Applicable NERC ReliabilityStandards

Applicable NERC Reliability
Standards
Ryan D. Quint, NERC
Power Plant Model Verification and Testing Workshop
September 2016
Disclaimer
ͻ The material presented here is not intended as compliance

guidance. Its sole purpose is to describe the technical aspects of
related reliability standards to power plant modeling and
verification.
ͻ For questions related to compliance, please contact NERC
Compliance Department directly.
2 RELIABILITY | ACCOUNTABILITY
MOD Standards Framework
NERC MOD Standards

Modeling, Data, and Analysis
MOD-032
Data for Power System Modeling and Analysis
System Modeling Plant Modeling

- Load Forecasts - MOD-025-2 – Generator Capability
- System Components - MOD-026-1 – Volt/Var Control
- Reactive Devices - MOD-027-1 – Power/Frequency Control
- Transfers
MOD-033
Steady-State and Dynamic System Model Validation
Applicability of MOD Standards
Standard Applicability Effective

MOD-025-2 Generator Owners (and Transmission Owners) 7/1/2016
• Generators and synchronous condensers – 20 MVA
• Aggregate plant/Facility – 75 MVA
MOD-026-1 Generator Owners and Transmission Planners 7/1/14

and EI or QI: Individual or Aggregate > 100 MVA (R1,3,4,5,6)
MOD-027 WI: Individual or Aggregate > 75 MVA
7/1/2018
ERCOT: Individual > 50 or Aggregate > 75 MVA (R2)
MOD-032-1 Balancing Authorities Generator Owners 7/1/2015
Planning Coordinators Resource Planners (R1)
Transmission Planners Transmission Owners
7/1/2016
Transmission Service Providers LSE (R2,3,4)
MOD-033-1 Planning Coordinators, Reliability Coordinators, and 7/1/2017

Transmission Operators
*Directly connected to the Bulk Electric System (BES) with gross nameplate rating
Implementation Plans
Standard Implementation
Order approving standard on 3/20/2014
Effective Date (all requirements) – 7/1/2016
Implementation Plan:
MOD-025-2 • 40% after 2 years
• 60% after 3 years
Order approving standard on 3/20/2014
Effective Dates:
• R1, R3, R4, R5 – 7/1/14
MOD-026-1
• R2 – 7/1/18
&
Implementation Plan:
MOD-027-1
MOD-025-2:
Verification and Data Reporting of Generator Real
and Reactive Power Capability and Synchronous
Condenser Reactive Power Capability
MOD-025-2
Requirements
ͻ Ensure accurate information available for planning models

Generator gross and net Real (P) and Reactive (Q) Power capability
Synchronous condenser Q capability
ͻ Implementation:
FERC approved 3/20/14; Order effective 7/1/16
40% after 2 yrs, 60% after 3 yrs, 80% after 4 yrs, 100% after 5 yrs
ͻ Requirements:
R1: GO provides its TP with verification of P capability
R2: GO provides its TP with verification of Q capability
R3: TO provides its TP with verification of Q capability
Verify in accordance with Attachment 1
Submit completed Attachment 2
MOD-025-2
Attachment 1
ͻ Verification Specifications:
First verification for each unit must be a staged test
If operational data for different points recorded on different days,
designate earliest as verification date
Periodicity:
o At least every 5 years; or within 12 months of discovery of change affecting
P or Q capability > 10% and expected to last > 6 months
New units must be verified within 12 months of commercial ops date
o Applies to existing units that have been in long-term shut down
P capability and full load Q capability testing should be performed at
same time; however, separate testing is allowed.
Q capability testing should be scheduled for time advantageous for unit
being verified to demonstrate its Q capabilities
o While TOP takes measures to maintain plant’s system bus voltage at
scheduled value within tolerance
MOD-025-2
Attachment 1
ͻ Verification Specifications
Units < 20 MVA at plant > 75 MVA aggregate - individual or aggregate
Units > 20 MVA must be verified individually
Check auxiliary equipment in-serve for expected normal operation
AVR in-service for Q capability verification
Operational data within 2 years prior to verification date acceptable as
long as meets criteria and 90% of previous staged test
o Previous test must have demonstrated at least 50% of D-curve Q capability
o If previously test was unduly restricted, then another staged test required
MOD-025-2
Attachment 1
Pmax 2.1. P and Q capability

over-excited at normal
Qmax expected maximum P
output:
ͻ Synchronous condenser:
Pmax, Qmax for minimum
of 1 hour
ͻ Variable generating units:
Qmax at time of
verification; should have
at least 90% units online
MOD-025-2
Attachment 1
Pmin
Pmax 2.2. Q capability* max
Qmax over- and under-excited
Qmax for:
ͻ Normal Pmin, as soon as
limit is reached (nuclear
exempt)
ͻ Normal Pmax, as soon as
limit is reached
Qmin Qmin * Other than wind & solar PV
MOD-025-2
Attachment 1
Recorded data for verification:

1. Gross P and Q capabilities
2 2. Scheduled voltage provided by TOP, if applicable
3 3. Voltage at high- and low-side of GSU and/or
6,7 interconnection transformer(s)
4. Ambient conditions at end of verification period
3
that GO requires for corrections to P output for
1 different ambient conditions
5. Date and time of verification period (start and
end times)
6. GSU and/or system interconnection transformer
voltage ratio and tap settings
7. GSU losses (P and Q) if measurements taken at
high side
8. Whether staged test or operational data
MOD-025-2
Attachment 1
ͻ Considerations:
Verify at normal operating hydrogen pressure for hydrogen-cooled units.
Calculate GSU losses if measurements taken at high-side
o GSU losses may be estimated based on GSU impedance.
Simplified one-line diagram must show sources of auxiliary P and Q and
associated connections for each unit verified.
o Include GSU and/or system interconnection and auxiliary transformers.
o Show Q flow with directional arrows.
o If metering does not exist, provide engineering estimate.
If adjustment requested by TP, develop relationships between test
conditions and unit output so that the amount of P that can be expected
can be determined at different conditions, such as summer peak
conditions.
MOD-025-2
Attachment 2
ͻ One-line diagram
ͻ Table and summary of
verification information
ͻ Current and past test data
MOD-026-1:
Verification of Models and Data for Generator
Excitation Control System or Plant Volt/Var Control
Functions
MOD-027-1:
Verification of Models and Data for
Turbine/Governor and Load Control or Active
Power/Frequency Control Functions
MOD-026-1/MOD-027-1
Terminology
Resource Excitation control system or Turbine/governor and load

plant volt/var control control or active
function power/frequency control
Synchronous Includes generator, exciter, Includes turbine/governor and
Machine voltage regulator, impedance load control
compensation, and power
system stabilizer
Aggregate Includes voltage regulator & Includes active power/frequency
Generating reactive power control system control
Plant controlling and coordinating
plant voltage and associated
reactive capable resources
MOD-026-1/MOD-027-1
Requirements
R1: Each TP provides information to the GO upon request:

List of models acceptable to TP
Block diagrams and/or data sheets for acceptable models
Model data for GO’s existing units
R2: GO provides verified generator dynamic model(s) for each unit
Model verified by GO using one or more models acceptable to TP
Each verification includes the following:
o Unit’s model response matches recorded response (next page)
o Manufacturer, model number (if available), and type of system
– e.g., digital vs. analog, static vs. rotating exciter, plant controls
– e.g., turbine type, boiler type, fuel type, manufacturer and controls
o Model structure and data
– e.g., block diagram, time constants, gains, limits, generator data
o Outer loop controls – blocked or nonfunctioning controls or modes of
operation that limit response
MOD-026-1/MOD-027-1
System Events and Tests
Standard MOD-026-1 MOD-027-1

System “Voltage excursion from a Frequency excursion event, with
Event measured system disturbance” unit operating in frequency
– size not specified, should responsive mode:
have noticeable perturbation to • EI: ǻf +]
terminal voltage • TI: ǻf +]
• WI: ǻf +]
• QI: ǻf +]
Staged “Voltage excursion from a • Speed governor reference
Test staged test” – for example, change with unit on-line
voltage reference step test* • Partial load rejection test**
with unit online and PSS on/off
* PSS Off tests verify excitation system models while PSS ON tests verify PSS models.
** Differences in control modes between testing and final simulation model need to be identified. Most
controls change gains or have a set point runback which takes effect when the breaker opens. This can
skew results of load rejection tests if not properly accounted for and understood.
MOD-026-1/MOD-027-1
Requirements
R3: GO provides written response to TP after receiving from TP:

Notification that model is not usable
Comments identifying technical concerns with verification documents
Comments and supporting evidence indicating modeled response does
not approximate recorded response for three or more events
Response will include either technical basis for maintaining model,
model changes, or plan to perform verification
R4: GO provides revised model or plans to perform PPMV within
180 days of making changes to controls or equipment that alters
response characteristic.
MOD-026-1/MOD-027-1
Requirements
MOD-026-1:
R5: GO provides response to TP within 90 days following receipt of
technically justified* request to perform model review, including:
Details of plans to verify model
Corrected model data including source of revision
* TP demonstrates simulated vs. measured response does not match
MOD-027-1 / MOD-026-1:
R5/R6: TP provides written response to GO within 90 days of
receiving verified model that model is usable or not usable,
including:
Initializes without error
No-disturbance simulation results in negligible transients
Exhibit positive damping
09/21/2016 | ATLANTA, GA
ISO-NE System Operations

Perspective on Power Plant
Model Verification
Bilgehan Donmez
SENIOR ENGINEER | REAL-TIME STUDIES
ISO New England Is Part of a Larger Electric
Power System
NERC
North American Electric Reliability Corporation
• Eastern Interconnection spans Québec Interconnection
from Rocky Mountains to East NPCC
Northeast
Coast and Canadian Maritimes Power
Coordinating
Council
– Primarily alternating-current
(AC) transmission
– New England linked to rest of
Eastern Interconnection via
transmission ties to New York
and New Brunswick
• Tied to Québec only through

direct-current (DC) transmission
Eastern
• 2003 Blackout ushered in wide- Interconnection
area monitoring and mandatory Western
Interconnection
reliability standards ERCOT
Interconnection
2
System Details
• 7.1 million retail electricity customers drive

the demand for electricity in New England
(14.7 million population)
• Region’s all-time summer peak demand set
on August 2, 2006 at 28,130 MW
• Region’s all-time winter peak demand set
on January 15, 2004 at 22,818 MW
• Energy efficiency and behind-the-meter solar
slow the growth in summer peak demand to
0.3% annually and reverse the growth in
overall electricity demand to -0.2% annually
Note: Without energy efficiency and solar, the region’s peak demand is forecasted to grow 1.1% annually and the region’s overall electricity demand
is forecasted to grow 1.0% annually. Summer peak demand is based on the “90/10” forecast for extreme summer weather.
3
A Range of Generation and Demand Resources Are
Used to Meet New England’s Energy Needs
Existing and Future Resources
• 350 generators in the region (MW)
35,000
• 31,000 MW of generating capacity 31,000
30,000
• 11,500 MW of proposed
generation in the ISO Queue 25,000
– Mostly natural gas and wind
20,000
• 4,200 MW of generation has
retired or will retire in the next 15,000
11,500
five years
10,000
• 600 MW of active demand

5,000
response and 1,900 MW of energy 2,500
efficiency with capacity supply 0

obligations Existing Existing Demand Proposed
Generation Resources Generation
4
New England Has Significant Wind Potential
• Population and electric demand are

concentrated along the coast in central
and southern New England
• Most of the new wind farms (current and
upcoming in the interconnection queue)
are coming to the northern part of the
New England system where we have
limited short circuit strength.
• 12,000 MW of onshore and offshore wind
potential
– Preliminary screening eliminated wind
sites near urban areas and sensitive
geographic locations Electricity Demand
(e.g., Appalachian Trail)
• Transmission will be required to connect
potential wind resources to load centers Wind zones
in New England
5
New England’s Transmission Grid Is the Interstate
Highway System for Electricity
• 8,600 miles of high-voltage transmission

lines (115 kV and above)
New
• 13 transmission interconnections Hydro
Brunswick
to power systems in New York and Québec

Eastern Canada
• 16% of region’s energy needs met
by imports in 2015
• $7.88 billion invested to strengthen
transmission system reliability since
2002; $4.18 billion planned New
York
• Developers have proposed multiple
transmission projects to access
non-carbon-emitting resources
6
Ensuring Reliable Power System Operations Is a
Major Responsibility
• Maintain minute-to-minute
reliable operation of
region’s generation and
transmission system
• Perform centralized
dispatch of the lowest-
priced resources
• Coordinate and schedule
maintenance outages
• Coordinate operations with
neighboring power systems
7
Operational Limits
• ISO-NE Operation Support Services perform thermal, voltage and

stability studies to come up with operating limits for the control
room.
• Being at the tail end of the Eastern Interconnection, north to south
transfers in the northern part of our system is transient stability
limited. In New England, we have a few dozen stability based
SOLs/IROLs respected in real time.
• With our stability limits, modeling plays a key role. If we have
models that, we believe, do not do an adequate job of capturing
facility performance, we can keep a new plant from coming on-line
or limit its output.
• We can also limit an existing plant’s output to the point of keeping
it offline, and these could be indefinitely, because using these
models puts us in an “undefined system condition”.
8
Why is it important to have accurate models?
• To ensure our transient stability simulations provide accurate

system representation
• To prevent restricting units due to poor plant models
9
Operational Use of PMUs
• Real-time Oscillation Alarm --> Email

– Operation support engineers
• Write real-time operating procedures
• Communicate with the plant
• Post Event Analysis

– Data: PMU, SCADA, DFR
– TSAT simulations
– PhasorAnalytics
• Frequency Response Monitoring
– System
– Individual generator
– In-house developed Excel tool
– Evaluating PNNL’s Frequency Response Tool
10
How do we use PMUs? Examples…
• Event detection: Oscillation monitoring

– Example 1: Forced oscillations detected by PMUs
• Post-event analysis and power plant model validation
– Example 2: Sustained oscillations detected by PMU due to high gain
settings of the exciter system
11
12
Example 2: Event Details
• Normally cleared single phase to ground fault on a nearby 345

kV line followed by a 5 second automatic reclosing
• Poor damping at the plant was detected by the PMU
oscillation detection logic
• High speed (5 second) reclosing exacerbated the plant
oscillations
13
Unit Details
• Combined cycle plant:

2 gas turbines and 1
steam turbine
• ~550 MW total output
• PMU available at POI
14
PMU Data: Line Flow and Total Plant Output
15
Example 2 is a Success Story in using PMU data
to correct Plant Models
• We worked with the generator owner to analyze the results.
• The generator owner hired a consultant to check the
parameters of the dynamic model and updated the exciter
model parameters to better represent the plant behavior.
• The generator owner also decided to install a power system
stabilizer, which is scheduled to go in-service by the end of
September.
16
What is next?
• MOD 26 and 27
• MOD 33
• Online Transient Stability Analysis
17
Power Plant Model Validation
(MOD-26 & MOD-27)
• Batch Model Validation – 19 generators/plants with 1 click
– A MATLAB tool calls TSAT playback function
PMU & TSAT Real Power (MW) at POI

-1200
-1210
(MV )
Real Power (MW)
-1220
P
PMU Curve
ti
-1230
TSAT Volt-Theta Curve
R
TSAT Volt-Freq Curve
-1240
0 5 10 15 20 25 30 35
Time (s)
PMU & TSAT Reactive Power (MVar) at POI

130
120
110
Reactive Power (MVar)
100
90
PMU Curve
TSAT Volt-Theta Curve
80
TSAT Volt-Freq Curve
70
0 5 10 15 20 25 30 35
Time (s)
18
System Wide Model Validation (MOD-33)
• In-house developed bridging tool – Java

– Carries over real-time SE statuses to the planning case
– Compare power flow and simulation results with SE/PMU data
– Identify mismatches and investigate model accuracy
SE Case
NX-9 Bridging Modified

Mapping Tool Planning Case
Planning
Case
19
Pilot On-line Transient Stability Assessment
• Real-Time security
assessment Fixed data (updated periodically)
Network Dynamic
Generator
model in data in
mapping
planning planning
table
• Post event analysis Updated in real time
case case
– Good match with PMU EMS Snapshot Data Preparation Tool (DPT)
Dynamic
equivalent for
external area
data for the events Updated in real time

Fixed data (updated periodically) (optional)
tested Auxiliary files (e.g.,

interface file, contigency DSA Manager
file, transfer file, etc.)
– Event recreation
– Perform what-if analysis
TSAT Server 1 TSAT Server 2 TSAT Server n
20
21
Power Plant Model Validation
and Performance Monitoring
presented by Dmitry Kosterev
Bonneville Power Administration
NERC Power Plant Model Validation Workshop

September 2016
Power System Models
• Accurate power system models are required for
reliable and economic power grid operations
• Two large-scale power outages in August 10, 1996

the West on July 2 and August 10
1996
• Observed generator responses
were different from what was
expected
• Loss of confidence in abilities to
establish safe operating limits
• Interties were de-rated temporarily :
– California-Oregon AC from 4,800 to Failure of the power system model to
3,200 MW
reproduce what happened
– Pacific HVDC from 3,100 to 2,000 MW
Benefits of Generator Testing in WECC
WSCC required all generators >20 MVA be tested for
model validation
Benefits of WECC generator testing program are

indisputable:
• Great majority of dynamic models needed data revisions
• Structural model errors were detected – e.g. static exciters
represented with models used for rotating dc exciters
• Errors in control settings were identified and corrected, e.g.
error in PSS wiring
Need to sustain model validation efforts was evident

Governor Response Modeling
Generator testing program alone was not enough …
Validation studies done for August 10 1996 outage

showed optimistic frequency response
• WSCC implemented an interim measure – block governors
on all thermal units loaded above 90% of their capacity
In early 2000s, WECC was developing a frequency

response criteria
• System tests were done to validate governor response
across the West (May 2001 – all AGC systems were
suspended, planned dropping Grand Coulee and Hoover
generation)
• Validation studies showed the frequency response was still
optimistic
An effort was launched
in 2002 to classify
power plant operating
practices:
Frequency - Responsive
Under Load Control
Frequency Non-
responsive, or
Baseloaded
June 14 2004 Westwing Event
Malin Frequency, June 14 2004 West Wing event
60.1
Frequency response
60
representation improved
59.9 greatly
Frequency [Hz]
Actual
59.8
Malin Freq
59.7
59.6
COI Power, June 14 2004 West Wing Disturbance
6000
59.5
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Time [sec] 5500
5000
4500
Power [MW]
Actual
4000
COI
3500
Reasonable correspondence 3000
between actual and simulate 2500
power pick-up
2000
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Time [sec]
Frequency
response is not just
about system
frequency
Distribution of
frequency response
also affects post-
contingency path
loading and voltage
stability.
It is an issues in
WECC because of
unbalanced
frequency response
distribution
2006 WECC Generating Unit Model
Validation Policy
• WECC formalized its model validation requirements in
Generating Unit Model Validation Policy
• Data Requirements
– Data must be provided using grid simulator models
• Baseline model development
– Required for new plants, when equipment changes are
made, or when a model is found to be in error
• Periodic model verification
– Done every 5 years to make sure that models are up to date
• Reactive limit testing
NERC Reliability Standards
2007 – NERC started the development of model verification
standards
2013 – NERC approved:

– NERC MOD-025 – active and reactive power capabilities
verification
– NERC MOD-026 – generator and excitation control model
verification
– NERC MOD-027 – generator turbine control model
verification
2014 – NERC approved MOD-033 requiring plant operator

to provide accurate power plant model data
Baseline Model Development 10
Baseline Model Development vs. Periodic
Model Validation
• Baseline model development

• Needed to establish the correct model structure
• Needed to create initial model data set
• Periodic model verification

• Done to ensure that the models stay accurate and up-to-
date AFTER a good model baseline is developed
• Should not be a substitute for baseline model
development
Baseline Model Development
Equipment Model
Speed Reference
ZREF uop Gmax
Z 1 1 1
6 1 + s Tf
6 Tp s
ucl Gmin
Damping PILOT
SERVOMOTOR
s Tr
Rt 1 + s Tr
Rp
• Needed to establish the correct model structure

• Inspection of equipment and control settings
• Some staged tests are required
• Disturbance monitoring can complement model development
12
Mechanical Digital
Many modern digital controls map directly to models

13
Thermal Hydro
Practical limitations usually exist in large thermal plants

14
Baseline model development is an essential first step to
establish correct model structure and to develop an
initial model data set that includes:
• Inspection of equipment and control settings
• Some staged tests may need to be done
• Disturbance monitoring can be used to complement
baseline model development
Disturbance Monitoring 16
Disturbance Monitoring
POI Plant
POI Generator
Substation
DDR DDR
EFD
V I V I
IFD
17
Disturbance Monitoring
Plant Disturbance Monitoring

• Plant disturbance monitoring can complement baseline
model development
• Several mature applications are available
• EPRI PPPD is used widely by GOs in North America
POI Disturbance Monitoring

• POI disturbance monitoring can be used for periodic
model verification per MOD-026 and 027 once a good
baseline model is developed
• POI monitoring provides Transmission Planner an
independent way to validate model and generator
performance during disturbances
Using Disturbance Recordings for MOD
• Disturbance recordings can be used for compliance
with MOD-026 and 027 Standards
• All major grid simulators (PSS®E, PSLF, PowerWorld)
have disturbance play in capabilities
• Model verification must be done over multiple
disturbances
• Applications for data and model management are
available and mature
Using POI Disturbance Recordings
MODEL
Power Plant
Voltage and Frequency
MWs and MVARs

G
Controllable
Voltage and
Frequency
Voltage and frequency are played in as inputs

Active and reactive power are used as “measure” of success”
20
Successful model verification
Voltage and frequency are inputs

Active and reactive power are “measures of success”
Blue line = actual recording

22
Red line = model
Un-successful model verification
Voltage and frequency are inputs

Active and reactive power are “measures of success”
Blue line = actual recording

23
Red line = model
When Model and Actual Disturbance Data
Do Not Match
• Check powerflow model (and specifically transformer
impedances and taps)
• Check dynamic data record against test report (and
specifically decimal points)
• Beware of model data conversion and minus signs
• Most common model issues:
– Power System Stabilizer models
– Turbine control mode of operation / governor models
– Operating head in hydro turbines
– Generator inertia
– Deficiencies in model structure
• Things we do not normally model:
– Automatic Generation Controls, secondary controls, UEL
24
Transmission Planners
• MOD Standards require Transmission Planners to check
usability of the GO-provided models, but not model
accuracy
- Sets of automated tools are developed to stress-test
model usability, including voltage and speed reference
step responses, fault simulations, and by playing in
various disturbances
• It is a good practice for Transmission Planners to

– Perform independent model verification
– Transmission planners need to understand models and
associated data and how it affects
Transmission Planners:
Going Beyond the Standard
BPA approach has been “Trust but Verify”
• Systematic deployment of DDRs at power plants following 1996
outages
• Developed and deployed Power Plant Model Validation (PPMV)
application for model validation
• Helped many BPA customers with MOD compliance
• About 60% of models did not match disturbance recordings
even after baseline tests were performed
• Helped to drive model verification efforts, resulting in
significant model improvements
• Disturbance-based model verification studies are performed
following each large disturbance – I know how my generating
fleet performed within minutes after a system event!
DDR Deployment at BPA
BPA has PMU disturbance monitoring:
- Conventional –
- 13 plants, 133 generators, 21,645 MW of capacity
- Wind –
- 12 plants, 1,200 MW of generation
Data is streamed continuously to BPA control centers
Engineering and real-time applications developed and

implemented
27
Success Stories – Plant A
BEFORE AFTER
Blue = Actual, Red = Model
28
Success Stories – Plant B
BEFORE-2014 AFTER-2015
180 180
175 175
Active Power [MW]
Active Power [MW]

170 170
165 165
160 160
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Time Time
20 20
10 10
Reactive Power [MVAR]
0 0
-10 -10
-20 -20
-30 -30
-40 -40
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Time Time
29
Success Stories – Plant C
BEFORE-2014 AFTER-2015
240 238
238 236
Active Power [MW]

Active Power [MW]
236
234
234
232
232
230
230
228 228
226 226
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Time Time
10 10
5 5
0 0
-5 -5
-10 -10
-15 -15
-20 -20
-25 -25
-30 -30
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Time Time 30
Detecting Control Abnormalities
After baseline is established, performance data can be used to detect
generator control abnormalities
Power
700
Unexpected action from
680
660 Stabilizer failure plant MW controller

640
Power
Active Power [MW]
550
620
600 540
580
530
560
Active Power [MW]

540 Observed 520
Expected
520
510
500
2 3 4 5 6 7 8 9 10
Time (sec) 500
50 Observed
490 Expected
Reactive Power (MVAR)
480
0 10 20 30 40 50 60 70 80 90 100
0 Time (sec)
-50
Abnormal runback in
reactive power
-100
0 5 10 15 20 25 30 35 40 45 50
Time (sec)
31
Wind Power Plant Model Verification
32
Wind Power Plant Model Verification
Wind power plant dynamic models are complex and not
well understood by utilities or plant operators
– 7 models, more than 130 parameters
Phased approach is adopted in WECC:

• First, develop, validate and stress-test “typical” data
sets for various wind generator make and models
• Second, plant-specific data is updated with actual

settings and validated against tests and disturbances
33
Wind Power Plant Model Validation Example
Input:
POI Voltage
Measure of success:
POI Reactive Power
Measure of success:
34.5-kV substation
voltage
Summary
Accurate models are required for reliable and economic power
system operations
Baseline model development is needed to establish the model

structure and initial data sets
Using disturbance data is an effective method of model

verification, disturbance play-in functions are available in all grid
simulators, tools for model and data management are mature and
available
Transmission planners are encouraged to use disturbance

monitoring for independent model verification and performance
monitoring – tools are available
Collaboration between TPs, GOs and OEMs is encouraged on

model verification and performance/disturbance monitoring 35
Thank You
Power Pla nt Model Verifica tion a nd Testing Workshop
“Ba seline” Testing

• Synchronous Genera tors
– OC Sa tura tion Tests
– Inertia Tests
– Impeda nces a nd Time Consta nts Verifica tion Tests
• Excita tions Systems (a nd PSS)
– Va r Rejection Tests
– OC or On-Line AVR Step Tests
– Frequency Response Tests
• Turbine Controls
– Speed Reference Step Tests
September 2016
Gas Turbine and Steam

Turbine Modeling a nd Testing
September 2016
September 2016
Governor Test via Load Reference Step
Load Reference Step Test Sequence
0.5
+0.4
0.4
0.3
TNR Change (%)
0.2
0.1 0.0
0.0 0.0 0.0
0.0
-0.1 -0.2
-0.2
-0.3 -0.4
-0.4
-0.5
00:00
01:00
02:00
03:00
04:00
05:00
06:00
07:00
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
Time (mm:ss)
September 2016
6
Februa ry 22013
013
Governor Test via Load Reference Step
CTG Governor Test via Loa d Reference Step
September 2016
Sept. 2016
Sept. 2016
10
Sept. 2016
Practicalities of
Genera tor a nd Excita tion
Testing
Examples of Unexpected Events

• Don’t ha ve signa ls you need
• Didn’t get the signa ls you expected
• Ca n’t opera te a s you would like
12
September 2016
Don’t have the signals you need – only Ib
13
September 2016
Didn’t get the signals you expected
14
September 2016
Can’t operate as you would like
15
September 2016
Before You Do The Test

Two Often-Overlooked Test Prepa ra tion Exa mples

• Rea ctive Ca pa bility Curve with limits a nd protection
• CT Pha se Ca libra tion
17
September 2016
18
September 2016
CT Phase Errors
• P = f (cos) Q = f (sin)
• Therefore, Q more sensitive to CT pha se errors when
opera ting a t typica l power fa ctors.
Actual Measured Measurement Error
V (kV) I (kA) I Phase (deg) P (MW) Q(MVAR) pf CT Phase Error (deg) P (MW) Q(MVAR) P (MW) Q(MVAR)
18.000 5.556 0.0 100.00 0.00 1.00 2.4 99.91 4.19 -0.09 4.19
18.000 5.556 2.0 99.94 3.49 1.00 2.4 99.71 7.67 -0.23 4.18
18.000 5.556 4.0 99.76 6.98 1.00 2.4 99.38 11.15 -0.38 4.17
18.000 5.556 6.0 99.45 10.45 0.99 2.4 98.93 14.61 -0.52 4.16
18.000 5.556 8.0 99.03 13.92 0.99 2.4 98.36 18.05 -0.67 4.13
18.000 5.556 10.0 98.48 17.36 0.98 2.4 97.67 21.47 -0.81 4.11
18.000 5.556 12.0 97.81 20.79 0.98 2.4 96.86 24.87 -0.96 4.08
18.000 5.556 14.0 97.03 24.19 0.97 2.4 95.93 28.23 -1.10 4.04
18.000 5.556 16.0 96.13 27.56 0.96 2.4 94.89 31.56 -1.24 4.00
18.000 5.556 18.0 95.11 30.90 0.95 2.4 93.73 34.86 -1.38 3.96
18.000 5.556 20.0 93.97 34.20 0.94 2.4 92.45 38.11 -1.51 3.91
18.000 5.556 22.0 92.72 37.46 0.93 2.4 91.07 41.31 -1.65 3.85
18.000 5.556 24.0 91.35 40.67 0.91 2.4 89.57 44.46 -1.78 3.79
18.000 5.556 26.0 89.88 43.84 0.90 2.4 87.96 47.56 -1.91 3.73
18.000 5.556 28.0 88.29 46.95 0.88 2.4 86.25 50..6
50 60
50.60 -2
2.0
04
-2.04 3.66
66
66
3.66
18.000 5.556 30.0 86.60 50.00 0.87 2.4 84.43 53
5 3.58
58
53.58 -2
2.17
17
-2.17 3.58
58
3.58
19
Sept. 2016
Online Voltage Reference Step Test

CTs with 0.8 degree pha se error
20
September 2016
Experience in Testing –
A Consulta nts Perspective
Additional Requirements for Consultants

• Knowledge of ma ny different equipment types
• Efficient Project Execution
• Longevity
22
July 2013
23
September 2016
24
September 2016
25
September 2016
Thank You!
Dan Leonard
GE Energy Consulting
daniel1.leonard@ge.com
518-385-0165
Power Plants
Overview
John Undrill
NERC Workshop
September 2016
Does the plant type (hydro/thermal) matter?
If the issue is grid voltage control -

Not very much
The voltage control capability and behavior of

a low speed hydro generator is very similar to
that of a 3600rpm generator
Excitation systems are different in detail but

similar and equivalent in function
Generator protection differs in detail but

not in function
Does the plant type (hydro/thermal) matter?
If the issue is frequency control -

Very much
Gas turbines can change real power output

very quickly initially but must be rate limited
outside a responsive band
Steam turbine plants can change real power

by a small amount very quickly but then must
be rate limited
Most hydro plants are somewhat slower off

the mark but can maneuver over a wide range
much more quickly than thermal plants
Does the plant type (wind) matter ?
If the issue is voltage control

Yes - of course
Generator types are evolving - some can exert
practical reactive power control - some cannot
Interconnection conditions typically require

primary control at the point of interconnection
The point of interconnection is typically more

remote from the generator than in hydro/thermal
plants
Reactive power control effort may be exerted

by device (eg Statcom) separate from the
turbine-generators
Does the plant type (wind) matter ?
If the issue is frequency control

Yes - of course
Generator types are evolving - some can exert
practical real power control - some cannot
Unlike hydro/thermal plants maneuvering

domain is variable due to uncontrollable
natural causes (mother nature)
Nevertheless - control of real power at

practically useful rates of change is
possible
Temperature change issues should be minimal
Other issues - perhaps aerodynamic stability - could turn out to be

significant as experience builds up
Does the plant type (photovoltaic) matter ?
If the issue is voltage control

Yes - of course
If the issue is frequency control

Yes - of course
Converter types are evolving - some can exert

practical real/reactive power control - some cannot
Operational issues associated with providing

primary and secondary control are largely
unknown at present - may arise on much
faster time scales than those of interest in
electromechanical plants
Why should we be deeply concerned about the modeling
of ‘old-tech’ hydro and thermal plants when the
new incoming plants are wind and solar ?
The sun does not shine at night

Even mariners do not fully understand how mother nature
programs the wind
When nature does not provide sun, wind, water -

we will have to burn fuel
Even with ‘rapid’ growth and penetration forecasts -
The control of the power system will depend on and be

strongly influenced by hydro and thermal machines for
most of the way to our present planning horizons
Synchronous Generators
(Ef d − vt )
Ii =
Xd
Ef d Vt sinδ
Pe =
Xd
Vt (Ef d − vt )
Qe =
Xd
Real three phase Idealized d-q axis
machine machine
Generator
characteristics
Magnetization curve
The open circuit

magnetization curve
sets the base for the
per unit representation
of the machine -
if it is wrong -
everything is wrong
Generator
characteristics
Capability curve
Generator
characteristics
Vee curve
Generator
model
Steady state
operation
Sets up initial
conditions for
dynamic
simulations
Generator
model
Dynamics
Standard
Maxwellian
Simple principle
Devilish details
Used in derived
form (via Park
transformation) in
simulations
Basis of genrou/
gentpj
Generator model
T(s) torque applied at coupling

w(s) rotational speed
Efd(s) voltage applied at field winding terminals
ifd(s) current at field winding terminals
v(s) voltage at stator winding terminals
i(s) current at stator winding terminals
Z(s) operational impedance
Power Plant Modeling Basics
Mark W. Baldwin PPMVT Workshop 09-20-2016

PPM Basics – Dominion Perspective
2
'RPLQLRQ¶V)RVVLODQG+\GUR)OHHW§JHQHUDWRUV
Fleet Resides in 3 NERC Regions:

1) SERC
2) RFC
3) NPCC
Operated by 2 RTO’s
1) PJM
2) ISO-NE
Presentation Summary
3
¾ Stability Planners – The Significance of

Stability Studies
¾ Instabilities Associated with Power Plants
¾ Power Plant Modeling Basics – Generating

Unit Natural Response
Transmission Stability Planners
4
¾ Run Studies to Identify System Instabilities –

disturbance tests
¾ Apply Corrective Action in Real Time

Operations
¾ Order Transmission System Upgrades

(Element Additions, Modifications) Years or
Decades Ahead Operations
Generating Unit Instabilities
5
A case of Voltage
Instability –
inadequate leading Adjacent
MVAR capability Bus Cold
Line Loads
230kV
UT SST
Generator
with UEL G
Adjacent Plant Loads
6
A case of Voltage
Instability –
inadequate leading
MVAR capability
UEL attempts to
drive MVARs
NORTH by
increasing field
current
7
Adjacent Bus
Cold Loads
MW’s
A case of Speed Unknown
Instability –
inadequate MW
capability 230kV
UT SST
12 MW
GTG
G
Adjacent Plant Loads
PPM Basics - Overview
8
Natural Response of Generating

Unit Must Be Captured in Order to
Analyze Stability
Energy Storage Media Must be

Modeled to Capture Generating
Unit Natural Response
Energy Storage Media is primarily

a) Flux
b) Rotating Mass
PPM Basics – Synchronous Machine
9
A synchronous machine
model captures effects of
both flux and kinetic energy
Flux effects are captured

in a suite of inductances
and field winding
characteristics
Kinetic effects are

captured in machine
speed and inertia But both the flux and speed drivers must be modeled:
a) Voltage regulator / exciter
b) Governor / turbine
Voltage Regulator / Exciter Models
10
Voltage Regulator DC Commutator Exciter

Feedback Signal
Operator Input Exciter Output

DC Exciter Signal Flow Diagram
Turbine/Governor Models
11
Feedback signal
governor
Fuel valve Turbine
Operator characteristic characteristic
input
Turbine/Governor Signal Flow Diagram

Turbine/Governor Models (cont’)
12
Load-based
temperature Exhaust
limit temperature

Turbine/Governor Models (cont’)
13
Turbine
damping

PPM Basics Summary
14
Current Modeling Practices
15
¾ Compensator (voltage) and PSS Models always

available but compensators are not included in one
RTO’s Approved Models List (AML)
¾ UEL and other limiter models are sometimes
available but no limiters are included in RTO AML
¾ Governor deadbands very frequently not modeled on
STG’s. One RTO has 11 approved T/G model types –
only one of the 11 features db.
¾ Fast-valve schemes, FB load controllers never
modeled (though PSS/E does have T/G model types
that feature fast-valve)
Reference Material
16
¾ IEEE Technical Report, PES-TR1, “Dynamic Models

for Turbine-Governors in Power System Studies,”
Power System Dynamic Performance Committee, et
al. January 2013
¾ IEEE Std 1110 IEEE-2002 (R2007), Guide for
Synchronous Generator Modeling Practices and
Applications in Power System Stability Analyses
¾ IEEE Std 421.5-2016, IEEE Recommended Practice
for Excitation System Models for Power System
Stability Studies
¾ PSS/E 33.3 Model Library (and associated manuals),
Siemens PTI, September 2012
PPM Basics
17
Questions
Modeling Improvements
Initiative Update
Mohamed Osman, NERC
September 2016
Advanced Analytics
Goals:
ͻ Improve interconnection-wide models
ͻ Address modeling gaps
ͻ Develop modeling guidance
ͻ Understand changing grid characteristics
ͻ Increase industry awareness and expertise
ͻ Collaborate with entire utility industry
ͻ Engage with other industry groups
Case Quality Metrics Assessments
ͻ Case Quality: Reasonableness of the data for individual Element

models that comprise the powerflow and dynamics cases.
Powerflow Metrics: assessment of quality of powerflow base case as
developed by interconnection-wide case creation entities
Dynamics Metrics: assessment of quality of associated dynamics case –
component models
ͻ Assessments
Phase 1 (2015): Initial development of criteria and testing; main focus on
powerflow case
Phase 2 (2016): Re-assessment of Phase 1 metrics; additional metrics, focus
on dynamics
Phase 3 (2017): At a minimum, assessment of Phase 1 and 2 metrics on new
cases (trending); possible new metrics
Dynamics Metrics
Phases 1 and 2
Phase Dynamics
Phase 1 • Generator netting
(2015) • Generator classical model representation
• Consistence reactance values
Phase 2 • Generator inertia and time constants
(2016) • Saturation factors and “severe” factors
• Generator speed damping coefficient
• Turbine-governor time constants
• Turbine power development fractions
• GAST models
• DC exciter self-excitation parameters
• WT3E electrical wind speeds
Phase 2 Metrics Assessment
Observations
ͻ Generator Reactive Limits

All interconnections, units at Qlim
ͻ Generator Reactive Limits Power Factor
Unreasonable Qmax or Qmin based on Pmax
ͻ Generator Time Constants
ͻ Generator Inertia Constants
Units with very small inertia constant
ͻ Saturation Factors
Many units with saturation factors outside reasonable (and severe) ranges
Phase 2 Metrics Assessment
Observations
ͻ Speed Damping D
Non-classical gen models with non-zero damping value
ͻ Turbine-Governor Power Development Fractions
Units with governor model with fractions that do not add to 100%
ͻ GAST Model
NERC developing Modeling Notification for GAST; predominant in EI
ͻ Self-Excitation Parameter for DC Exciters
Many units with small positive rather than small negative value
Example Dynamic Data Results
Dynamics Metrics
Case SUM WIN SLL
Phase Metric Perf Score (%) Perf Score (%) Perf Score (%)
Gens without Models 142 / 3402 4.17 189 / 3347 5.65 153 / 3401 4.50
Netted Gens with Models 3 / 2616 0.11 14 / 2161 0.65 2 / 1892 0.11
Phase I Netted Generators 12 / 2616 0.46 22 / 2161 1.02 7 / 1892 0.37
Gens with Classical Models 3 / 4063 0.07 3 / 3985 0.08 3 / 4063 0.07
Inconsistent Reactances 110 / 3351 3.28 123 / 3359 3.66 110 / 3342 3.29
Inconsistent Time Constants 213 /3351 6.36 217 / 3359 6.46 213 / 3342 6.37
Unreasonable Inertia Constants 622 / 3354 18.55 622 / 3362 18.50 619 / 3345 18.51
Unreasonable Saturation Factors 1869 / 3351 55.77 1884 / 3359 56.09 1859 / 3342 55.63
Severe Saturation Factors 243 / 3351 7.25 255 / 3359 7.59 243 / 3342 7.27
PSS but no Excitation 18 / 1726 1.04 22 / 1711 1.29 12 / 1718 0.70
Phase II Inconsistent Speed Damping 313 / 3351 9.34 340 / 3359 10.12 312 / 3342 9.34
Inconsistent Lead-Lag Time Const 24 / 1444 1.66 34 / 1454 2.34 24 / 1430 1.68
Erroneous Power Dev Fractions 21 / 224 9.38 21 / 238 8.82 21 / 224 9.38
GAST Models 30 / 2337 1.28 31 / 2319 1.34 30 / 2322 1.29
DC Exciter Self-Excitation Errors 52 / 665 7.82 52 / 696 7.47 52 / 662 7.85
Inconsistent Type III Wind Speeds 1 / 102 0.98 1 / 77 1.30 1 / 103 0.97
Appendix Materials
Supporting Industry based on Findings
1.4
1.2
0.8
0.6
0.4
Exponential Characteristic
0.2
Quadratic Characteristic
• Tool developed by the USACE for

Unsaturated Characteristic
0
0 0.5 1 1.5 2 2.5 3
L I
ad fd
fitting open circuit and V-curve

test data for GENTPJ model
• Posted to NERC SAMS page (HERE)
with instruction manual (HERE).
MOD-032 Feedback Loop
MOD-032
Case Quality
Designee
Assessment
Review
Updated
New Base
Model
Cases
Checks
Coordinate
w/ Entities
Plant Level Controls and Protection
Modeling Improvements
ͻ Ensure models sufficient to recreate plant behavior; explore

known/historical events and required modeling improvements to
represent beforehand and post-mortem.
Identify modeling gaps and possible new models to address gaps
Plant-level, turbine, and boiler controls and protection
Short- and mid-term post-disturbance behavior
Modeling Notifications
ͻ Notifications posted HERE

ͻ Proactively address modeling
issues used across industry
Engineering Develop Feedback Notify

Identify Problems
Analysis Guidance with SMEs Industry
Case Fidelity
Renewable Energy Modeling
ͻ Growing penetration of renewables across North America

ͻ Growing importance for accurate and robust modeling
ͻ Clear outcome and next step of NERC ERSWG and DERTF
ͻ Need concerted focus industry-wide on renewable models
ͻ Share and develop modeling practices
ͻ Educate system planners on renewables models available
ͻ Ensure uniform robust implementations
ͻ Tackle new or updated renewables modeling techniques
ͻ Understand impacts and concerns with renewables in weak grids
Plant Modeling Focuses
ͻ Non-uniform software implementation issues

ͻ Frequency response and ‘baseload flag’ implementation
ͻ Auxiliary equipment limit modeling
ͻ Phasing out of obsolete models
ͻ Improvements in case fidelity
Where Can We Help?
Disturbance-Based PPMV
Ryan D. Quint, NERC

September 2016
MOD Standards Drivers
ͻ Disturbance-based model verification is one method to verify

models used for planning and operating the BPS
Not a required practice; rather an alternative method
May be an effective means of compliance with MOD standards,
depending on circumstances
ͻ MOD-026-1 language, for example:

…“model response matches the recorded response…from either a
staged test or a measured system disturbance”
…“modeled response did not match the recorded response to a
transmission system event”
…“following receipt of a technically justified unit request” from the TP…
o Technical justification: …“demonstrating that the simulated unit or plant
response does not match the measured unit or plant response.”
Disturbance-Based PPMV Overview
Data Recording Locations
PPMV Playback Model
ͻ Major software vendors now have “playback model”

ͻ Injects voltage and frequency (or angle) into simulation
ͻ Separates tested model from rest of system at point of
measurement
Disturbance-Based
Verification Comparison
PLAY-IN DATA (INPUTS)
MEASURES OF SUCCESS (OUTPUTS)
Disturbance-Based PPMV Process
1. System Disturbance Measurements
2. Steady-State Case Creation
3. Dynamics Case Creation
4. Transient Stability Simulation
5. Model vs. Actual Comparison
Independent Verification
ͻ Which data is correct? Turns out neither were correct…

ͻ “1 good measurement is worth 1000 expert opinions”
What a Good Model Looks Like
ͻ Approximates general shape of response very well

ͻ Minor differences between events
What a Bad Model Looks Like
ͻ Does not approximate general shape of response well

ͻ Substantial differences in comparison (between events)
Verification vs. Calibration
ͻ Verification: yes/no check on response of model vs. actual

Effective means for ensuring modeled response approximates actual
Proven to identify plant issues such as controller failure or mode of
operation different than modeled
Should be used over multiple events
vs.
ͻ Calibration: tuning model to match measured response
Needs to be supervised by engineering judgment
Requires very good understanding of plant behavior and baseline model
Should not be used to replace baseline testing
Should not solely rely upon numerical curve fitting techniques
Beyond Standards:
Performance Monitoring
Reliability Guideline:
PPMV using PMUs
ͻ Guideline approved by NERC

Planning Committee Sept 2016
ͻ Overview of mechanics and
considerations for performing
disturbance-based verification
ͻ Appendix of available playback
model and tools
ͻ First step in supporting industry
with model verification
Workshop Objectives
Ryan D. Quint, NERC

September 2016
Workshop Objectives
ͻ Bring together array of industry expertise

ͻ Understanding plant dynamic behavior, modeling, and controls
ͻ Baseline testing procedures
ͻ Operational limitations
ͻ Manufacturing perspective
ͻ Utility experience
ͻ Industry tools and practices
ͻ Technical considerations for MOD-025, -026, and -027
ͻ Reliability benefits of accurate modeling
ͻ Path forward and next steps
Key Takeaways and
Next Steps
Ryan D. Quint, NERC
September 2016
Key Takeaways
ͻ Modeling
Accurate modeling critical for reliable operation and planning of the bulk
power system
Power plant controls and technologies are evolving and the models need
to reflect these changes accordingly
Plant-level controls can and should be modeled, as appropriate, if they
affect the unit in the dynamics timeframe
Accurate models essential for reliable and economic operation of the
bulk power system
ͻ Model Verification
Generator testing programs have benefits beyond standards compliance
Sustained model verification proven beneficial for GOs and TPs
“Baseline testing” needs to be defined and/or used appropriately
o Development of a baseline model via test and/or calculation is essential
Key Takeaways
ͻ Model Verification (cont.)

Reverification of that baseline model can be either by additional testing
or measurement-based approaches
Planners should require sufficient, clear, and comprehensive reports of
the development of a baseline model and/or reverification – “trust but
verify”
Technical challenges with MOD standards that should be addressed from
engineering standpoint – lessons learned, examples, guidance, etc.
ͻ Disturbance-Based PPMV
Reliability Guideline developed and approved by NERC PC
Used as a yes-no verification step of power plant model performance
Collaboration and coordination between T and G has proved very
effective
Proving effective for renewable models as well, more focus needed from
wind and solar
Key Takeaways
ͻ Plant Performance
Boiler and turbine controls/interactions should be considered in more
detail to determine an appropriate level of modeling, as applicable
Frequency response and adequate modeling of FR a continued concern
with continued NERC focus
Plant-level controls should not interfere with sustained primary
response of governor and excitation systems
ͻ Transmission Planning and Operations Perspective
Auxiliary protection and performance needs to be better understood by
the planners in system studies – understand vulnerabilities of ride-
through limitations
System operators can and have limited or shut off units if models are
inadequate to represent reality – need representative models for
reliability
Potential Next Steps
ͻ Develop standards guidance material for developing and

verification of models as per MOD standards
ͻ Task force focusing on model verification (testing) and
generator capability
ͻ Regional workshops on verification and testing for related MOD
standards
ͻ Coordinate more closely with manufacturing industry – create
forum for open bi-directional engineering discussion
ͻ Focus areas:
Plant auxiliary limitation modeling
Removal of obsolete/non-representative models
Case quality correction – saturation, model parameters, limits, etc.
Logistics
ͻ Survey to workshop participants

Workshop delivery & material
Technical comments/considerations that NERC should look into
ͻ Materials will be posted shortly to NERC website
Will send email to participants with link and list of registrants – contact
information redacted
Overview of EPRI
PPPD Tool
Anish Gaikwad
agaikwad@epri.com
Pouyan Pourbeik
ppourbeik@peace-pllc.com
NERC Workshop
September 2016
© 2016 Electric Power Research Institute, Inc. All rights reserved.

BACKGROUND ON PPPD
2007 EPRI started a fresh approach to R&D on power plant
model validation – not based on any previous R&D
2007 – developed first proto-type of the Power Plant

Parameter Derivation (PPPD) tool, for automating model
parameter derivation based on staged testing
2008 to 2009 – extended in a novel way, to allow the use of

on-line data (either disturbance data or on-line step tests)
Since 2010 EPRI has supported the tool through a User’s

Group, which current has 31 members (utilities, ISOs, etc.) –
3 are international members
2
PPPD CAPABILITIES & FUNCTIONALITY
Power Plant Parameters Derivation (PPPD) is a stand alone
executable software tool
Licensed through the EPRI Modeling and Model Validation

Tools User’s Group (MMVTUG)
The tool allows user’s to validated generator, exciter and

turbine-governor models using either:
– Measured field testing data (e.g. off-line tests)
– On-line voltage regulator (and turbine-governor speed-ref) step tests
– Digitally recorded disturbance data during system disturbances
while unit is on-line (e.g. from PMU or DFR)
3
HOW IT IS APPLIED
The novel aspect is the ability to use on-line disturbance
data for model validation
It can also be used to process data collected during
“emulated disturbances”, e.g. on-line voltage regulator step
tests, digital turbine-governor speed-reference step tests
Grid Disturbance Measured
Vt (pu)
0.96 Fitted
0.95
0 20 40 60 80 100 120
PMU/DFR
Ifd (pu)
2
at Plant 1.95
0 20 40 60 80 100 120
2.1
Vfd (pu)
2
1.9
0 20 40 60 80 100 120
Time (seconds)
4
DATA RECORDING AND CAPTURE
A B C
Three phase generator

stator voltages from
secondary of PTs
PT Two (or three) generator stator currents from

secondary of CTs
CT
Digital Fault
Generator Recorder
(can be a
dedicated DF or
an inbuilt DFR in
Field current measured
exciter)
Field across the field shunt
Winding
Digital Fault Recorder

(DFR)
HV GSU (from PMU or DFR)

Excitation System
Field voltage measured

From excitation system output
5
VALIDATION PROCESS
Collect Disturbance
Data from DFR
PPPD Scale data into per

unit format and
import into software
tool
Automated Software Tool
Adjust parameters Simulate event

based on non-linear
recursive least-
square algorithm
Compare simulated
No and measured data.
Do they adequately
match?
Yes
Finished
6
KEY BENEFITS OF PPPD
Many practical aspects of generator model validation (based
on over a decade of experience in generator testing) are
built into the tools functionality
Can be used with both on-line and off-line data
Simple to learn and use, with GUI
IEEE standard exciter models built-in; outputs in GE PSLFTM

and Siemens PTI PSS®E format
User’s group to share experience and learning
7
EXCITER/GENERATOR RESPONSE
On-line, disturbance monitoring
On-line, full-load, step test
8
EXCITER/GENERATOR RESPONSE
(NO GENERATOR FIELD DATA)
0.99
0.98
0.97
Voltage (pu)
0.96
0.95
Measured
0.94 Optimized Fit
Initial Estimate
0.93
0 10 20 30 40 50 60
Time (seconds)
9
GOVERNOR RESPONSE
Gas Turbine Response fit

with GGOV1 using PPPD
105
Measured
Fitted
100
95
P (MW)
90
85
80
75
0 50 100 150 200 250
Time (seconds)
10
GOVERNOR RESPONSE
Large steam-turbine
11
EXPERIENCE WITH PPPD
User’s Group since 2010
Presently 31 group members (utilities, ISOs, GOs) – 3 are
international
Amongst the members well over 100 synchronous
generators validated (generator + exciter)
In the past couple of years UG members have started
working on turbine-governor model validation and tens of
units validated
Applied both using on-line disturbance monitoring (DFR or
PMU) and on-line step-tests
Applied in WECC, ERCOT and EI
12
CONTINUED WORK
The tool is constantly updated – in 2016 Version 9.0 will be
released
Updates to PPPD in 2016 include:

– Addition of wind and PV “generic” models (previously EPRI REMV)
– Addition of static var system “generic” models (previously EPRI
SVSMV)
13
REMV TOOL – WILL BE PART OF PPPD
IN VERSION 9.0
So far applied to dozens of WTGs and a few PVs

Also, applied to two Wind Power Plants
14
SVSMV TOOL – WILL BE PART OF
PPPD IN VERSION 9.0
PMU
Vt VtI
It ItT
SVS
Over half a dozen SVCs/STATCOMs validated
15
OTHER TOOLS
The MMVTUG includes also load modeling tools, which are
outside the scope of the current discussion
EPRI LMPD – allows to estimation of load parameters for the

composite load model using disturbance data
EPRI LCET – allows for creation of composite load model

data cards based on climate and demographics data, etc.
16
Together…Shaping the Future of Electricity
17
Generator Baseline Testing
Instrumentation and Steady State Calculations
Greg Brooks
US Army Corps of Engineers
Walla Walla District
NERC Power Plant Model Verification and Testing Workshop
September 21, 2016
US Army Corps of Engineers

BUILDING STRONG®
Agenda
Data Collection
Generator Steady State Test Procedures

and Parameter Fitting
BUILDING STRONG®
Data Collection
BUILDING STRONG®
Measurement Points
CT IT
PT
VT
Field
Vfd
Stator
Ifd
Turbine Gate Position

Blade Angle (Kaplan)
BUILDING STRONG®
General Considerations
Use calibrated transducers
Ź The meters may not be
Your test gear is the cheap part
Ź Minimize setup time
• Good connectors (don’t wire your gear in the field)
• Labeling and color coding
• Use good cable management
Ź Have a bulletproof, efficient calibration and
signal check procedure
BUILDING STRONG®
Noise Mitigation
Shielded Twisted pairs
Ź Reduces EMI
Keep unshielded leads short and twisted
Use balanced differential signals
BUILDING STRONG®
Stator Values Transducers
VT – Use a little transformer
Ź Scales and Isolates
IT - Passive Clamps
Ź No batteries
Ź Contribute some phase shift
• Correct with linear interpolation
• Park Transformation
• Find the power angle, add the correction, then
recalculate P and Q using S and the new power
angle
BUILDING STRONG®
Field Values
Avoid clearance issues in the exciter
Ź Measure at the meters or slip rigs
Use adequate isolation transducers
ŹVERY IMPORTANT!
BUILDING STRONG®
Transducer Calibration
Be calibrated, the meters on the board
may not be.
Voltage drop when ifd isn’t measured
directly at the shunt
Ź Use an analog meter and don’t lift any leads
RC
A Vm, Rm Shunt
RA
BUILDING STRONG®
Governor Values
Gate and Blade (Kaplan)
Ź Auxiliary outputs from digital governors
• Be careful of voltage references!
Ź Linear string transducers
BUILDING STRONG®
Data Acquisition
Sample Rate: 2 kHz
Simultaneous or Sequential
Ź The closer the better
Differential inputs reject common mode
noise
16 Bit is sufficient
Real time data
BUILDING STRONG®
Essential Requirements
Acquisition Hardware
Acquisition Software
Data Processing Software
Simulation Software
6 hours of time on the unit at varying
operating points
Lots of time back in the office
BUILDING STRONG®
Generator Steady State Test
Procedures
&
Parameter Fitting
BUILDING STRONG®
Generator Model
GENTPJ
BUILDING STRONG®
Steady State Parameters
GENTPJ
ifdo - air gap line field current at 1.0 p.u. VT

S1.0 - open circuit magnetic saturation factor at 1.0 p.u. VT
S1.2 - open circuit magnetic saturation factor at 1.2 p.u. VT
Xd - D axis synchronous reactance, pu
Xq - Q axis synchronous reactance, pu
Xl - stator leakage reactance, pu
kis - stator current saturation factor
BUILDING STRONG®
Open Circuit Saturation Curve
1.2
Duration - 20 minutes
Terminal Voltage, per unit

1
0.8
Purpose 0.6
0.4
Ź Estimate ifdo, S1.0, S1.2 0.2
Hazards
0
0 200 400 600 800
Field Current (Amps)
Ź Stationservice voltage limits (if connected)

Ź Extended time at low VT
Ź Exceeding maximum VT
BUILDING STRONG®
Procedure
Generator Status: offline at rated speed
Manually adjust VT in steps
Ź Only adjust in one direction until extreme is reached
Example
Ź From 14.8 kV down to 8 kV, then up to 14.8 kV
Ź Use 200 V steps waiting for 5 seconds between
Record VT and ifd
BUILDING STRONG®
1.2
ifdo
1
Hysteresis
0.8
0.6
Measured
Air Gap Line
Saturation Curve
0.4
Parameters: ifdo, S1.0, S1.2

0.2
0
0 100 200 300 400 500 600 700 800 900
Field Current (Amps)
BUILDING STRONG®
1.4
Steps
1.2 1) Estimate ifdo ifd1.2 itest1.2
2) Saturation Factors
1 ifdo itest1.0
0.8
_1.0=( _ 1.0 _ 1.0)/ _ 1.0
_1.2=( _ 1.2 _ 1.2)/ _ 1.2

0.6
0.4
0.2
Air Gap Line Saturation Curve
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Field Current, per unit
BUILDING STRONG®
Quadratic Model
. .
.
= =
. .
< =

Ź = 1+
Vary parameters in blue to find best fit
BUILDING STRONG®
V-Curve & Reactive Limits
0.6
0.5
Reactive Power, per unit

0.4
Duration - 60 minutes 0.3
Purpose 0.2
Ź Estimate Xd, Xq, Xl, Kis 0.1
Ź Verify reactive limits 0

0 0.5 1 1.5 2 2.5
Hazards
Ź Motoring
Ź System conditions put limits beyond safe operation
• Excessive VT, IT, or ifd
• ifd below 1/3 of ifd at 0 VAR, and MW at test
BUILDING STRONG®
V-Curve & Reactive Limits
Procedure
Perform at +0, 70%, and 90% MW
Start at 0 MVAR
Slowly lower MVARs toward lower limit
Ź Stop at reactive limit or discomfort
Slowly raise MVARs toward upper limit
Ź Stop at reactive limit or discomfort
Slowly lower MVARs to 0 MVAR
Adjust MW and repeat
BUILDING STRONG®
Generator Capability Curve
V-Curve & Reactive Limit Test
0.8
Reactive Power, Per Unit
0.3
Stator Heating Limit

Active Power, Per Unit
Overexcitation Limiter
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Underexcitation Limiter
-0.2
+0% Power Test
70% Power Test
90% Power Test
-0.7
-1.2
BUILDING STRONG®
V-Curve
0.6
Parameters: S1.0, S1.2, Xd, Xq, Xl, kis

0.5
Absolute eactive Power, per unit
0.4
0.3
Measured P=0%
0.2 Measured p=70%
Measured P=90%
Model
0.1
0
0 0.5 1 1.5 2 2.5
BUILDING STRONG®
Automatic Curve Fitting
Never a substitute for engineering
Gets you close very quickly
Starting point
Ź Factory Specifications
Ź Prior test data
Ź Pay attention to the per unit base
BUILDING STRONG®
Starting Point
Manually Find ifdo
Manually Find ifdo
Manually Find ifdo
Manually Find ifdo
Manually Find ifdo
Manually Find S1.0
Manually Find S1.0
Manually Find S1.0
Manually Find S1.0
Automatically Solve
Use the Solver For the Rest
Questions?
BUILDING STRONG®
Generator model
Generator model - genrou
Generator model - gentpj
Generator models - genrou gentpj
Different ancestry
genrou
- used at GE
- descended from - gensal
- analog computer modeling
gentpj
- used at Westinghouse
- probably originated in digital form
Different assumptions in deriving useable digital models

from the basic Maxwellian starting point (several slides ago)
Both use open circuit magentization curve as the basis for representing
saturation throughout the structure
Use different approaches to recognizing effect of saturation
Differing assumptions regarding behavior of currents in rotor circuits
genrou assumes sudden change in stator current

affects ‘both’ rotor circuits instantaneously
gentpj assumes sudden change in stator current

affects only ‘surface’ rotor body circuit
instantaneously - with delay in effect on field/deep
rotor body circuit
genrou and gentpj are both approximations
They use different assumptions and result in slightly different approximations
Stator current affects magnetic saturation in two distinct ways

a) it modifies the ‘marco’ spatial distribution of flux around the
machine
b) it modifies the ‘micro’, spatial distribution of flux density in
the stator teeth
genrou and gentpj with Kis=0 both recognize effect a)

gentpj with Kis>0 has been introduced to provide approximate
representation of effect b)
genrou and gentpj are far from the only possibilities for generator modeling
in fact, I prefer and a different model for much of my own work
Stator current affects magnetic saturation in two distinct ways

a) it modifies the ‘marco’ spatial distribution of flux around the machine
b) it modifies the ‘micro’, spatial distribution of flux density in the stator teeth
genrou and gentpj with Kis=0 both recognize effect a)

gentpj with Kis>0 has been introduced to provide approximate
representation of effect b)
1 1
2vbus1
1vt1
0.95 0.95
0.9 0.9
0 5 10 15 20 0 5 10 15 20
180 100
160 80
1pg1
1qg1
140 60
120 40
0 5 10 15 20 0 5 10 15 20
6 3.5
5
3
1efd1
1if1
4
2.5
3
2 2
0 5 10 15 20 0 5 10 15 20
3 2
1.5
2
1elmv1
1elmt1
1
1
0.5
0 0
0 5 10 15 20 0 5 10 15 20
Time, sec Time, sec
junkf.cha / junkj.cha
Generator develops synchronizing and
damping torques
Generator develops synchronizing and
damping torques
5 5
4 4
3 3
Ks(f)
Ks(f)
2 2
1 1
0 0
10 -3 10 -2 10 -1 10 0 10 1 10 2 10 -3 10 -2 10 -1 10 0 10 1 10 2
1.4 0.5
1.2
0.4
1
0.8 0.3
Kd(f)
Kd(f)
0.6 0.2
0.4
0.1
0.2
0 0
-3 -2 -1 0 1 2
10 10 10 10 10 10 10 -3 10 -2 10 -1 10 0 10 1 10 2
Frequency, Hz Frequency, Hz
Strong network connection Weak network connection

We do care about damping !!!
Generator Low Voltage
Ride-Through
NERC Power Plant Model Verification and Testing

September 20-21, 2016
Robert J. O’Keefe
American Electric Power
22
TPL-001-4 R3.3.1.1 & R4.3.1.2 – Generator Low
Voltage Tripping
“Tripping of generators where simulations show

generator bus voltages or high side of the GSU
voltages are less than known or assumed generator
low voltage ride through capability. Include in the
assessment any assumptions made.”
33
Voltage PRC-024-2 Attachment 2

44
Voltage Tripping
If PRC-024-1 R2 & Attachment 2 applies to generator

voltage tripping relays only (Might perhaps be better titled “Generator Voltage
Relay Setting Curve”)…
What about other possible mechanisms of generator low

voltage tripping?
Specifically, drop out of critical plant aux equipment?
Auxiliary bus UV protection & equipment contactor drop out are the UV
tripping mechanisms identified by AEP Generation
55
Voltage Tripping
PRC-024 SDT retreated from comprehensive voltage ride-

through applicability and dealt exclusively with generator relay
settings that affect voltage ride-through
Left unaddressed are voltage ride-through questions related

especially to plant auxiliary equipment essential to unit
operation, or required by environmental regulations for unit
operation
66
Voltage Tripping
AEP Generation Remarks:
“Addressing PRC-024-1 for the generators alone is a massive task; adding the auxiliary
equipment would require an implementation plan years longer than what is approved”
“There is no definitive standard or guide in industry regarding contactor drop out,

however, 82% is a reasonable estimate based on Generation experience”
“There is not a great deal of certainty around when units will trip. I expect that all units
will be different”
“Testing for ride through was perceived to be quite difficult which is why the drafting
team, urged by industry comment, did not require it. Some additional measures were
added into the [PRC-024] standard at one point but had to be taken back out due to
industry push back”
77
Voltage Tripping
Good generating plant design provides for supply of

critical plant loads from generator bus, NOT THE
SWITCHYARD OR HIGH SIDE OF GSU!
If that is the case, may be rare that auxiliary equipment

trip voltage would be reached outside of close-in faults
88
Challenges In Determining Generator Ride-
Through In Extreme Event Scenarios
• Wide differences in generating unit performance / Incomplete data
• Auxiliary bus UV protection:
• UV settings, time delays, whether a unit even has auxiliary bus protection vary widely
• Known UV settings at or below 72 percent
• Time delays vary between instantaneous to 10 seconds or more
• Contactor dropout of critical generating unit auxiliary equipment
• UV level below which AEP Generation experience says contactor dropout may be
expected: 82 percent
• No standard on contactor dropout
My Interim TPL recommendation: 85 percent at generator terminals

exceeding normal fault clearing duration
99
Sample Events
Following are plots of generator bus voltage (low side of GSU) from
stable delayed clearing fault scenarios (754 Data Survey) and FIDVR
events
What may be concluded about generator

ride-through in these instances?
10 10
Will it Trip?
1.2
Per Unit Voltage
Time
Scale:
0 – 2.5
seconds
.20
11 11
Will it Trip?
12 12
Will it Trip?
13 13
Will it Trip?
14 14
Will it Trip?
15 15
Will it Trip?
16 16
Will it Trip?
17 17
Will it Trip?
18 18
Will it Trip?
19 19
Will it Trip?
20 20
Will it Trip?
21 21
Will it Trip?
22 22
Will it Trip?
23 23
Will it Trip?
24 24
Will it Trip?
25 25
Will it Trip?
26 26
Will it Trip?
27 27
Will it Trip?
Such cases not commonly encountered in

normal course of planning or operations…
Yet TPs still need clear indication on ride-

through capability of generating resources
28 28
TPL-001-4 R4.5 – Simulation of extreme events
“If the analysis concludes there is Cascading

caused by the occurrence of extreme events, an
evaluation of possible actions designed to reduce
the likelihood or mitigate the consequences of the
event(s) shall be conducted.”
29 29
What are some practical steps to
minimize the uncertainty?
ͻ Guideline on sourcing critical plant auxiliary

equipment to harden against LV events?
ͻ Guideline on plant auxiliary LV ride-through
correlated to transient stability?
ͻ Existing generation
ͻ New generation
30 30
HOW DO WE BETTER
DEFINE / PREDICT
CONVENTIONAL PLANT LV
RIDE-THROUGH WITHOUT
IMPOSING EXCESSIVE
BURDEN ON GENERATORS?
Thank you!
Power Plant Modeling
Generator Control Systems
Shawn Patterson
Bureau of Reclamation
September 20, 2016
Simplified Generator
Field Winding
Stator
N
+
DC Voltage (from exciter)
S -
Rotor
Excitation Control System
Portion included in exciter models
Manual
Control DC
Field Current 3 phase
AC voltage
Exciter
(DC
Voltage Generator
Source)
Terminal
Raise/ Automatic Voltage
Lower Voltage Feedback
Command Regulator
Reference
Adjuster
Voltage Regulators Use
Feedback Control
Difference between
Desired and Actual
Output
AVR Exciter + Generator Generator Terminal Voltage

Desired
Output Output
+ System to be
Controller
controlled
Feedback
Rotating DC Exciters - Older Systems
Non-continuously acting voltage regulator
AVR
Changes resistance
Field
Pilot
Exciter
PE Rheostat EXC
GEN
(Exciter for
Main Exciter) Main Exciter
Exciter Field Main (Generator) Field
Feedback
Rotating (DC) Exciter
Setpoint
Block Diagrams,
Transfer Functions,
B Laplace Transforms Saturation
Magnetic Amplifier Systems
Manual Control
Field
Rheostat EXC
GEN
Amplidyne
(or Mag Amp)
AVR
Induction Motor PMG Amplidyne

SCR Pilot Exciters
Static Pilot Exciter
From
Station
Service
EXC
GEN
or
Generator
Terminals
Rotating main exciter

Amplification Stage
Controller Options
At least 4 model options to choose from

Rotating AC Exciters
Rectifier may be stationary or rotating,
controlled or uncontrolled
DC
Source EXC
GEN
If controlled here, Direct control of main
slower response field here (faster)
AVR
AC Rotating Exciter with Rectifier
At least 12 model options to choose from
Static Systems
Static Exciter (SCR Bridge)
Rectifier Bridge
Supplied by
Main Generator
Terminals or
GEN
Other AC Power
Source
AVR
Limiter Input Options
PSS Input
Q
MACHINE CAPABILITY
MVAR
out Stator
Heating Limit
P
MVAR
in
Stability Control
Response
Timing
Activation Stability Control

Reactive Current Compensation
G Allows
VAR
sharing
between
G parallel
generators
Vc = Vt+jXcIt
Voltage Droop
Power System Stabilizer
Most models are entirely digital,

i.e., all model parameters
determined by settings file
Analog systems have a simpler model

Governors Use
Feedback Control
Difference between
Desired and Actual
Output
Gate/Valve
Opening
Governor Prime Mover - Generator Frequency (Power)

Desired
Output Output
+ System to be
Controller
controlled
Feedback
Mechanical Governor
Gate Motion
Controller Hydro
Mechanical Governor
Valve Motion
Steam
Electro – Hydraulic Governor
Electronics
(analog or digital)
Hydro
Hydraulics
Prime Mover
Water Flow to Mechanical Torque
Hydro
Prime Mover
Steam Pressure to Mechanical Torque
Valve position
Optional Controller(s)
Steam
Complex Control Systems
Gas Turbine
Outer Loop Control Systems
Speed Governor
Speed (System Frequency) Power increase raises

Deviation frequency
Speed Adjust
+
Governor Turbine Power
Other Setpoints Power Setpoint

Load, Temperature,
Etc., Controllers
Generic Model
For Steam or
Gas Turbines
Unresponsive Governor Model
References
• IEEE Standard 421.5-2016, IEEE Recommended
Practice for Excitation System Models for Power
System Stability Studies
• IEEE PES Technical Report PES-TR1, Dynamic Models
for Turbine-Governors in Power System Studies,
January 2013
Primary Frequency Response Initiative
NERC Power Plant Model Verification and
Testing
Troy Blalock
NERC RS Chairman
September 21, 2016
Why Primary Frequency
Response Is Important
ͻ Essential for Reliability of the Interconnections

Cornerstone for system stability
Line of defense to prevent Under Frequency Load Shedding(UFLS)
Prevent equipment damage
ͻ Essential for System Restoration
Droop response is critical in restoration efforts
Hydro units and gas turbines are some of the first units to be restarted
ͻ Compliance with NERC Standards BAL-003-1, BAL-001
Prevent future regulations related to generator frequency response
performance
ͻ To accurately predict system events (Transmission Models)
The Capability Goal
Every BES Generator should have a working governor and be set

in accordance with Frequency Response Guideline for system
reliability and system restoration and provide primary
frequency response between Pmin and Pmax.
Exemptions Nuclear, existing wind solar, or possibly some others
(environmental. etc)
LGIA and SGIA should be modified to require the same for all
future generation.
Reliability Guideline: Primary
Frequency Control
• Posted 12/15/2015
Next Step: Generator Survey
• Three Interconnections: Eastern, Western and Quebec conduct

periodic generator surveys based on selected events over
multiple years to measure generators individual performance.
Generator owners would complete and provide survey data for all
synchronized BES generators exceptions: startup or shut down
exempted.
Surveys should be sent within timely manner of selected frequency
excursions. RS representatives in coordination with NERC shall select
events with consideration of Point C or Nadir and/ or below
recommended deadband with potential the maximum number of units
on line between Pmin and Pmax.
Recommended at least two event a years spaced out in such a way to
allow generator to address any issues discovered.
Survey will be “voluntary” but could be followed up with NERC Section
800 data request for any survey responses not provided.
Troy Blalock, P.E.
South Carolina Electric & Gas Company
601 Old Taylor Rd.
Cayce, SC 29033
Mail Code J42, Cayce, SC 29033
Office #: 803-217-2040
Cell #: 803-206-8984
Fax #: 803-933-8220
Email: Jblalock@scana.com
What is the Reliability Issue
Majority of generators in BA’s fleet/ East and West

Interconnections are currently incapable of providing primary
frequency responsive and we continue to commission new
generation and the trend continues.
Problems
1) Dead Bands Settings
2) Lack of coordination with plant
DCS or other outer loop
controls
Turbine OEM Reach Out to Customer Base
2012 Frequency Response Initiative
Report
Page 37, Frequency Response Initiative Report, October 2012
What has been learned – Outer
Loop Control
Coordination with plant DCS is a requirement when

operating in MW Set Point Coordinated Control.
Improvements
No Frequency Algorithm
in DCS
3 -175 MW GE7FA Gas Mark VIe Turbine
3/3/2015
Frequency Algorithm in
Plant DCS
11 RELIABILITY
R ELIABILIITY | A
ACCOUNTABILITY
CCOUNTABILITY
Eastern Profile Changing ? New Trend
03/22/16 11:32 EDT 06/05/16 17:15 EST
1111 MW Trip 918 MW Trip

Power Plant Modeling:
Benefits for non-controls People
Brandon Bouwman
U.S. Army Corps of Engineers – Hydroelectric Design Center
September 20, 2016
Brandon’s Three Rules for Controls Settings
1. DO NOT break anything

There is absolutely no point to power plant model validation if the
generator doesn’t work
2. DO NOT piss off the operator

Operators will “re-calibrate” what is not understood or a nuisance
3. Set tuning and operate as-needed/requested

We can all agree what makes good sense for the system:
Reliable power delivery is good for the country, wallets, and environment
for everyone involved
Benefits of Power Plant Model Validation
1. Actually operate a unit over its full range – can find issues, mis-coordination, etc.
2. Observe unit behavior – find areas of rough running, resonance, maintenance,

and other issues that may be mitigated or avoided
3. Check that expectations for the plant match reality by verifying models, system
layout, capabilities, etc.
Know the limits of a computer code vs. the vagaries of real, high-definition life.
Benefits of Power Plant Model validation
Operator experiences and exposure is diminishing due to:
SCADA
Automation
“Black Box” approach
Training reductions
?
?
?
Case Study 1: Ultimate Step Test
Know what you’re
doing before you
do it.
Testing sounds like

fun, interesting work,
but misplacing a
decimal point can be
nearly (or actually)
catastrophic
Case Study 2: Sticky Wickets
As-found…
TThat looked weird…
So we did it again
TThis one looks better

Third time – that’s more like it!
Reason?
Case Study 3: Power-gate Curves
Know your variables! Block

diagrams don’t always have
everything included
Case Study 4: Operator Training
This is an 80 MVA machine

Case Study 4: Operator Training
Much better synchronizing with

proper speed matching
Case Study 5: Limiter / UEL coordination
Plots received from transmission authority show

significant ~1Hz oscillation
Event occurs at the
same
Location on the
capability curve each
time;
Interaction with limiter
1. Event occurs at the
same location on the
capability curve each
time
2. Events involve
interaction with limiter
3. Out of 14 identical
units with the exact
same digital exciter
settings, only 2 exhibit
this behavior
4. Confirmed not
SCADA
Reduced limiter gain

solved the issue, but no
root cause found.
Conclusion:
The point of all of these examples is to understand both the
model of your controls and the
reality of your plant.
1. Models, controls systems, and machinery are not separate.
2. They’re also not infallible.
3. Validation is not a burden, but an opportunity to deeply check our systems as a whole.
ERCOT Experience on Power Plant Model
Validation
José Conto
NERC Technical Workshop:

Power Plant Model Verification and Testing
Sept 20-21, 2016
Agenda
• ERCOT
• System Studies
• Model Validation
2
The ERCOT Region
At a glance
• About 90% of Texas load - 24 million consumers
• Competitive-choice customers: 75% of load
• More than 7 million electric-service ID’s
(premises)
• More than 46,500 circuit miles of high-voltage
transmission
• More than 550 generating units
• More than 77,000 megawatts (MW) of expected
available summer peak capacity
• Record peak demand: 71,197 MW (Aug. 11, 2016)
PUBLIC 3
The ERCOT Grid
Transmission Investment and Development

• In 2015, $1.1 billion in transmission added
• 1,093 circuit miles of transmission improvements completed
• 1,514 circuit miles of transmission planned
• $5.6 billion under development over the next five years
Generation Development
• Over 54 GW new generation added since 1999, 148 units decommissioned
• 16,189 MW generation committed for the future
• Over 61 GW of active generation requests under review,
o Includes more than 24 GW of wind (January 2016)
o Solar in queue: 1,417 MW in 2016, 1,765 MW in 2017
Solar and Wind Generation

• 288 MW of installed solar capacity
• Nearly16 GW of installed wind capacity
• Wind generation record: 14,023 MW (Feb. 18, 2016)
• Wind penetration record: 48.28 percent (March 23, 2016)
PUBLIC 4
Planning Studies
• Transmission Planning Criteria

– Full interconnection model (+ 69 kV network)
– Voltage stability margins, transient voltage

response, damping criteria
– UVLS, UFLS and Load Resources (market

program where load trip at 59.7 Hz)
– Combined Cycle unit governor model
PUBLIC 5
Studies outside the box
• LVRT and HVRT studies on new interconnection
• Small Signal Stability study
• Panhandle Wind Power expansion
• SSR studies for all new interconnections.

Screening studies by ERCOT, full studies by TP’s
PUBLIC 6
PMUs in ERCOT
• Total 98 PMUs @ 42
locations (Dec. 2015)
Future PMUs
- New generating Facilities over 20 MVA
- FACTS devices
- Transmission Facility (associated constraint)
PUBLIC 7
Generator Dynamic Model Validation
• PMU’s provide high accuracy measurements with unit online for

comparison against simulation output
• Iterative process will produce a model that matches the actual dynamic
behavior of the equipment in the field
• Automation of the process and use of parallel-computing techniques

greatly eases the workload associated with matching simulation
models with measured response
• PMU’s can be used to satisfy generator model validation requirements

set in NERC standards MOD-026-1 and MOD-027-1
PUBLIC 8
Single machine outage – Frequency Event
138 kV bus
Red: PMU Blue: Simulation
After droop adjustment
Red: PMU Black: Simulation
PUBLIC 9
PUBLIC
1
0.96
0.98
1.02
1.04
1.06
1.08
1.1
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
2.20
2.40
2.60
2.80
3.00
3.20
3.40
3.60
3.80
4.00
4.20
4.40
PMU_V
4.60
4.80
5.00
Wind Generator Model Validation
Simulation_Default
10
Model Translation – Wind Turbine PSSe UDM to TSAT
Generator Terminal Voltage (93018) Generator Active Power (93018) Generator Reactive Power (93018)
Q (MVAr)
P (MW)
V (pu)
time (s) time (s) time (s)
TSAT TSAT TSAT
Bus Frequency (154)

f (Hz)
time (s)
TSAT
PUBLIC
Thanks!
José Conto
Principal, Dynamic Studies
ERCOT System Planning
(512) 248-3141
Jose.Conto@ercot.com
12
Excitation system
Essential to provide magentizing current to

generator field winding
Typically power rating is about 1/4 - 1/3 percent of

generator MVA rating
Must be capable of
today - fast control of megawatt-level real power
readily practical with electronics
legacy - fast control of multi-kilowatt level real power
prior to about 1945 - had to be electromechanical
1945 - about 1975 - had to be electromagnetic
Excitation system
Control genrator terminal voltage
Manages sharing of reactive power load among

generators on a common bus or three winding transformer
Limits high field current of prevent excessive heating of generator field

winding (OEL)
Limits low field current to avoid undesirable effects of underexcitation (UEL)

Static excitation system
Controlled rectifier Hundreds of volts - thousands of amps

Excitation power transformer 1 - several MVA
Inner control loop
Voltage regulator Digital - milliwatts - watts
Brushless rotating exciter system
Uncontrolled rectifier Hundreds of volts - thousands of amps

Brushless exciter Many kilowatts - few megawatts
Voltage regulator Digital - milliwatts - watts
Real Power Dynamics
Real Power Dynamics - Gas turbine plant
Gas turbine plant
P-I-D governor Gas turbine engine

Acceleration limit controller Temperature limit controller
Load controller
Combined cycle plant
Combined cycle plant
Real power dynamics - Hydro plant
Beware of obselete models
There has been extensive new model development
Reflects current digital ‘front end’ practice

Improves hydraulic servo modeling
Improves Francis/Kaplan turbine modeling
Hydro plant
Baseline Testing
Excitation Systems
Shawn Patterson
Bureau of Reclamation
September 21, 2016
Model Development
• base \’bas\ n. 1 a: the bottom support of something; that on which a
thing stands or rests b: a fundamental principle or groundwork;
foundation; basis
• vt. 1: to establish, as a fact or conclusion
Identify individual model components

Not a black box fit!
Every parameter must be derived on a technical basis
Measured or calculated
Then measure the overall output response to an input
Compare the complete model response to measurement

General Principles
• Control theory basics (Diff. Eqs., s-domain, transfer functions)
• The more data, the better the model
• Break the models into as many parts as possible
• Test signals and measurement devices must have higher bandwidth
than the equipment/models
• Understand the limitations (signal accuracy, model fidelity, system
changes, noise, hysteresis, bandwidth…)
• There are no exclusively right or necessary ways to do things
• but there are many ways to get things wrong!
Transfer Function
Input System Output

U(s) G(s) Y(s)
G(s) = Y(s)/U(s)
To identify a system by comparing outputs,

model inputs must be identical to test inputs
The excitation control system is a closed loop
In the middle of this model is the generator

So first get this right
Open Circuit Saturation Characteristic
- Base Field Current

- Generator Gain
Then this
Saturation Characteristic under load
Steady State Model

This must be right
Generator Transient Model
T’do
T’do can be identified anytime a changing

Vt is measured offline with a known
Efd, e.g., VAR rejection, exciter shutdown,
step in Efd (manual control mode), etc.
Exciters - get rotating exciters first
Calculate
Measure or get
From manufacturer
Control Loops
Break the problem down
• Isolate components if possible
• Open loop tests if possible
• Offline tests remove power system influence
• Sometimes possible to measure components while unit is shut down
• Test AVR first without PSS, Compensation
• Testing at reduced load removes effect of local mode oscillation
• In general, linear elements should be measured, nonlinear elements
are calculated
Model Parameters
• Linear parameters
• gains, time constants
• Small signal tests/analysis
• Step and frequency response tests
• Nonlinear Parameters
• limits, gain multipliers
• Large signal tests, or calculate
Calculated and measured Calculated
Measured
If analog: measure (or calculate and measure ) Adjust parameters as necessary

If digital: transfer time constants directly to model based on complete model response
Small signal tests
• Step response
• Frequency Response
• Input into Vref

• Input scaling typically 1 volt / percent terminal voltage
• +/- 1 volt usually small enough to maintain linearity, but get good
signal/noise
• Input signals must be frequency rich

• Measurements must be sampled at greater than 10 times the model
bandwidth
Frequency Response
• Input sine signals into Vref summing junction into voltage regulator
• Step through many frequencies over the response range of the
system
• 0.05 Hz to 10 Hz
• Measure Vt for output signal
• Bode plot for response of Vt/Vref
• Entire closed loop response of exciter/generator
Transfer Function
Input System Output

u = U sin(wt) G(jw) y = Y sin(wt+f)
Frequency Response of System is G(jw)
Gain at w = |G(jw)| = Y/U, Phase at w = f

Step Response
A step input into the

AVR summing junction
provides a frequency rich
signal
Measure and compare all signals!

(Efd, Ifd, MW, MVAR)
Small Signal Responses
• Ensure response is essentially linear
Offline
Note if signals
+/- 1 % steps are limited
Into Vref
Efd should be fairly symmetrical,

and have no flat spots
Frequency Response
Some models
are better than others
at frequencies higher
3 Hz
IEEE standard models

only “guaranteed” up
to 3 Hz
Offline response eliminates system (and other) influences

on the results
Strictly speaking, this model represents an integral-of-accelerating-power
stabilizer (dual input), which is always digital, but is often a generic model for analog
circuits as well
An analog PSS consists of only circled portions, and should be tested/validated

in its entirety, each portion separately
Digital systems can be tested to various extents based on design
Why test a 100% digital component?

Depends on your level of trust with the equipment
Trust but verify!

PSS Frequency Response
40
Measured
20 Model Frequency response of PSS
Gain (dB)
can be modeled very well

0
-20
-2 -1 0 1 2
10 10 10 10 10
Freq (Hz)
100
Phase (deg)
-100
-200
-2 -1 0 1 2
10 10 10 10 10
Freq (Hz)
Local mode oscillation

is difficult to match
Step response with PSS disabled to compare damping when enabled

Local mode oscillation

is the only measurement
available to verify PSS
effectiveness
Step response with PSS to compare damping when enabled

Reactive Current Compensation
Step Response with

compensation enabled
and disabled to verify
effective Xcomp
Challenge limit with a step input.

Note levels, timing, reset value, and
response characteristics. Perform test
in simulation
Timing
Pickup
Reset
Alternatively, this model can

be populated by careful analysis
of the model and excitation
system manual
Test the model thoroughly

Same approaches as
overexcitation limiter.
This can be performed

during MOD-025
testing
References
IEEE Standard 421.5-2016, IEEE Recommended Practice for
Excitation System Models for Power System Stability Studies
IEEE Task Force on Generator Model Validation Testing of the Power

System Stability Subcommittee, “Guidelines for Generator Stability
Model Validation Testing,” IEEE PES General Meeting 2007,
paper 07GM1307
Power Plant Model Validation:
BPA Experience
NERC Workshop
September, 2016
Steve Yang
Bonneville Power Administration
1
BPA Overview
• Bonneville Power
Administration (BPA) is a
federal Power Marketing
Agency in Pacific Northwest
• BPA markets power from 31
Federal dams and the
Columbia Generating Station
Nuclear Plant
• BPA operates more than
15,000 miles of
transmission, including
4,735 miles of 500-kV lines
BPA operates several large paths in the Western Interconnection – California

Oregon AC Intertie (4,800 MW), Pacific HVDC Intertie (3,100 MW), Northern
Intertie (3,100 MW), and Montana Intertie (2,200 MW)
2
Outline
• Dynamic Disturbance Recorders for

PPMV at BPA
• PPMV tools and process with a case
study and examples
• When models and reality don’t agree
• Summary
3
Dynamic Disturbance Recorders (DDRs)
for Model Validation at BPA
• Phasor Measurement Unit (PMU)
– Output voltage and current phasors, calculated frequency,
active and reactive power
– Data can be archived locally and streamed to a control center
in real time
– Minimum filtering is preferred for model validation
• Point on Wave Recorders

– Technologies:
• DFRs with continuous recording capabilities (BEN6000)
• Portable Power System Monitor (PPSM)
– Output: point on wave voltages and currents
– Data is archived locally, retrieved via FIN network
– Phasor calculations are done off-line
– Preferred technology for electronically connected equipment
Modern DDRs can combine Phasor and Point on Wave recording

4
DDRs at Point-of-Interconnection (POI)
• BPA has installed DDRs at power plant POIs since 1996
• BPA developed Power Plant Model Validation (PPMV) application using
DDR data and GE PSLF play-in function (added in 2001)
• BPA requires PMU installation for all new generation including the wind
Substation Power Plant
Point of
Interconnection
V I
G
G
Record:
- POI bus voltage
- POI bus frequency
- Power plant MWs and MVARs
DDR needs to be placed at Power Plant POI 5

BPA’s DDR Coverage at POI
• Conventional –
• 14 plants,
• 133 generators,
• 21,345 MW of generation
• Wind –
• 12 plants
• 1,200 MW of generation
• More to be added
6
BPA PPMV Tools
• BPA PPMV
• Sequence of GE PSLF EPCLs and MATLAB programs
• BPA-PNNL PPMV
• Stand-alone data management program and
automated PSLF interfaces
• EPRI PPPD
• Stand-alone MATLAB based software
7
PPMV Process: Initial Set-up
Extract power plant data from WECC
basecase and build a sub-system
Enter plant information
8
PPMV Process: Event Notifications
• BPA frequency event detection program

• BPA event detector
• WECC JSIS event alerts
• DFR triggers
• StreamReader trending displays
9
PPMV Process: Collect and Review Data
10
PPMV Process: Run Validation
11
Initial Results
12
When models and reality don’t agree
• Model validation studies need to be done over several events
• It is very rare to get a good match for all events because

operating mode can be different (e.g. generator is ramping,
generator on AGC responding to a contingency)
• Frequent reasons for systemic model mismatch include:

• Wrong powerflow data
• PSS status / gain
• Inertia time constant
13
Transformer Name Plate and Tap Setting
Check transformer impedances and taps
14
New Transformer Data
15
PPMV Process: Review Results
16
PPMV Process: Generate Report
17
WECC Re-Certification of Centralia
Power Plant
18
Calibration
• EPRI Power Plant Parameter Derivation (PPPD) is most

mature, a user group is established including 23
participants
• Bernie Lesieutre @ University of Wisconsin uses a unique
approach of pattern matching – which is useful to
provide insight in model inaccuracies
• Others:
• MATLAB
• University of Texas – Particle Swarm Optimization
• PNNL – Kalman filter
• Georgia Tech – super-calibrator
• Idaho Power developed in-house optimizers
19
Calibration
Before calibration After calibration
• PPMV can complement model development and calibration, there are successful
case studies. However, engineering expertise and knowledge of generator
controls are essential
• Beware of curve fitting exercises 20
If All Fails Repeat Baseline Test
21
After Re-test
MOD-027
MOD-026
22
Going Beyond the Standards
23
Performance Monitoring and
Detection of Control Failures
• PMU monitoring provides detection of generator abnormalities
Power
700
680
Unexpected action
660
640
PSS failure from plant controller
Active Power [MW]
620
600
Power
580 550
560
540
540 Observed
Expected 530
520
Active Power [MW]

500 520
2 3 4 5 6 7 8 9 10
Time (sec)
510
50
500
Observed
490 Expected
Reactive Power (MVAR)
0 480
0 10 20 30 40 50 60 70 80 90 100
Time (sec)
-50 Abnormal runback in

reactive power
-100 24
0 5 10 15 20 25 30 35 40 45 50
Time (sec)
Summary
• PPMV using Dynamic Disturbance Recorders
– Provide a cost-effective approach for periodic model verification as
required by MOD-026,MOD-027 and WECC policy
– Give Transmission Planners independent way to validate power plant
dynamic performance and models
– Can be used in real-time performance monitoring
• Several technology solutions are available
• Model validation tools are mature and well-developed

– A number of success stories
– Disturbance model validation is a part of regular business process at
BPA
25
Thank You!
Contact information:
Steve Yang (hyang@bpa.gov)
26
GE Perspective
Power plant modelling & verification
e
Observations
Observations Global trends
• Plant operating in load control with no • Loss of inertia from renewable growth
frequency bias
• Reduction in primary response due to lag in • Faster + higher primary response
plant level load control loop or requirements
communication
• Assumption that AGC target contains • Reserve margin capability/enforcement
frequency bias necessary
• Co-ordination AGC – Plant control –Turbine • Droop setting reductions ( ie IESO 3 %
governor requirement for GT in CCGT)
• Frequency signal resolution limitations in
plant controller • Dead band reductions or modification
• Large steam governor dead band vs boiler
stability vs hydraulic system capability • Proliferation of different requirements
Missteps Common questions

• Disabling plant load control to sustain • Plant upgrades : where to set 4% droop …
primary response ISO, prevailing ambient, max rated capability,
handling of peaking features
• Applying primary response in two zones
without co-ordination • Role of ACE vs plant response
e
MOD-025
Generator Real and Reactive Power Capability
x Many units may not have operated with as many (+) or (-) VARs in their lifetimes,
so they could be pushed to their limits
x Prior to testing: Make sure all excitation system components are in good operating condition to
handle the field current and var variations
¾ Brushless exciter: Inspect diodes, fuses, armature, connections, etc
¾ Static exciter: Inspect thyristors, fuses, collector, brushes
¾ Confirm proper coordination between AVR under-excitation limiter (UEL) and loss-of-field
(40) protection (already required per NERC PRC-019)
x Consider having AVR expert on-site during testing in case any unexpected alarms or limits occur
x Testing for over-excited (lagging) reactive power:
¾ VAR changes only on tested unit will often result in reaching generator or switchyard
over-voltage limits
¾ By absorbing VARs on adjacent and neighboring generators, switchyard voltage can be
held down, allowing tested generator to achieve more lagging reactive power
x Testing for under-excited (leading) reactive power:
¾ VAR changes only on tested unit will often result in reaching generator or switchyard
under-voltage limits or “shrunken” UEL limit based on reduced voltage
¾ By increasing VARs on adjacent or neighboring generators, switchyard voltage can be
propped up, allowing tested generator to achieve more leading reactive power
Page 1 NERC PPMVT Workshop – Manufacturer Panel Confidential © Siemens Energy, Inc. 2016. All rights reserved. September 2016
MOD-026
Generator Excitation Control System Model Verification
x Model comparison with test results is more productive when excitation system
model reflects:
¾ Actual AVR settings, rather than “typical” settings
¾ Updated IEEE model, when appropriate.
Example: IEEE Type ST4B and ST6B models for static excitation systems with
P-I control became available in 2005 in IEEE 421.5-2005 standard
x Testing typically includes AVR signal injection into AVR voltage summing junction via:
¾ Test points on analog AVRs
¾ Software signals on most modern digital AVRs
x Types of excitation systems which can be tested, modeled, simulated and validated
by Siemens:
¾ All Siemens excitation systems including RG3, Thyripol, Thyrisiem
¾ All legacy Westinghouse excitation systems, including WTA, WTA-300, WDR,
MGR, PRX, Magastat
¾ Selected third party excitation systems (Cutler-Hammer, Basler, Semipol, etc.)
¾ Other systems with coordination between plant owner and AVR OEM.
MOD-026
Generator Excitation Control System Model Verification
Model validation of Siemens Thyripol excitation system

Terminal using IEEE Type ST6B
Voltage
Field
Voltage
MOD-027
Turbine/Governor Model Verification
• Settings affecting turbine-governor models and model validation
in modern turbine control systems:
¾ Proportional gain (affects droop)
¾ Integral gain (affected by reset or repeat time constant)
¾ Dead band (set in terms of rpm)
• Signal injection into frequency error summing junction via software in modern
digital control systems
• Types of turbine control systems which can tested, modeled, simulated and
validated by Siemens :
¾ All Siemens controls systems, including
SPPA-T3000, SPPA-T2000, TXP, Simatic S5 and S7 systems
¾ All legacy Westinghouse control systems including AEH*, DEH*, WDPF
* Cannot perform with signal injection
¾ All Siemens turbines with non-OEM controls, including
Allis Chalmers, Westinghouse
¾ Selected third party control systems
MOD-027
Turbine/Governor Model Verification
SGT603 (rev0) Model

for Siemens
SGT6-2000E(8)
Gas Turbine
Establishing the Baseline Model
John Undrill
NERC Workshop
September 2016
You should not assume that the equipment in the plant matches
the data used in initial setup of the model
It is important that the equipment, and at-site documentation, be

seen by people who know what to look for
The I&C engineer at the plant should/may know the transducers,

wiring, and terminations
but
He may never have seen a transfer fucntion diagram of his

equipment and has never run a dynamic simulation of a power
system
Operating practices are important
Voltage regulator should be in auto-voltage mode

PSS should be ON
Turbine may be in
Droop mode, responsive to frequency without

overriding action by load controller
Droop mode, responsive to frequency but

with overriding action by load controller
Load controller may have/not have frequency bias

Practical points
Plots of results must be displayed quickly after the

test is done
Must respect the white knuckle limits of the plant

personned
Offsets in plots are a fact of life - ambient contitions

change - input output curves should match in
principle - fit in great detail is likely to be misleading
Drawings may be old, wrong, completely lost
Beware of instruments that have recent calibration

stickers on them but, neverthess, but reading wrong
by large factors
Manufacturer’s field personnel will be familiar with their equipment
and able to plug in their ‘engineering’ terminals
Then they can (may be able to)

- look up parameter values
- read operating values (eg exciter field amps)
- show some transfer function information
- apply well defined test disturbances
- record dynamics tests with good bandwidth
They will/may know details of their equipment that are site specific,
undocumented, disabled, nonstandard additions
There is no value in an dedicated (or automated) effort to match

a transfer function model to test results if the ‘as operating’
configuration of the plant does not match the transfer function model
Risks are small but must be understood
Old equipment that is working correctly can break if disturbed

- brittle terminal blocks, dodgy switches
- wait for replacement parts - delay in return to service
New equipment that has worked correctly in one operating

condition may fail to work, or work incorrectly, in different
conditions required by test
- delay in return to service
Protection may be set tighter than it is thought to be
Unexpected trips do not inspire confidence
Costs of delays in return to service can far exceed the direct

costs and fees of testing/modeling
Cautions and precautions
The generator capability curve does not define the prudent

range of testing
Auxiliary bus conditions are as important as conditions at

the generator terminals and at the high voltage bus
Protection should not be altered - neither settings not status
Test plan should respect both protection and alarm limits

Basics of Power System Control
Real Power
NERC Workshop
September 2016
John Undrill
1
Basics of Power Systems Dynamics
dω
Newtonian dynamics 2H = Tengine − Telec
dt
dω Pengine − Pelec
2H =
dt ω
Engine power and Pelec = Penom + Delec ω

load can vary with frequency
Pengine = Psnom + Dengine ω
dω Psnom − Penom − (Dengine − Delec )ω

Basic equation of motion 2H =
dt ω
Basics of Power Systems Dynamics
Basic equation of motion
No governor, constant ΔPlnom

Δω =
engine power (Dengine + Dload )
Proportional governor ΔPenom = Kp Δω
ΔPlnom
Governor gain = Kp Δω =
Kp + Dengine + Dload
ΔPlnom ΔPlnom
= =
Kp + D D + R1
This describes the steady state frequency offset but does not describe
the dynamic behavior of the system
Governing Droop
The governor
- does not control speed
- does not control power
- does control the relationship between

speed and power
(special cases)
off line - power is zero

- controls speed
tied to infinite bus - speed is fixed

- controls power
Turbine Load Control
Adjust frequency by Adjust power allocation by

collective adjustment of differential adjustment of
speed-load references speed-load references
Governing Droop - Proportional Governor
Want this Build this
n = nref − R Pengine
Kp (nref − n) = Pengine
R = 1/Kp
n = nref − R Pengine
Note that we have used valve stroke as a surrogate for engine power
Proportional governor
This is oversimplified -
This is more realistic -
Would always be stable if it
Four 90deg phase shifts
could be built -
Small steam turbine / small gas turbine
Could not be built
5 percent load increase on isolated engine
0.0 Divx 10 Divy 10 Time 25.0

Name Min Max Col CpPu F Fa Fb Fc Fd
0 s0 0.5 0.7 1 1.0 1 0.0 0.0 0.0 0.0
1 s1 0.5 0.7 2 1.0 1 0.0 0.0 0.0 0.0
2
3
s2
s3
0.995
0.995
1.005 3
1.005 ____
4
1.0
1.0
1
1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
power
Kp = 20 speed Kp - 25
Droop = 0.05 Red Tv = 0.05

Green Tv = 0.15
Red valve Blue Tv = 0.25
Green power Cyan Tv = 0.35
Blue speed 1
2
s1
s2
0.5
0.995
0.7 ____
1.005
1
1
1.0
1.0
1
1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Kp = 33, 25, 20
Red R = 0.03
Green R = 0.04
Blue R = 0.05

2 s2 0.995 1.005 1 1.0 1 0.0 0.0 0.0 0.0
Proportional - Integral Governor
At node 0 error = nref − n − Rv

When error is zero n = nref − Rv
Could achieve droop by using feedback of valve stroke or electrical power
(Valve stroke is a surrogate for engine power)

Control of Load Sharing
P-I Governor - Choice of Droop Value
R = 0.05 R = 0.005
Operator
adjusts
speed-load
references
to adjust
allocation of
power output
red
0.0 Divx 10 Divy 10 Time 100.0 0.0 Divx 10 Divy 10 Time 100.0
Name Min Max Col CpPu F Fa Fb Fc Fd Name Min Max Col CpPu F Fa Fb Fc Fd
1 s1 0.0 1.0 2 1.0 1 0.0 0.0 0.0 0.0 1 s1 0.0 1.0 2 1.0 1 0.0 0.0 0.0 0.0
2 s2 0.98 1.02 1 1.0 1 0.0 0.0 0.0 0.0 2 s2 0.98 1.02 1 1.0 1 0.0 0.0 0.0 0.0
6 s6 0.0 1.0 3 1.0 1 0.0 0.0 0.0 0.0 6 s6 0.0 1.0 3 1.0 1 0.0 0.0 0.0 0.0
0.0 Divx 10 Divy 10 Time 250.0 0.0 Divx 10 Divy 10 Time 250.0
Name Min Max Col CpPu F Fa Fb Fc Fd Name Min Max Col CpPu F Fa Fb Fc Fd
1 s1 0.0 1.0 2 1.0 1 0.0 0.0 0.0 0.0 1 s1 0.0 1.0 2 1.0 1 0.0 0.0 0.0 0.0
2 s2 0.98 1.02 1 1.0 1 0.0 0.0 0.0 0.0 2 s2 0.98 1.02 1 1.0 1 0.0 0.0 0.0 0.0
6 s6 0.0 1.0 3 1.0 1 0.0 0.0 0.0 0.0 6 s6 0.0 1.0 3 1.0 1 0.0 0.0 0.0 0.0
R enewable Energy Modeling
NERC PPMVT Workshop
Matt Richwine
September 20, 2016
W ind Plant Model Network
Equivalent
Station Turbine
System Transformer(s) MV Collector Transformer Equivalent
Equivalent Equivalent WTG
System
R2, X2 R1, X1, B1

Rt2, Xt2 Rt1, Xt1
HS SS LS SS HS WTG
WTG Bus
Xfmr Bus Xfmr Bus Xfmr Bus
Bus typically used for voltage regulation
• Recommended network for representing wind farms in positive sequence RMS models for
transmission planning
• Aggregation of the collector and turbines has been established for many years in
industry[1]
2
NERC Model Workshop
9/20/2016
Generic Renewables Models
Generic RMS Models
Renewables Models, 1st Generation
Renewables Models, 2nd Generation (WECC)
Renewables Models, IEC61440-27 (ENTSO-E)
Exciter-like
functions
Generator-like
functions
• Not all modules are
required for all wind
turbine types
• Some modules are also
used for solar plant
representation Governor-like
functions
2nd Generation Renewables Model (developed by WECC) 3

NERC Model Workshop
9/20/2016
T ypes of Renewables Models
Very Detailed Less Detailed
Typical Applications
• Equipment Interactions • Interconnection • Bulk Transmission
• Very Weak Systems Studies Planning Studies
• Special Purpose Studies • Special Transmission
Planning Studies
Selection of Model Depends on Application

4
NERC Model Workshop
9/20/2016
V oltage Reference Step Test
• Wind plant is in voltage regulation
mode at the 345kV bus
• +1% voltage reference step at the POI
• Second Generation Generic Models used
for validation (WECC Model)
• Response without voltage droop
• In satisfaction of NERC MOD-026-1
Verification of Models and Data for Plant
Volt/Var Control Functions
Vref/Vreg
or Qref/Qgen Plant Level
Control Vterm
fref/freq and Pgen
Plant_pref/Pgen
Qgen
Iqcmd
Qref Iq
Electrical Ipcmd Generator/
Pgen Ip
Torque Pref Control Converter
Prefo Control
Pord
Z
Zref Pgen
Drive-Train
Pm
Pitch T
Aero
Control
5
NERC Model Workshop
9/20/2016
P lant Interactions V1reg
G1 X1 Xg Vinf
X2
Challenge:
V2reg
Plant volt/VAr (MOD-026) model validation can be
confounded by other local voltage-regulating devices G2
(wind farms, solar plants, STATCOMs, SVCs, etc). Factors
include:
• Weak grids (Xg >> X1, X2): Interaction likely
• Non-dominant plant (X1 >> X2): G1 hard to isolate
• Non-linear response (ie. deadbands): Non-linear
plant must be fully modeled or disabled
Occurrence:
• Wind plants are often located on weak grids
• Wind plants are often clustered where wind is good
• SVCs/STATCOMs are often used in weak grids for
voltage support
• Deadbands are common in volt/var controls
Critical to Understand the Local System

6
NERC Model Workshop
9/20/2016
P lant Interaction in Test Data
Linear System – Good Test Data
Good Linearity – Slope

indicative of Grid Strength
Noisy Region
Non-Linearity – Unusable Test Data

Control
Shunt Interaction
Switching
Noisy Region
7
NERC Model Workshop
9/20/2016
P MU Measurements Vref +
+
Verr PI Iq X
Vbus
Volt/VAr Challenge: Vmeas
Imposing a voltage (from a PMU recording) at V1reg to

validate G1’s model open the voltage regulation control PMU
loop
V1reg
G1 X1 Xg Vinf
X2
PMU V2reg
G2
Mitigation Option:
Imposing a voltage (from a PMU recording) farther up-
stream can leave the feedback loop closed and
improve model validation
However, data recorded farther up-stream often
includes more interacting equipment – becomes
difficult to isolate one device
V1reg PMU
G1 X1 Xg Vinf
Test data courtesy of BPA
X2
V2reg
8
G2 NERC Model Workshop
9/20/2016
P MU Measurements
Important to check when using PMU
data for Volt/VAr model validation:
1. No non-linearity in the local system
(close enough to interact significantly
with the plant under test)
2. No in-feeds from other equipment
3. Sufficient impedance between the

point-of-regulation and the PMU
point-of-measurement (or a way to
deal with errors from open-loop
control)
9
NERC Model Workshop
9/20/2016
Questions?
T hank You!
Matt Richwine
GE Energy Consulting
Matthew.Richwine@ge.com
518-385-4407
11
NERC Model Workshop
9/20/2016
References
[1] E. Muljadi, C.P. Butterfield, A. Ellis, J. Mechenbier, J. Hocheimer, R. Young, N. Miller, R. Delmerico, R. Zavadil, J.C. Smith, ”Equivalencing the
Collector System of a Large Wind Power Plant”, presented at the IEEE Power Engineering Society, Annual Conference, Montreal, Quebec,
June 12-16, 2006.
[2] Richwine, M.P.; Sanchez-Gasca, J.J.; Miller, N.W., "Validation of a second generation Type 3 generic wind model," in PES General Meeting |
Conference & Exposition, 2014 IEEE, July 2014
12
NERC Model Workshop
9/20/2016
Capacitor Bank Switching Test
• Shunt capacitor switching test (10MVAr rated

bank at HV bus)
• Voltage regulation mode
• Wind plant responds by reducing reactive
power when the bank is switched in,
increasing when switched out
• Power (wind speed) decreasing
• 3 models validated[2]:
• GE models
• Generic – 1st generation
• Generic – 2nd generation
13
NERC Model Workshop
9/20/2016
A ctive Power Response – GE Model
• Frequency step injection test – release of an over-frequency step (or fast
ramp) test
• Plant responds according to its frequency droop curve (example shown)
• Wind speed is not fed into the model
• Close match of active power achieved
Example Droop Curve

14
NERC Model Workshop
9/20/2016
A ctive Power Response – Generic
• Second Generation Generic Model –
active power response is at the plant
controller module (REPC_A)
• Frequency step injection of 200mHz
• Frequency response settings are 5%
droop, 36mHz deadband
• Measured wind speed is not fed in to
the model
• Power declines quickly but recovers
more slowly when released
• Power is “wind following” outside the
event…larger and slower variations
• Power is more regulated during the
event… smaller deviations (as
expected)
15
NERC Model Workshop
9/20/2016
© Copyright 2013, First Solar, Inc.
Introduction
• Test results from NREL-DOE test on FS PV plant
• 3 Tools/Platforms
a) PNNL-BPA Power Plant Model Validation (PPMV) Tool
b) GE PSLF Tuning (Manual)
c) EPRI Renewable Energy Model Validation (REMV) Tool
• Conducted 3 tests
a) Under-Frequency Test
b) Capacitor Switching Test

c) Reactive Power Control Test
• PMU Data recorded at 30 Hz
Figure Courtesy - Dmitriy Kosterev, Steve Yang, Pavel Etingov, “PMU-based application for power plant model validation (PPMV)”, NASPI Working Group Meeting, March 23-24, 2015
Event 1 –Under-Frequency Test
• Frequency time series from actual event in ERCOT measured on November 29, 2011 was used
• Centralized Power Plant Controller (PPC) emulated this frequency response
• Plant operated in curtailed mode with 50% of plant capacity left as reserve
• Very aggressive droop setting of 1.67% was implement during test

3
Under-Frequency Test Results
PPMV Tool GE PSLF
REMV Tool

4
Event 2 –Capacitor Switching Test
• “Staged test” data used for this model validation task
• Capacitor bank located at medium voltage collector bus engaged during test
• 5 MVAR reactive power target set in SCADA HMI
• Inverters could not meet reactive power target within its power factor capability range
• PPC engaged 4.75 MVAR capacitor bank automatically to meet the target
• Capacitor switching dynamics adequately captured in simulation

5
Capacitor Switching Test Results
PPMV Tool GE PSLF

REMV Tool
6
Event 3 –Reactive Power Control Test
• “Staged test” data used for this model validation task
• Centralized Power Plant Controller (PPC) was set in VAR control mode
• During test, set point for reactive power was changed from 0 MVAR to roughly 1.6 MVAR and
back to 0 MVAR
• PPC issued reactive current command to achieve desired reactive power at POI.

7
Reactive Power Control Test Results
PPMV Tool GE PSLF
REMV Tool

8
Conclusions & Future Work
• PV plant model validation using staged test data and natural event data is presented
• Existing PV plant models as approved by WECC REMTF/MVWG group is capturing plant dynamics
adequately under natural events and under staged tests conditions
• System operators could provide actual event data to help verify model adequacy and validate model
performance when subjected to real-time grid disturbances
• Similar model validation exercise for another PV facility rated above 250 MW is under progress,
results will be presented when validation is complete.
• More tests like over-frequency event will also be tested in future

9
Questions ?
Sachin Soni
Sachin.soni@firstsolar.com
623-255-9981

10
Sept 20, 2016
NERC PPMVT Workshop

Generator Governor Frequency Response
2004 NERC Operating Policy 1, Generation Control and
Performance, Section C
“Governor limits – Turbine control systems that provide

adjustable limits to governor valve movement (valve
position limit or equivalent) should not restrict travel more
than necessary to coordinate boiler and turbine response
characteristics.”
Auxiliary equipment capacity limits.

Process deviation limits
Operator define unit load limits.
© ABB Group
September 30, 2016 | Slide 2
Generator Governor Frequency Response
Primary Frequency Control Consideration
© ABB Group
September 30, 2016 | Slide 3
Basics of Power System Control
Reactive Power
NERC Workshop
September 2016
John Undrill
1
Control and sharing of reactive power load requires droop
Collective adjustment of voltage references changes voltage
Differential adjustment of voltage references changes distribution of reactive power

Basics voltage regulator
Ef d = K(Vref − vt )
(Ef d − vt )
Q=
Xs
K Xs
vt = vref − Q
K +1 K +1
Ef d = K(Vref − (vt − Xc Q))
(Ef d − vt )
Q=
Xs
K Xs − KXc
vt = vref − Q
K +1 K +1
Ef d = K(Vref − (vt − Xc Q))
(Ef d − vt )
Q=
Xs
K Xs − KXc
vt = vref − Q
K +1 K +1
Make K >> 1
Then K/(K+1) ~= 1
Then
vt = vref + Xc Q
Rising series current compensation
Could we -
Choose X_c = X_t
vt = vref + Xc Q then the AVR controls

high side bus voltage
vbus = vref − Xt Q
Bad idea !!!!
vbus = vref − (Xt − Xc )Q

MVAR adjustment - Raise Qa - Lower Qb
No compensation Drooping compensation Xc=-.11

Good response Quicker response
1 sec time constant Less sensitive
Rising compensation Xc=-.11
Slow response
Excessively sensitive sensitive
Rising compensation Xc=-.13

!!!!
Emerson Process Management
Power & Water Solutions
Steam Turbine Controls Modeling
Unit Control Modes – Coordinated Controls
• Coordinated Mode
– Developed in late ‘70s and early ‘80s to
answer a long-standing controls
problem
– Boiler and steam turbine want to fight
each other (Throttle Pressure vs. MWs)
and didn’t “coordinate” with each other
– Both Boiler Master and Turbine Master
receive a common “unit” setpoint which
limits disparities between steam
generation and steam use (flow)
– Tuning and configuration of these
closed-loop controls determines how
responsive the unit will be during a
system frequency event (boiler slow,
turbine fast)
Unit Control Modes – Turbine Following & Sliding Pressure
• Turbine Following Mode
– Steam turbine valves respond to
Throttle Pressure changes while the
boiler firing rate determines MWs
– Not responsive to frequency events
(boiler too slow)
• Sliding Pressure Mode
– Steam turbine valves are either wide
open or locked in a position
– Minimizes throttling enthalpy losses
– Most efficient mode
– Boiler firing controls both Throttle
Pressure and MWs
– Not responsive to frequency
Turbine Protection Circuits Affecting Frequency Response
• Speed Regulation
– Pinches modulating steam valves closed
proportionally to overspeed
• Power/Load Unbalance (PLU)
– Overspeed protective circuit looking at
mechanical vs. electrical load
– Slams valves without tripping to prevent
overspeed on sudden load loss
• Overspeed Protection Control (OPC)
– Slams governing valves at 103% rated
speed, but allows re-opening
• Fast Valving / Early Valve Actuation
– Rapid valve closure at high load loss to
prevent overspeed and rotor angle instability
First Solar Power Plant Controller
VChadliev@firstsolar.com
First Solar Plant Control System Architecture
x Modern solar PV plants can contribute to the reliability and
efficiency of grid operation through sophisticated Power
Plant Controller by offering the following capabilities:
o Voltage / VAR control and/or Power Factor regulation

o Fault ride-through
o Real power control, ramping, and curtailment
o Primary frequency regulation
o Frequency Droop Response

x Centralized Solar PV pant controller can also play an
important role during NERC’s MOD 026-1 & 027-1
verification process
2
First Solar Plant Control System Architecture
ͻ Monitor conditions at the POI and compare PCS PLC
them with the Set Points
e e-
ͻ Send commands to inverters to PCS PLC -
make output adjustments and achieve
POI requirements e--
zzz
zzz
Grid
zzz
e-
Point
of Inter- zzz
connection
(POI) Control Signal e-

PCS PLC
Voltage
MW
MVAR zzz
PCS PLC zzz
Closed Loop ~100 ms Cycles zzz
zzz
Electricity Flow
e-
e- 3
Agua Caliente 290MW (AC)
Palo Verde Nuclear

Generating Station
Hassayampa
Agua Caliente Substation
California
Hoodoo Wash
IInc.
Arizona
First SSolar,
Substation
2013, Fi
i h 2013
Copyright l
North Gila 500kV Palo Verde-Hassayampa
Substation Transmission Line
©C
4
Agua Caliente - Typical Plant Operating Day (March 19, 2014)
1.2 110%
Power (PU)
0.9
Power (PU) and Reactive Power (PU)
Voltage (PU) and Power Factor

105%
0.6
Voltage (PU) Power Factor

0.3 100%
0.0
Reactive Power (PU)

95%
-0.3
-0.6 90%
4:00 8:00 12:00 16:00 20:00
Time of Day 5
Agua Caliente Voltage Support – Event on March 21, 2014

6
Agua Caliente Voltage Support – Event on March 21, 2014
1.2 110%
500kV Palo Verde -
Hassayampa line Power (PU)
taken out of service
0.9
Power (PU) and Reactive Power (PU)
105%
0.6
Voltage (PU)
100%
Voltage (PU)
PU))
Night
0.3
Shutdown

95%
0.0
The PV plant is asked to
Reactive Power (PU)
support voltage by APS
Operators.
-0.3 90%
4:00 8:00 12:00 16:00
Time of Day 7

Applicable NERC ReliabilityStandards

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Applicable NERC ReliabilityStandards

Transféré par

Droits d'auteur :

Formats disponibles

Applicable NERC Reliability

ͻ The material presented here is not intended as compliance

NERC MOD Standards

System Modeling Plant Modeling

Standard Applicability Effective

MOD-026-1 Generator Owners and Transmission Planners 7/1/14

MOD-033-1 Planning Coordinators, Reliability Coordinators, and 7/1/2017

ͻ Ensure accurate information available for planning models

Pmax 2.1. P and Q capability

Qmin Qmin * Other than wind & solar PV

Recorded data for verification:

Resource Excitation control system or Turbine/governor and load

R1: Each TP provides information to the GO upon request:

Standard MOD-026-1 MOD-027-1

R3: GO provides written response to TP after receiving from TP:

ISO-NE System Operations

• Tied to Québec only through

• 7.1 million retail electricity customers drive

• 600 MW of active demand

efficiency with capacity supply 0

• Population and electric demand are

• 8,600 miles of high-voltage transmission

to power systems in New York and Québec

• ISO-NE Operation Support Services perform thermal, voltage and

• To ensure our transient stability simulations provide accurate

• Real-time Oscillation Alarm --> Email

• Post Event Analysis

• Event detection: Oscillation monitoring

• Normally cleared single phase to ground fault on a nearby 345

• Combined cycle plant:

PMU & TSAT Real Power (MW) at POI

TSAT Volt-Freq Curve

PMU & TSAT Reactive Power (MVar) at POI

• In-house developed bridging tool – Java

NX-9 Bridging Modified

data for the events Updated in real time

tested Auxiliary files (e.g.,

NERC Power Plant Model Validation Workshop

• Two large-scale power outages in August 10, 1996

Benefits of WECC generator testing program are

Need to sustain model validation efforts was evident

Validation studies done for August 10 1996 outage

In early 2000s, WECC was developing a frequency

Under Load Control

Reasonable correspondence 3000

between actual and simulate 2500

2013 – NERC approved:

2014 – NERC approved MOD-033 requiring plant operator

• Baseline model development

• Periodic model verification

ZREF uop Gmax

• Needed to establish the correct model structure

Many modern digital controls map directly to models

Practical limitations usually exist in large thermal plants

Plant Disturbance Monitoring

POI Disturbance Monitoring

MWs and MVARs

Voltage and frequency are played in as inputs

Voltage and frequency are inputs

Blue line = actual recording

Voltage and frequency are inputs

Blue line = actual recording

• It is a good practice for Transmission Planners to

Data is streamed continuously to BPA control centers