Investigating Power System

Primary and Secondary Reserve

Interaction under High Wind
Power Penetration
Yingchen Zhang, Jin Tan, Ibrahim Krad,
Rui Yang, and Vahan Gevorgian
National Renewable Energy Laboratory

Erik Ela
Electric Power Research Institute

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at [Link]/publications.

Technical Report
NREL/TP-5D00-64637
December 2016

Contract No. DE-AC36-08GO28308

 Investigating Power System
Primary and Secondary Reserve
Interaction under High Wind
Power Penetration
Yingchen Zhang, Jin Tan, Ibrahim Krad,
Rui Yang, and Vahan Gevorgian
National Renewable Energy Laboratory

Erik Ela
Electric Power Research Institute
Prepared under Task No. WE14.9C01

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at [Link]/publications.

National Renewable Energy Laboratory Technical Report

15013 Denver West Parkway NREL/TP-5D00-64637
Golden, CO 80401 December 2016
303-275-3000 • [Link]
Contract No. DE-AC36-08GO28308
 NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government.
Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty,
express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the United States government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.

This report is available at no cost from the National Renewable Energy

Laboratory (NREL) at [Link]/publications.

Available electronically at SciTech Connect http:/[Link]/scitech

Available for a processing fee to U.S. Department of Energy

and its contractors, in paper, from:

U.S. Department of Energy

Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831-0062
OSTI [Link]
Phone: 865.576.8401
Fax: 865.576.5728
Email: reports@[Link]

Available for sale to the public, in paper, from:

U.S. Department of Commerce

National Technical Information Service
5301 Shawnee Road
Alexandra, VA 22312
NTIS [Link]
Phone: 800.553.6847 or 703.605.6000
Fax: 703.605.6900
Email: orders@[Link]

Cover Photos by Dennis Schroeder: (left to right) NREL 26173, NREL 18302, NREL 19758, NREL 29642, NREL 19795.

NREL prints on paper that contains recycled content.

List of Acronyms
ACE area control error
AGC automatic generation control
DFIG doubly-fed induction generator
FESTIV Flexible Energy Scheduling Tool for Integrating
Variable Generation
KERMIT Renewable Energy Modeling and Integration
Tool
MAFRIT Multi-Area Frequency Response Integration
Tool
PFR primary frequency response
PSLF Positive Sequence Load Flow
PSS/E Power System Simulator for Engineering

iii
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
Executive Summary
Power system frequency needs to be maintained close to its nominal value at all times to
successfully balance load and generation and maintain system reliability. Adequate primary
frequency response and secondary frequency response are the primary forces to correct an
energy imbalance at the second-to-minute level. As wind energy becomes a larger portion of
the world’s energy portfolio, there is an increased need for wind to provide frequency
response. This paper addresses one of the major concerns about using wind for frequency
regulation: the unknown factor of the interaction between primary and secondary reserves.
The lack of a commercially available tool to model this has limited the energy industry’s
understanding of when the depletion of primary reserves will impact the performance of
secondary response or vice versa. This paper investigates the issue by developing a multi-area
frequency response integration tool with combined primary and secondary capabilities. The
simulation is conducted in close coordination with economical energy scheduling scenarios to
ensure credible simulation results.

iv
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
Table of Contents
1 Introduction ........................................................................................................................................... 1
2 Model Development and Validation .................................................................................................... 4
3 Case Studies ......................................................................................................................................... 8
4 Results ................................................................................................................................................. 12
5 Conclusion .......................................................................................................................................... 14
References ................................................................................................................................................. 15
Appendix: PSLF System Data .................................................................................................................. 16

v
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
List of Figures
Figure 1. Overall architecture of the MAFRIT simulation platform............................................................. 3
Figure 2. Four-area system in PSLF ............................................................................................................. 4
Figure 3. Four-area system in MAFRIT ....................................................................................................... 5
Figure 4. Comparison of rotor speed and active power of all generators in PSLF and Simulink ................. 6
Figure 5. Comparison of terminal voltages (magnified on the right) of all generators in PSLF and
Simulink ................................................................................................................................... 6
Figure 6. Droop curve of the wind turbine generation .................................................................................. 7
Figure 7. Wind output (MAFRIT) with and without AGC for the 1-hour simulation .................................. 8
Figure 8. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 24
hours and the display of system load, frequency, and ACE ..................................................... 9
Figure 9. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 24
hours and the display of system load, frequency, and ACE under 18% wind energy
penetration.............................................................................................................................. 10
Figure 10. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 1 hour
showing the system load, frequency, and ACE under 18% wind energy penetration............ 10
Figure 11. Wind power and system frequency when load increases by 50 MW at t = 2,504 s .................. 12
Figure 12. Wind power and system frequency when load increases by 50 MW at t = 1,900 s .................. 13
Figure A-1. Four-area system ..................................................................................................................... 17
Figure A-2. Illustration of the legends ........................................................................................................ 18

List of Tables
Table 1. Simulation System Summary .......................................................................................................... 5
Table A-1. Summary of the System ............................................................................................................ 18
Table A-2. Bus Data ................................................................................................................................... 19
Table A-3. Line Data .................................................................................................................................. 21
Table A-4. Transformer Data ...................................................................................................................... 22
Table A-5. Generator Data .......................................................................................................................... 23
Table A-6. Load Data ................................................................................................................................. 24
Table A-7. Synchronous Generator Model and Parameter ......................................................................... 25
Table A-8. Exciter Model and Parameter ................................................................................................... 26
Table A-9. Governor Model and Parameter................................................................................................ 27
Table A-10. Generator Data of DFIG ......................................................................................................... 28
Table A-11. Turbine Data of DFIG ............................................................................................................ 28
Table A-12. Converter Data of DFIG ......................................................................................................... 29
Table A-13. Control Data of DFIG ............................................................................................................. 29
Table A-14. Parameter of Simplified Wind Power Plant Model ................................................................ 30

vi
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
1 Introduction
To successfully manage the nation’s bulk electric power system, the balance of generation
must be maintained at all timescales. The timescale at which this occurs will dictate the
operation needed to ensure that the system is in balance. An imbalance between generation
and load can overload transmission lines and cause unscheduled power flows, voltage
magnitude fluctuations, and electrical frequency deviations. A severe frequency deviation can
lead to a partial system failure or worse, a cascading failure (e.g., blackout). Electric power
system operators use a variety of scheduling and control techniques to maintain the electricity
frequency close to its nominal value at all times. An interconnected power system must have
adequate resources to respond to a variety of contingency events to ensure rapid restoration of
the frequency. Primary frequency response (PFR)—also called primary control reserve [1]
and frequency responsive reserve [2]—is the capacity available for automatic local response
to frequency excursions through turbine speed that adjusts to counter-frequency deviations to
stabilize the frequency [3]. PFR occurs shortly after an event and acts to stabilize the
frequency deviation to a steady-state level. Although primary control is a function of local
controllers responding to a frequency deviation, secondary freqeuency response control is a
centralized control directed by the system operator. It utilizes automatic generation control
(AGC) to restore the system frequency to its nominal value and keep the interchanges
between balancing authority areas to their scheduled levels.

Wind power has the capability to provide both forms of responses through active power
control. To provide upward reponse to correct under-frequency conditions, wind power plants
track their maximum available power and schedule their output below the maximum power
point [4]. This is different from conventional generators, which can increase or decrease their
fuel flow to control their power output willingly. With increased wind power penetration in
all major North American interconnections, there is an increased need to expand the use of
frequency control capabilities that can be provided by wind energy technology [5]. If
designed correctly, active power control from wind power can perform better than
conventional generators in terms of speed and accuracy because most existing wind power
plants interface with power grids through power electronics devices [6].

One of the challenges of using wind power for frequency regulation is the unknown outcome
when primary and secondary reserves interact [5]. In current system operations, PFR reserve
is not scheduled in unit commitment and real-time dispatch because it is assumed that it is an
inherent function of conventional generators (via direct control of fuel flows through steam
valve, gate valve, or combustor) and that there is always ample supply. However, with
increased penetrations of wind, this may not always be the case. Under current operating
mechanisms, if wind power is enabled to provide both primary and secondary responses but
only secondary reserve is scheduled through the ancillary services market, if an event
happens that deployes PFR reserves there will be a deficiency on the energy that restores the
system frequency response to its nominal value. Also, under circumstances in which the
secondary reserve is under-procured, primary response will automatically act to further
restore frequency. As a result, the remaining primary response may not be available to
stabilize the frequency after a sudden loss of generation.

Mitigating these challenges requires a tool that can realistically model the interactions
between primary and secondary control. There are tools that are designed to implement AGC
in an extended system simulation. For example, the Power System Simulator for Engineering
(PSS/E) is a software tool that simulates power system steady-state and dynamic
performance. In 2011, PSS/E rolled out a new function called “extended term dynamic

1
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
simulation,” which extended the time frame for dynamic simulations to virtually unlimited
and implemented slow-moving controls, such as AGC, switched shunts, and transformer tap
changing [7][8]. PSS/E uses implicit integration to solve differential equations with changing
time steps. This function offers the means to change simulation time steps by choosing
different modes; however, when the user choses large time steps, the fast responses will be
filtered. The unit allocation is calculated using unit base point, unit maximum regulation, and
the ramp rate.

KEMA, Inc., published the Renewable Energy Modeling and Integration Tool (KERMIT) [9]
to simulate power system frequency behavior during a 24-hour time interval. The tool
incorporates an AGC model that responds to non-fault events, such as a generator trip, load
shedding, and variation of renewable resources. KERMIT is designed to simulate a power
system’s dynamic resources and provides simulations from 1 second to 25 hours. This tool
incorporates production models, such as balancing market clearing, scheduling ramp look-
ahead, generation scheduling, and load/renewable resource forecasting; and system dynamic
models, such as inertias, turbine governors, excitations, and AGC. However, the major
drawback is that KERMIT does not include a detailed network model; it represents the
control areas as nodes in a reduced network. The inter-area flows are calculated by relative
phase angles between areas. (The phase angles are decided by the area frequency and
instantaneous mismatch of the area’s mechanical-electrical power.) As a result, KERMIT
does not have individual dynamic responses for each generator, and the entire network is
lossless. KERMIT was utilized to study the impact of renewable generation on the California
grid [10] [11]. This study primarily looked into the variability associated with renewable
resources and how energy storage provides AGC and can help mitigate variability; thus, the
study focused on secondary reserves and the system response. A similar study was done on
the PJM grid to assess the effectiveness of AGC in a frequency regulation market [12]. All of
the KERMIT-related studies focused on secondary reserves and AGC control without a
balanced focus on PFR.

The Flexible Energy Scheduling Tool for Integrating Variable Generation (FESTIV) tool
developed by researchers at the National Renewable Energy Laboratory has focused on
impacts of area control error (ACE) and frequency with a detailed inclusion of electricity
market operation (including security-constrained unit commitment and economic dispatch)
[14]. It is a forward-looking, electricity scheduling simulation software, and it includes
security-constrained unit commitment, security-constrained economic dispatch, and AGC
sub-models. Each sub-model’s output serves as the input to subsequent sub-models.
However, it does not model the dynamic behavior of PFR or the impacts of multi-area
interconnections.

In this study, we address this challenge by designing a Multi-Area Frequency Response

Integration Tool (MAFRIT) in Simulink to simulate the frequency dynmaic of the grid and
using FESTIV to provide economical scheduling. In MAFRIT, we model a power system that
has multiple balancing authority areas, and we calculate the ACE based on the frequency
deviation of each area and the tie-line flow deviation from the schedule. We decide the AGC
participation factors based on an economic reallocation from the FESTIV scheduling. All of
the generator set points in MAFRIT are updated every 5 minutes according to the energy
scheduling. Within each 5 minutes, they adjust their output automatically based on the AGC
signal to balance the load in a 4-second resolution. The overall architecture of the simulation
platform is displayed in Figure 1.

2
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Figure 1. Overall architecture of the MAFRIT simulation platform

Section 2 describes the model development and validation. Section 3 and Section 4 provide
case studies about the interactions between wind primary and secondary reserves. Section 5
provides a discussion and conclusion.

3
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
2 Model Development and Validation
MAFRIT integrates PFR (primary freqeuncy response) with secondary frequency response
(AGC). It simulates the power system dynamic response in a full-time spectrum with variable
time steps, from milliseconds to minutes and hours to days. It is capable of simulating both
normal and event conditions, and it can represent real power system operations and thus
evaluate the adequacy of primary and secondary reserves. This unique interaction between a
turbine governor model and a novel AGC model places special emphasis on electric power
systems that have high penetrations of wind generation. To ensure the credibility of the
model, a demonstration model provided by GE’s Positive Sequence Load Flow (PSLF)
dynamic simulation software is used and translated into the Simulink platform. The PSLF
system layout is displayed in Figure 2, and the system layout modeled in MAFRIT is shown
in Figure 3.

Figure 2. Four-area system in PSLF

4
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Figure 3. Four-area system in MAFRIT

A summary of the PSLF system is displayed in Table 1. Details about the system
parameters—including buses, lines, transformers, generators, and loads—are included in the
appendix.

Table 1. Simulation System Summary

Number
Areas 4
Buses 18
Lines 18
Transformers 6
Generators 4
Loads 7

The dynamic model of MAFRIT has been validated against the PSLF simulation, and the
results are displayed in Figure 4 and Figure 5. Figure 4 shows the rotor speed and real power
output of all the generators after a sudden increase in load. Figure 5 shows the terminal
voltage magnitude of all of the generators after a fault. The performance of the Simulink
model closely matches the results of the PSLF simulation.

5
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 1.02 550

1.01
Speed1/p.u.

Pe1/MW
500

450
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

1.02 1100

1.01 1000
Speed2/p.u.

Pe2/MW
900
1

800
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

1.02 550

1.01 500
Speed3/p.u.

Pe3/MW
450
1

400
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

1.02 1100

PSLF PSLF
Simulink 1000 Simulink
1.01
Speed4/p.u.

Pe4/MW
900
1

800
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Time/sec Time/sec

Figure 4. Comparison of rotor speed and active power of all generators in PSLF and Simulink

1.1 1.1

1 1
G1 Voltage/p.u.
G1 Voltage/p.u.

0.9 0.9

0.8 0.8
0 10 20 30 40 50 60 70 80 90 100 19 19.2 19.4 19.6 19.8 20 20.2 20.4 20.6 20.8 21

1.1 1.1

1 1
G2 Voltage/p.u.
G2 Voltage/p.u.

0.9 0.9

0.8 0.8
0 10 20 30 40 50 60 70 80 90 100 19 19.2 19.4 19.6 19.8 20 20.2 20.4 20.6 20.8 21

1.1 1.1

1.05
G3 Voltage3/p.u.
G3 Voltage3/p.u.

1 1

0.95

0.9 0.9
0 10 20 30 40 50 60 70 80 90 100 19 19.2 19.4 19.6 19.8 20 20.2 20.4 20.6 20.8 21

1.1 1.1

PSLF PSLF
1.05
Simulink Simulink
G4 Voltage/p.u.
G4 Voltage/p.u.

1 1

0.95

0.9 0.9
0 10 20 30 40 50 60 70 80 90 100 19 19.2 19.4 19.6 19.8 20 20.2 20.4 20.6 20.8 21
Time/sec Time/sec

Figure 5. Comparison of terminal voltages (magnified on the right) of all generators in PSLF
and Simulink

In addition to the validated dynamic model, an AGC controller is modeled in Simulink. The
ACE calculation of the AGC controller is described below.

= , + , , = , =

6
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
where

i is a bias factor of area i

, is the tie-line power change of area i
is the frequency deviation of area i
is the synchronizing torque coefficient between areas i and j.
In addtion, a 360-MW doubly-fed induction generator (DFIG) based wind power plant [13] is
added in Area 4. The PFR of the DFIG is implemented according to the droop curve shown in
Figure 6. Under normal operating conditions with near-nominal system frequency the control
is set to provide a specified margin by generating less power than is available from the unit.
The reserve available (or “headroom”) is the available energy curtailed shown as the reserve
between the operational point and P0.

Figure 6. Droop curve of the wind turbine generation

7
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
3 Case Studies
The purpose of this study is to investigate the adequacy of the primary and secondary
reserves using MAFRIT. The credibility of the accuracy of the MAFRIT simulation lies in
the dynamic model benchmarking using PSLF, as discussed above. It also lies in the natural
connection to the energy scheduling tool FESTIV. It includes security-constrained unit
commitment, security-constrained economic dispatch, and AGC sub-models. Each sub-
model’s output serves as the input to subsequent sub-models [14]. The tool co-optimizes the
energy and reserves to closely reflect the way in which system operators schedule energy in
reality. This tool is completely configurable so that the effects of different operating
strategies can be explored. This study uses FESTIV to generate schedules for all generators in
MAFRIT. All generators are enabled to schedule regulating reserve. The scheduling assumes
perfect forecasting errors for wind power. To ensure that wind power will be scheduled for
reserve, its reserve cost is made lower than the other generators.

A close look at the 1-hour simulation results for the wind turbine generator’s response when
providing or not providing AGC is shown in Figure 7. When the wind turbine generator
tracks only energy scheduling, it can change its output very rapidly. When a wind turbine
generator helps restore system frequency, its output deviates from the energy schedule and
there is a need to withold energy to provde up-ramping. In reality, ramp limiters are added to
wind power plants, because their response can be very fast and increase frequency volatility.
In this study, ramp limiters are not considered.

Figure 7. Wind output (MAFRIT) with and without AGC for the 1-hour simulation

The 24-hour simulation results are displayed in Figure 8. The load profile displayed is the
aggregate of all loads at different parts of the system. The generators respond to load changes
by following the 5-minute energy schedule and using 4-second AGC; therefore, their outputs
deviate from the schedule. The wind turbine generator’s output is also limited by its
maximum power point tracking. This 24-hour simulation does not include any disturbances.
The frequency deviations are caused by load ramping and the quick responses of the wind
turbine to the dispatch signals.

8
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
Figure 8. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 24
hours and the display of system load, frequency, and ACE

In the above secnario, wind penetration as calculated by energy during the 24-hour period is
6%. To obtain a deeper understanding of the impact of system performance when wind is
providing regulation, an 18% penetration scenario is simulated. Figure 9 presents the 24-hour
simulation results for that scenario, and Figure 10 presents the magnified 1-hour version.

9
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Figure 9. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 24
hours and the display of system load, frequency, and ACE under 18% wind energy penetration

G1 G2
280 1200

260 1100
MW

MW

240 1000
schedule

Real output
220 900
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

G3 G4
240 700

220 600
MW

MW

200 500

180 400
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Wind Power System Load

200 2140

2120

100
MW

MW

schedule
2100
Real output

MPPT
0 2080
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

frequency ACE
60.2 0.05

60.1

0
HZ

pu

60

59.9 -0.05
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Time/Hour Time/Hour

Figure 10. Scheduling (FESTIV) and real output (MAFRIT) of the wind turbine generators for 1
hour showing the system load, frequency, and ACE under 18% wind energy penetration

10
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
Comparing the 18% penetration case to the 6% penetration case, the frequency and ACE
devation in higher wind peneration perform worse than in the lower wind penetration because
wind power varibilbity in the prior case is more significant than it is in the low penetration
case. Note that in the high penetration case, when the wind power is capped by the maximum
capacity of the wind power plant, the frequency is reduced. This means that the fast-change
wind power will fulfill the challenge of frequency control for the grid. In the future high wind
penetration case, more frequency responsive reserve is needed.

11
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
4 Results
To study the interaction between primary and secondary reserves, we applied disturbances at
different times during the 24-hour simulation. The primary focus is to assess the frequency
response when wind is providing primary response, AGC, or both. Figure 11 illusrates that
when wind is providing frequency regulation and an event happens, wind does not have
enough headroom to provide a full-scale primary response. Wind that provides only primary
response has a better frequency nadir. Wind that provides only AGC has a faster response to
restore frequency.
7 Wind Power
10
5
Wind noAGC noPri
Pwind order
Wind noAGC withPri

4 MPPT Wind withAGC noPri

Wind withAGC withPri

3
W

2 Schdule

1
2500 2520 2540 2560 2580 2600
Time/s

Frequency

60

59.9
HZ

Wind noAGC noPri

59.8
Wind noAGC withPri
Wind withAGC noPri
Wind withAGC withPri
59.7
2500 2520 2540 2560 2580 2600
Time/s
(a)

Figure 11. Wind power and system frequency when load increases by 50 MW at t = 2,504 s

12
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 7 Wind Power
10
3.5

3
Wind noAGC withPri
2.5 Wind withAGC withPri
Wind withAGC30delay withPri
2
W

1.5

0.5
1900 1920 1940 1960 1980 2000
Time/s

Frequency
60.05

60

59.95

59.9 59.9
HZ

59.85
59.85
Wind noAGC withPri
59.8
59.8 Wind withAGC withPri
Wind withAGC30delay withPri
59.75
59.75 1904 1906
1900 1920 1940 1960 1980 2000
Time/s
(b)

Figure 12. Wind power and system frequency when load increases by 50 MW at t = 1,900 s

In most system operations, the AGC signal is disabled for tens of seconds immediately
following a disturbance so that controllers can focus on restoring the system to equilibrium.
We simulated that effect by disabling AGC for 30 seconds after the disturbance occured, as
shown in Figure 12. Also, in this case, when the disturbance happens the wind power plant
has enough energy to provide both primary and secondary response in full scale. The result
shows that the the frequency nadirs for delayed and non-delayed AGC response are nearly the
same. Comparing this to the no-AGC case, the wind with AGC will reduce the restoration
time. In particular, in the case of the delayed AGC response, when the AGC kicks in 30
seconds after the event happpens, it brings the frequency back to the nominal value faster.

13
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
5 Conclusion
The proposed tool allows for a better understanding of the interaction between PFR and
secondary frequency response, which are responses typically simulated with separate tools.
An improved understanding of the interactions of these controls is important so that any
reliability issues that occur between the seams of these two time frames can be assessed.
Careful consideration of these interactions will improve power system reliability and aid the
design of control systems that will result in responses that are superior to those of
conventional thermal generation and still have very little effect on the loading and life of the
wind turbine and its components. A better understanding of the interaction between primary
and secondary frequency control on multi-area systems with and without wind power plants
providing both of these controls, and how they impact reliability compliance measures in
various grids, will help the industry move forward on PFR market designs.

14
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
References
1. ENTSO-E, UCTE, Operational Handbook Policy 1: Load-Frequency Control and
Performance (Brussels, Belgium, March 2009).

2. Western Electricity Coordinating Council, “White Paper on Frequency Responsive

Reserve Standard” (April 2005).

3. N. Miller et. al., “California ISO Frequency Response Study” (Final Draft),
(Schenectady, NY: GE, November 2011).

4. J. Aho et. al., “Controlling Wind Turbines for Secondary Frequency Regulation: An
Analysis of AGC Capabilities Under New Performance Based Compensation Policy,”
in the 13th International Workshop on Large-Scale Integration of Wind Power Into
Power Systems as Well as on Transmission Networks for Offshore Wind Power Plants
Proceedings (Berlin, Germany, November 2014).

5. J. H. Eto et. al., Use of Frequency Response Metrics to Assess the Planning and
Operating Requirements for Reliable Integration of Variable Renewable Generation
(Technical Report) (Berkeley, CA: Ernest Orlando Lawrence Berkeley National
Laboratory, December 2010).

6. E. Ela et. al., Active Power Controls from Wind Power: Bridging the Gaps (Technical
Report), NREL/TP-5D00-60574 (Golden, CO: January 2014).

7. L. Wang and D. Chen, “Automatic Generation Control (AGC) Dynamic Simulation in

PSSE,” Power Technology (February 2011). [Link].

8. L. Wang and D. Chen, “Experience with Automatic Generation Control (AGC)

Dynamic Simulation in PSSE,” Power Technology (August 2012).
[Link].

9. “KERMIT Dynamic Models,” KERMIT Documentation for CA ISO Training.

10. KEMA, Inc., Research Evaluation of Wind Generation, Solar Generation, and
Storage Impact on the California Grid (Project Report) (Sacramento, CA: California
Energy Commission, June 2010).

11. KEMA, Inc., Emission II Study of Advanced Storage used for Frequency Regulation
(Technical Report) (Albuquerque, NM: Sandia National Laboratories, December
2012).

12. KEMA, Inc., KERMIT Study Report: To Determine the Effectiveness of the AGC in
Controlling Fast and Conventional Resources in the PJM Frequency Regulation
Market (Technical Report) (Norristown, PA: PJM, December 2011).

13. E. Ela et. al, Active Power Control from Wind Power phase 1 Report (Technical
Report) (Golden, CO: )

14. E. Ela et. al., “A Flexible Power System Operations Simulation Model for Assessing
Wind Integration,” in IEEE Power and Energy Society General Meeting Prcoeedings
(Detroit, MI: July 2011).

15
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
Appendix: PSLF System Data
1. System layout
The four-area system given in GE’s PSLF is shown in Figure A-1.

16
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Figure A-1. Four-area system

The voltage levels for the buses shown in blue, red, and purple are 16 kV, 230 kV, and 500 kV, respectively. Bus 101, Bus 1, and Bus 2
are used as examples to illustrate the meaning of the legends around the buses.

17
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Figure A-2. Illustration of the legends

2. Summary of the system

Table A-1. Summary of the System

Number
Areas 4
Buses 18
Lines 18
Transformers 6
Generators 4
Loads 7

18
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 3. System power flow data
The power base for the system is 100 MVA.

A. Bus data
Table A-2. Bus Data

Voltage Base Bus VSched V Angle Vmax Vmin

Bus No. Bus Name Area
(kV) Type (p.u.) (p.u.) (deg) (p.u.) (p.u.)
1 NORTH-01 230 1 0.9897 0.9854 -5.08 1.0397 0.9397 1
2 NORTH-02 230 1 0.9586 0.9464 -12.91 1.0086 0.9086 1
3 NORTH-03 230 1 0.9675 0.957 -12.9 1.0175 0.9175 1
11 WEST--01 230 1 0.9795 0.9787 6.30 1.0295 0.9295 2
12 WEST--02 230 1 0.9097 0.9065 -9.35 0.9597 0.8597 2
13 WEST--03 230 1 0.9129 0.9099 -9.32 0.9629 0.8629 2
14 WEST--04 500 1 0.9241 0.9218 -9.62 0.9742 0.8741 2
21 SOUTH-01 230 1 0.9794 0.9776 -13.75 1.0294 0.9294 3
22 SOUTH-02 230 1 0.9854 0.9843 -11.35 1.0354 0.9354 3
23 SOUTH-03 230 1 0.9956 0.9951 -9.15 1.0456 0.9456 3
24 SOUTH-04 500 1 0.983 0.9818 -11.2 1.0330 0.9330 3
31 EAST--01 230 1 0.9977 0.9974 -11.38 1.0476 0.9477 4
32 EAST--02 230 1 0.9996 0.9992 -11.67 1.0496 0.9496 4
33 EAST--03 230 1 0.9998 0.9994 -10.95 1.0498 0.9498 4
101 NORTH-G1 16 0 1.0000 1.0000 0 1.0500 0.9500 1
111 WEST--G1 16 2 1.0000 1.0000 11.19 1.0500 0.9500 2
231 SOUTH-G1 16 2 1.0000 1.0000 -4.35 1.0500 0.9500 3
311 EAST--G1 16 2 1.0000 1.0000 -6.58 1.0500 0.9500 4

Bus no.: bus number

Bus name: bus name
Voltage base: the voltage base for the bus (kV)
Bus type: 0 - slack bus (voltage magnitude and phase angle fixed
1 - load bus (PQ bus, unconstrained voltage phase angle and magnitude)

19
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 2 - generator bus (PV bus, voltage control within generator limits)
VSched: scheduled voltage magnitude (p.u.)
V: actual voltage magnitude (p.u.)
Angle: actual voltage phase angle (degree)
Vmax: voltage checking upper limit (p.u.)
Vmin: voltage checking lower limit (p.u.)
Area: area to which the bus belongs

20
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 B. Line data
Table A-3. Line Data

First
From Bus To Bus
Resistance Reactance Susceptance Branch
Area
V Base V Base (p.u.) (p.u.) (p.u.) Rating
No. Name No. Name
(kV) (kV) (MVA)
1 NORTH-01 230 2 NORTH-02 230 0.01 0.05 0 600 1
1 NORTH-01 230 3 NORTH-03 230 0.01 0.05 0 600 1
2 NORTH-02 230 3 NORTH-03 230 0.01 0.05 0 600 1
2 NORTH-02 230 12 WEST--02 230 0.02 0.2 0.1 600 1
2 NORTH-02 230 13 WEST--03 230 0.03 0.3 0.15 600 1
3 NORTH-03 230 21 SOUTH-01 230 0.02 0.2 0.1 600 1
3 NORTH-03 230 31 EAST--01 230 0.05 0.5 0.25 600 1
11 WEST--01 230 12 WEST--02 230 0.01 0.05 0 600 2
11 WEST--01 230 13 WEST--03 230 0.01 0.05 0 600 2
12 WEST--02 230 13 WEST--03 230 0.01 0.05 0 600 2
14 WEST--04 500 24 SOUTH-04 500 0.06 0.4 0.2 2000 2
21 SOUTH-01 230 22 SOUTH-02 230 0.01 0.05 0 600 3
21 SOUTH-01 230 23 SOUTH-03 230 0.01 0.05 0 600 3
21 SOUTH-01 230 32 EAST--02 230 0.04 0.4 0.2 600 3
22 SOUTH-02 230 23 SOUTH-03 230 0.01 0.05 0 600 3
23 SOUTH-03 230 33 EAST--03 230 0.02 0.2 0.1 600 3
31 EAST--01 230 32 EAST--02 230 0.01 0.05 0 600 4
31 EAST--01 230 33 EAST--03 230 0.01 0.05 0 600 4

Resistance: branch section positive sequence resistance (p.u.)

Reactance: branch section positive sequence reactance (p.u.)
Susceptance: branch section positive sequence susceptance (p.u.)

21
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 C. Transformer data
Table A-4. Transformer Data

From Bus To Bus Primary Secondary First

PowerB
Winding Winding Tap R X B Branch
V V ase Area
Nominal Nominal Ratio (p.u.) (p.u.) (p.u.) Rating
No. Name Base No. Name Base (MVA)
V (kV) V (kV) (MVA)
(kV) (kV)
NORTH- NORTH-
230 101 16 600 230 16 1 0 0.1 0 600 1
01 G1
WEST- WEST-
11 230 111 16 1200 230 16 1 0 0.1 0 1200 2
01 G1
WEST- WEST-
13 230 14 500 100 230 500 1 0.01 0.05 0 100 2
03 04
SOUTH- SOUTH-
22 230 24 500 100 230 500 1 0.01 0.05 0 100 3
02 04
SOUTH- SOUTH-
23 230 231 16 600 230 16 1 0 0.1 0 600 3
03 G1
EAST- EAST-
31 230 311 16 1200 230 16 1 0 0.1 0 1200 4
01 G1

Power base: transformer power base (MVA)

R: resistance primary to secondary (p.u. value on transformer bases)
X: reactance primary to secondary (p.u. value on transformer bases)
B: Magnetizing susceptance (p.u. value on transformer bases)

22
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 D. Generator data
Table A-5. Generator Data

Bus Regulating Bus

Control PGen QGen Qmax Qmin VSched V
V V Area
Mode (MW) (MVar) (MVar) (MVar) (p.u.) (p.u.)
No. Name Base No. Name Base
(kV) (kV)
NORTH- NORTH-
101 16 0 523.4 110.8 999 -999 101 16 1 1 1
G1 G1
WEST-- WEST--
111 16 0 1000 298.3 500 -500 111 16 1 1 2
G1 G1
SOUTH- SOUTH-
231 16 0 500 50.6 250 -250 231 16 1 1 3
G1 G1
EAST-- EAST--
311 16 0 1000 72.6 500 -500 311 16 1 1 4
G1 G1

Control mode: 0 - voltage at the regulated bus is held constant within Q limits of generators specified by Qmax and Qmin
PGen: actual active power output of the generator (MW)
QGen: actual reactive power output of the generator (MVar)
Qmax: maximum reactive power output of the generator (MVar)
Qmin: minimum reactive power output of the generator (MVar)
Regulating bus: the bus at which the voltage is regulated by the generator
VSched: generator scheduled voltage magnitude (p.u.)
V: generator actual voltage magnitude (p.u.)

23
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 E. Load data
Table A-6. Load Data

Constant Current Constant Admittance

Bus Constant Power Load
Load Load Load
Area
V Base PLoad QLoad IPLod IQLod GLoad BLoad Type
No. Name
(kV) (MW) (MVar) (MW) (MVar) (MW) (MVar)
2 NORTH-02 230 0 0 0 0 334.93 0 0 1
3 NORTH-03 230 0 0 0 0 272.96 0 0 1
12 WEST--02 230 0 0 0 0 547.56 0 0 2
13 WEST--03 230 0 0 0 0 543.53 0 0 2
21 SOUTH-01 230 0 0 0 0 261.58 0 0 3
23 SOUTH-03 230 0 0 0 0 252.49 0 0 3
31 EAST--01 230 0 0 0 0 1005.15 0 0 4

Each load has three components: a constant power component, a constant current component, and a constant admittance component.

PLoad: active power consumed by the constant power component (MW)

QLoad: reactive power consumed by the constant power component (MVar)
IPLod: active power consumed by the constant current component at 1 p.u. voltage (MW)
IQLod: reactive power consumed by the constant current component at 1 p.u. voltage (MVar)
GLoad: active power consumed by the constant admittance component at 1 p.u. voltage (MW)
BLoad: reactive power consumed by the constant admittance component at 1 p.u. voltage (MVar)
Load type: 1 - non-conforming (fixed)
0 - conforming (scalable)

24
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 4. System dynamic data
A. Synchronous generator model and parameter
Table A-7. Synchronous Generator Model and Parameter

Capacity
Type Tpdo Tppdo Tpqo tppqo H D Ld Lq Lpd Lpq Lppd Lppq
(MW)
NORTH-G1 318 GENCC 6 0.05 1 0.05 3.5 0 1.5 1.4 0.32 0.5 0.2 0.2
NORTH-G1 282 GENCC 6 0.05 1 0.05 3.5 0 1.5 1.4 0.32 0.5 0.2 0.2
SOUTH-G1 600 GENSAL 6 0.05 0.05 2.8 0 1.05 0.69 0.32 0.2
EAST-G1 1200 GENSAL 6 0.05 0.05 2.8 0 1.05 0.69 0.32 0.2
WEST-G1 1200 GENSAL 6 0.05 0.05 2.8 0 1.05 0.69 0.32 0.2

Capacity(MW) Type ll s1 s12 ra rcomp xcomp accel pf qf

NORTH-G1 318 GENCC 0.12 0.03 0.29 0 -0.05 -0.01 0.4 0.53 0.5
NORTH-G1 282 GENCC 0.12 0.03 0.29 0 -0.05 -0.01 0.4 0.53 0.5
SOUTH-G1 600 GENSAL 0.12 0.03 0.3 0.005 0.0 0.0
EAST-G1 1200 GENSAL 0.12 0.03 0.3 0.005 0.0 0.0
WEST-G1 1200 GENSAL 0.12 0.03 0.3 0.005 0.0 0.0

GENSAL: Salient pole generator represented by equal mutual inductance rotor modeling
GENCC: Generator represented by uniform inductance ratios rotor modeling to match WSCC type F model; shaft speed effects are
neglected. Intended to model cross-compound machines represented as one generator in the load flow.

25
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 B. Exciter model and parameter
Table A-8. Exciter Model and Parameter

Capacity tr tb tc ka ta vrmax vrmin ke te

NORTH-G1 318 Exdc1 0.0 0.0 0.0 40.00 0.10 0.0 0.0 0.0 0.50
NORTH-G1 282 Exdc1 tr tb tc ka ta vamax vamin te
SOUTH-G1 600 Exac1 0.05 1.00 1.00 40.0 0.10 5.00 -5.00 0.5
EAST-G1 1200 Exac1 0.05 1.00 1.00 40.0 0.10 5.00 -5.00 0.5
WEST-G1 1200 Exac1 0.05 1.00 1.00 40.0 0.10 5.00 -5.00 0.5

kf Tf1 Tf2 E1 Se1 E2 Se2

NORTH-G1 Exdc1 0.05 1.0 0 2.2 0.1 3.5 0.4
NORTH-G1 Exdc1 kf Tf kc kd ke E1 Se1 E2 Se2
SOUTH-G1 Exac1 0.2 1.0 0.1 0.6 1.0 1.0 0.03 2.0 0.3
EAST-G1 Exac1 0.2 1.0 0.1 0.6 1.0 1.0 0.03 2.0 0.3
WEST-G1 Exac1 0.2 1.0 0.1 0.6 1.0 1.0 0.03 2.0 0.3

Exdc1: IEEE (1980) DC1 excitation system model with added speed multiplier
Exac1: IEEE type AC1 excitation system with added speed multiplier and with VR limits modified to agree with IEEE type AC1A.

26
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 C. Governor model and parameter
Table A-9. Governor Model and Parameter

Capacity Type k T1 T2 T3 uo uc pmax pmin T4 K1

20.0000 0.5 0.0 0.2 0.06 -0.5 1 0.0 0.4 0.3
K8 Db1 eps Db2 Gv1 Pgv1 Gv2 Pgv2 Gv3 Pgv3
0.0 0.0 0.0 0.0 0.0 0.0. 0.0 0.0 0.0 0.0
NORTH-G1 318+282 IEEEG1 K2 T5 K3 K4 T6 K5 K6 T7 K7
0.0 7.5 0.23 0.0 0.6 0.0 0.47 0.0 0.0
Gv4 Pgv4 Gv5 Pgv5 Gv6 Pgv6
0.0 0.0 0.0 0.0 0.0 0.0

Capacity Type rperm rtemp tr tf tg velm gmax gmin tw at

0.05 0.5 10 0.05 0.5 0.2 1 0 2.5 1.25
Pgv0 gv1 pgv1 gv2 pgv2 gv3 pgv3 gv4 pgv4 gv5
SOUTH-G1 600 0 0 0 0 0 0 0 0 0 0
EAST-G1 1200 HYGOV tblade dturb qnl ttrip tn tnp db1 eps db2 gv0
WEST-G1 1200 99 0.5 0.08 0 0 0 0 0 0 0
pgv5 hdam bgv0 bgv1 bgv2 bgv3 bgv4 bgv5 bmax
0 1 0 0 0 0 0 0 0

IEEEG1: IEEE steam turbine/governor model (with deadband and nonlinear valve gain added)
HYGOV: Hydro turbine and governor. Represents plants with straight-forward penstock configurations and electro-hydraulic governors
that mimic the permanent/temporary droop characteristics of traditional dashpot-type hydraulic governors.

27
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 D. Wind power plant data
There are two kinds of wind power plant models in MAFRIT: a DFIG power plant with phasor dynamic model and the simplified wind
power plant model using controlled current source and algebraic expression to present the relationship between the nominal wind speed
and power.

i. Phasor dynamic model of DFIG

The DFIG parameters are grouped in four categories: Generator data, Converters data, Turbine data, and Control parameter.

1. Generator data
Table A-10. Generator Data of DFIG

Pn (W) Vn (V) Fn (Hz) Rs (p.u.) Lls (p.u.) Rr’(p.u.) Llr’(p.u.) Lm (p.u.) H (s) F (p.u.) p
90*1.5e6 575 60 0.00706 0.171 0.005 0.156 2.9 5.04 0.0.1 3

[Pn Vn Fn]: [Nominal power, line-to-line voltage, frequency]

[Rs Lls]: [Stator resistance, stator inductance]
[Rr’ Llr’ Lm]: [Rotor resistance, rotor inductance, magnetizing inductance]
[H, Fm, p]: [Inertia constant, friction factor, pairs of poles]

2. Turbine data
Table A-11. Turbine Data of DFIG

Pm (W) w_a (p.u.) w_b (p.u.) w_c (p.u.) w_d (p.u.) Pm_c Vw_c Kp max
90*1.5e6 0.7 0.71 1.2 1.21 0.73 12 500 45 2

28
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 3. Converter data
Table A-12. Converter Data of DFIG

Converter maximum P
L (p.u.) R (p.u.) IL (p.u.) Ph_IL (deg) Vdc (V) Cdc (F)
(p.u.)
0.5 0.15 0.15/100 0 90 1200 90*0.1

4. Control data
Table A-13. Control Data of DFIG

(dV/ (dP/d (dI/d

Mode Vref Iq_ref Kp1 Ki1 Xs Kp2 Ki2 Kp3 Ki3 Kp4 Ki4 Kp5 Ki5 dt) t) t)
max max max
Voltage
300/ 100/ 0.00
regulatio 1.0 0 1.25 0 1 0.05 1 100 0.3 8 100 1 200
100 200 2
n

Mode: Voltage regulation/VAR regulation

[Kp1 Ki1]: Voltage regulation mode/grid voltage regulator gains
[Kp1 Ki1]: VAR regulation mode/reactive power regulator gains
[Kp2 Ki2]: Power regulator gains
[Kp3 Ki3]: DC bus voltage regulator gains
[Kp4 Ki4]: Grid-side converter current regulator gains
[Kp5 Ki5]: Rotor-side converter current regulator gains

ii. Simplified wind power plant model

The simplified wind power plant model presented here generates power using a linear relationship between the nominal wind speed and
nominal power. When the wind speed reaches the maximal value, the wind power plant trips from the grid. When the wind speed is
between the nominal speed and the maximal value, the power is fixed to 1 p.u.

29
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.
 Table A-14. Parameter of Simplified Wind Power Plant Model

Maximal Wind Speed

Nominal Power(MW) Nominal Wind Speed (m/s)
(m/s)
360 13.5 15

30
This report is available at no cost from the National Renewable Energy Laboratory at [Link]/publications.