flowcharts”, Version 2, November 16, 2012
(revised and provided in an email on 1/11/13).

A-1
Low-Voltage Active Current Management:

A-2
B
CURRENT LIMIT LOGIC
VDL1 is a piecewise linear curve define by four pairs of numbers:
{(vq1,Iq1), (vq2,Iq2), (vq2,Iq3), (vq4,Iq4),}

VDL2 is a piecewise linear curve define by four pairs of numbers:

{(vp1,Ip1), (vp2,Ip2), (vp2,Ip3), (vp4,Ip4),}

If (Pqflag = 0) % Q – priority
Iqmax = min {VDL1, Imax}
Iqmin = -1×Iqmax
Ipmax = min{ VDL2, Imax 2  Iqcmd 2 )
Ipmin = 0
Else % P – priority
Iqmax = min {VDL1, Imax 2  Ipcmd 2 }
Iqmin = -1×Iqmax
Ipmax = min{VDL2, Imax)
Ipmin = 0
End

B-1