nerctranslate this page modeling powerflow and...º@gpb'p@gè¯aun° x¡*da$t ‡Ê|†j...
TRANSCRIPT
North American Transmission Forum Modeling Activities
NERC Modeling Workshop
October 1-3, 2012
We’ve Come a Long Way
• 1970s and 80s – Real-time analytical tools • More sophisticated software • Faster, more powerful hardware
Then
Now
Our Models Don’t Always Fit Reality
• Case: Poor event simulation
• Case: Bad generator spec almost led to $3M transmission project
• Case: Data update errors cause project to stumble
For Example… 4000 MW Generation Failure Eastern Interconnection
Frequency
Time - Seconds 0.00 14.30 28.50 42.70 56.90
60.03
59.85
Legacy Model
Best Generic Model
Actual Event
Hurdles and Constraints
• Deregulation Data walls
• Legacy modeling practices Model errors • Companies’ models aren’t always compatible Takes time to fix
• Terminology confusion Wrong input parameters
• Equipment specs aren’t up-to-date Wrong results
Goal 1: Improve Model Accuracy Generator Specifications Worksheet
PMIN, QMAX
PMAX, QMAX
PMIN, QMIN PMAX, QMIN
Goal 1: Improve Model Accuracy PMAX, PMIN, QMAX, QMIN
Goal 1: Improve Model Accuracy Reference Documents and Diagrams
Goal 1: Improve Model Accuracy Generator Characteristics Reference Documents and Diagrams
Goal 2: Improve Model Compatibility
• Address Planning, Operations Planning, and Real-time operation – Within your company and with others
• Improve vendor software interoperability
• Adopt common data protocols
• Develop our own machine models – No “black-box” models!
• Move from bus-branch to equipment model
Involves vendors, manufacturers, etc. as stakeholders
Real-time Ops Planning Planning
Now Next hour through next 12 months
One year and beyond
Utility 1
Utility 2
Utility 3
Inconsistent definitions and poor representation between neighboring utilities
Inconsistent equipment specs and topology within utility models
Goal 2: Improve Model Compatibility Internal and External Compatibility and Accuracy
Goal 2: Improve Model Compatibility
Station A
Station B
Station A
Station B Bus-Branch Model
Equipment (Breaker-Switch) Model
Move to the Equipment Model
Goal 3: Improve Model-Development Procedures
Implement Chronological Model-building
Goal 4: Manage Expectations
• If we don’t fix this, NERC, FERC, DOE, will!
• Clarify who provides what data to whom, and when?
• Get the data right! (See Goal 1)
• It can’t be done overnight, but we’ve already started
• Develop training programs for those developing models
• Expect improvement, not perfection
• Must have executive-level support!
Approach
• Multi-faceted problem requires multi-faceted solution
• Meet with industry do-er’s & movers – NERC, ERAG, WECC, ERCOT, NATF & GO reps
– Identify practices, publish supporting documents
– Promote stakeholder cooperation & communication
– Avoid duplication, consolidate efforts where appropriate
• Develop a plan – Already started
– Some utilities already employ recommended techniques
• Get it done!
Schenectady 237 Miles
Questions?
WECC Composite Load Model
NERC Modeling Workshop
Prepared by WECC Load Modeling Task Force Presented by Dr.John Undrill
2
• In the first 40 years of digital simulations, the focus was on modeling the supply side • We have reasonably good power plant models
• Models data is to get more consistent as more plants have digital controls
• We have tools for power plant model validation using disturbance data
• Over the past decade, the focus has shifted on modeling the supply side
Power System Modeling
Changing Nature of Electrical Loads
Data Centers AC and Heat Pumps
Resistive Cooking Resistive Heating
Incandescent Lighting
Distributed Generation
Power Electronics
Share of total system load
Changing Nature of Electrical Loads
400
420
440
460
480
500
520
540
560
-10 0 10 20 30 40400
420
440
460
480
500
520
540
560
-10 0 10 20 30 40
Transmission voltage during a fault in an area with mainly resistive loads
Transmission voltage during a fault in an area with high amount of residential air-conditioning load
Changing Nature of Electrical Loads
• Electronic loads, VFDs, AC compressors, CFLs are increasing their penetration
• Resistive loads are phasing out
• Electrical Loads play much more influential role in power system stability – Load-Induced voltage stability
– Damping of inter-area power oscillations
• Do your planning studies reflect this ?
6
o
o
o o o o
o
o
o o o o
Simulations – instantaneous voltage recovery
This is what we thought would happen using old load model…
Need for Better Load Modeling
30 seconds
7
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
1.10
0 5 10 15 20 25 30Seconds
Volta
ge (p
u)
… and this is what actually happened
Reality – 30-second voltage recovery, 12 seconds below 80%
Need for Better Load Modeling
30 seconds
100%
75%
8
• 1980’s – Constant current real, constant impedance reactive models connected to a transmission bus o Reflected the limitation of computing technologies of that time
• 1990’s – EPRI Loadsyn effort o Several utilities use static polynomial characteristics for load
representation
• 1990’s – IEEE Task Force recommends dynamic load modeling o The recommendation does not get much traction in the industry
• 1996 – BPA model validation study for August 10 1996 outage: o Need for motor load modeling to represent oscillations and
voltage instability
History Of Load Modeling in WECC
2001 WECC “Interim” Load Model
• 2000 – 2001 – WECC “Interim” Load Model: • Presently used to plan and operate Western Interconnection
power system • 20% of load is represented with induction motors, the
remaining load is static, mainly constant current active, constant impedance reactive components
• Was the only practical option available in 2001 • “Interim” load model was intended as a temporary solution to
address oscillation issues observed at California – Oregon Intertie
• The model limitations and the need for a composite load model were recognized
10
• Late 1980’s – Southern California Edison observed events of delayed voltage recovery attributed to stalling of residential air-conditioners o Tested residential air-conditioners, developed empirical AC
models
• 1997 – SCE model validation study of Lugo event:
o Need to represent a distribution equivalent
o Need to have special models for air-conditioning load
History Of Load Modeling in WECC
Southern California Edison
Lugo Event – Load Modeling Lessons:
A. Need to represent a distribution equivalent
B. Need to have special models for residential air-conditioners
Model was used in Southern California for special studies using PTI PSS®E simulator
12
• 1994 – Florida Power published an IEEE paper, used a similar load model
• 1998 – Events of delayed voltage recovery were observed in Atlanta area by Southern Company, the events are analyzed and modeled
• Southern Company and Florida Power used in principle similar approaches to SCE’s and eventually WECC model
• The model use was limited to special studies of local areas
Load Modeling Efforts in the East
13
• 2005 – WECC developed “explicit” load model: o Adding distribution equivalent to powerflow case WECC-
wide
o Modeling load with induction motors and static loads
o Numerically stable in WECC-wide studies !
• 2007 – PSLF has the first version of the composite load model (three-phase motor models only)
• 2006-2009 – SCE-BPA-EPRI testing residential air-conditioners and developing models
• 2009 – residential air-conditioner model is added to the composite load mode
WECC Load Modeling Task Force
WECC Composite Load Model (CMPLDW)
Electronic
M
M
M 69-kV 115-kV 138-kV
Static
AC
12.5-kV 13.8-kV
UVLS
UFLS GE PSLF Siemens PTI PSS®E Power World PowerTech TSAT
Load Model Data
Electronic
M
Load Model Composition Data
M
M
Static
Load Component Model Data
Distribution Equivalent Data
UVLS and UFLS Data
M
69-kV 115-kV 138-kV
Distribution Equivalent Data
Electronic
M
M
M
69-kV 115-kV 138-kV
Static
M
X = 8% LF = 110 to 140% Tap = +/- 10%
∆V = 4 to 6% X/R = 1.5 PL < 7% B1:B2 = 3:1
V = 1.02 … 1.04
V > 0.95
B1 B2
R + j X
Bss
Electrical End-Use
Summer peak demand in California
0
1
2
3
4
5
6
7
8
Res. - AirConditioning
Com'l. - AirConditioning
Com'l. -InteriorLighting
Com'l. - Other Res. -Miscellaneous
Res. -Refrigerator
Com'l. -Ventilation
Res. -Cooking
Res. - Dryer Com'l. -Refrigeration
0
5
10
15
20
25
30
35
Peak Demand
Annual Consumption
Source: CEC Demand Analysis Office
Peak Demand (GW) Consumption (TWh)
Residential AC
Commercial AC
Lighting Refrigeration Ventilation
Commercial Buildings
20
< 1%4%< 1%5%
6%
26%
< 1%9% 4%< 1%
11%
33%
< 1%
Cooling
Ventilation
Refrigeration
Lighting
CEC California Commercial End-Use Survey Summer Peak Load
Cooling Units for Data Centers, Computer and Telecommunication Rooms
Cooling 10-25 hp compressor motors Roof-Top Direct Expansion HVAC
Central Cooling System Chiller 200-250 hp compressors
22
• Usually special design motors • Constant torque load • Roof-top Air-conditioning / Refrigeration:
o Compressors will have contactors, disconnect at 40 to 50%, reconnect at 45 to 55%, very little time delay
o Very likely to restart immediately • Large Chillers:
o Compressors will trip for a delayed voltage sag and likely to lock out
Compressor Motors
23
• Usually NEMA Design B motors • Speed-dependent torque load • Fan inertia is 0.5 to 1.0 second range, pump inertia is 0.1 to
0.2 range
• Many of pumps and ventilation fans use Variable Frequency Drives for improved efficiency
Fans and Pumps
24
• Increasingly used in commercial buildings circulating pumps, fans, etc
• Behave as a static constant power load • Power factor
o Almost unity at positive sequence o 0.75 RMS because of harmonics
• Trips at 60 to 70% voltage
• BPA and SCE have tested a few and continue testing VFDs
Variable Frequency Drives
Fluorescent Lights
• Vast majority of fluorescent lights are electronic ballast • Significant harmonics • Active power is almost constant current • Reactive power is capacitive and constant current
4 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.1-200
-100
0
100
200
Time
Vol
tage
Fluorescent Voltage and Current
4 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.1-4
-2
0
2
4C
urre
nt
0 20 40 60 80 100 120 140-10
0
10
20
30
40
50
60
Voltage [V]
Rea
l Pow
er [W
]
0 20 40 60 80 100 120 140-30
-25
-20
-15
-10
-5
0
5
Rea
ctiv
e P
ower
[VA
R]
Fluorescent Powerwaveform voltage sensitivity
Residential Loads
Residential Air-Conditioning
1-phase A/C Compressor Motors are Prone to Stall
7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 120
50
100
150
200
250Voltage (Volts)
7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 120
50
100
Current (Amps)
Time (sec)
7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 120
5
10
15
20Active Power (kW)
7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 120
5
10
15
20
Time (sec)
Reactive Power (kVAR)
Dip to 55% for 3 cycles
Stall Thermal Trip
Single-phase AC compressors stall for a short voltage sag and remain stalled even when the voltage is recovered
To find out WHY, SCE, BPA and EPRI tested more than 30 AC units
29
Test Findings: Compressor Motor Steady-State Loading
80 85 90 95 100 105 110 1152.6
2.8
3
3.2
3.4
3.6
Pow
er (k
W)
Ambient Temperature (F)80 85 90 95 100 105 110 115
0.56
0.58
0.6
0.62
0.64
0.66
Sta
ll V
olta
ge (p
er u
nit)
• Compressor loading and stall voltage depend on the ambient temperature
• Compressor motors have high power factor ~0.97 when running
(a) A/C Compressor Motors are non-symmetric
R
S
C Run Capacitor
Thermal Relay
Auxiliary Winding Main
Winding
1 phase supply, 2 windings
capacitor-run motor
31
(b) Compressor Motors Inertia is Very Low
310 mm
75 mm
E.g. 3.5-ton compressor motor: Weight: 4.6 kg
H = 0.03 – 0.05 seconds
32
(c) Compressor Load Torque in very cyclical
360 720
Torq
ue
Rotor Position
It is very possible that the motor stalls at the next compression cycle
33
Compressor Motor Tests – Power-Voltage Trajectories
0 50 100 150 2000
2000
4000
6000
8000
10000
12000
Voltage [V]
Com
pres
sor R
eal P
ower
[W]
Real Power
RUN
STALL
STALL
115F110F105F100F95F90F85F80F
0 50 100 150 2000
2000
4000
6000
8000
10000
12000
Voltage [V]
Com
pres
sor R
eact
ive
Pow
er [V
AR
]
Reactive Power
RUN
STALL
STALL
115F110F105F100F95F90F85F80F
* note motor load and stall voltage increase with temperature
34
Single-Phase Motor Models
• Three-phase motor models cannot represent behavior of single-phase motors with the same data set: – Stalling phenomenon (3-phase motor model usually stalls at much lower
voltages)
– Real and reactive power when stalled
– Steady-state sensitivities of real and reactive power with respect to voltage and frequency
• Single-phase motor models exist but require point-of-wave simulations – Not acceptable for positive-sequence grid simulators
Phasor Model (MOTORC)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1-15
-10
-5
0
5
10
15
time
aux.
win
ding
cur
rent
Dynamic Phasors model sinusoidal waveforms with changing magnitudes and phase.
The dynamic phasor motor model resembles traditional positive sequence motor models (motorw), but includes effects of asymmetric machine design, and unbalanced operation.
36
Performance Model
0 0.2 0.4 0.6 0.8 1 1.20
1
2
3
4
5
6Real Power
Rea
l Pow
er (p
er u
nit)
Voltage (per unit)
RUNSTALL
STALL
0 0.2 0.4 0.6 0.8 1 1.20
1
2
3
4
5
6Reactive Power
Rea
ctiv
e P
ower
(per
uni
t)Voltage (per unit)
RUN
STALL
STALL
Motors stall when voltage drops below Vstall for duration Tstall A fraction Frst of the aggregated motor can restart when the voltage exceeds Vrst for duration Trst
37
Model Benchmarking
• GE PSLF has MOTORC and performance LD1PAC models
• Siemens PTI PSS®E has performance ACMTB model • WECC with support from DOE is in process of
benchmarking single-phase motor models – PSCAD model with detailed 3-phase representation of a feeder – Positive sequence load model with MOTORC dynamic model – Positive sequence load model with BPA performance model
Event Validation Studies with LD1PAC Model
Load Composition
40
• Apply your judgment !
• California Energy Commission: 2006 California Commercial End-Use Survey (CEUS)
• LBNL Reports on Electricity Use in California • PNNL DOE2 building simulations • BPA-PNNL End-Use Load Characterization
Assessment Program (ELCAP) • BPA Building Data • Load shapes provided by WECC members
Load Composition Information
California Commercial End-Use Survey
• 15 climate zones in California • Four seasons • Typical, Hot, Cold, Weekend • 24-hour data
0 5 10 15 20 250
2
4
6
8
10
12
14x 10
5
FCZ10 Season:Summer Day:Typical
AirCompMotorsProcessMiscOfficeEquipIntLightExtLightRefrigCookingWaterHeatVentCoolingHeating
Data is available on CEC web-site
< 1%4%< 1%5%3%
25%
< 1%8%
3%< 1% 10%
41%
< 1%
FCZ10 Season:Summer Day:Typical Hour:
16
AirCompMotorsProcessMiscOfficeEquipIntLightExtLightRefrigCookingWaterHeatVentCoolingHeating
bott
om
top
42
• CEC / BPA / DOE hired PNNL to develop load composition model
• Detailed models of various building types • WECC developed mapping from end-uses to model
components • Inputs are:
o City, climate conditions, # of buildings for a given feeder • Output:
o Load shapes and load composition data
PNNL Load Composition Model
0.00
5.00
10.00
15.00
20.00
25.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Load
(MW
)
Hour of day
ZIP
Motor-D
Motor-C
Motor-B
Motor-A
Electronic
City, climate data
Buildings
Estimated Load Composition and Load Shapes
44
• There is a balance between precision and the amount of effort required to maintain the data sets
• Based on the understanding developed using PNNL tool, WECC developed a simplified LCM version to create default data sets
• Produces load composition data for summer (normal, peak, cool), shoulder (normal) and winter (normal) days
WECC Load Composition Model
45
WECC Load Composition Model
Spreadsheet is available from WECC
WECC Approach
47
• WECC asked utilities to provide “Climate Zone IDs” with the base cases for every load larger than 5 MW
• WECC region was divided into 12 climate zones
• “Climate Zone ID” includes a climate zone + type of
substation (RES/COM/MIX/RAG)
• Industrial loads are identified explicitly and have
their own identifiers
Climate Zone IDs
48
Climate Zones
NWI
NWV
NWC` RMN
HID
DSW
NCC
NCV
SCC SCV
NWC – Northwest coast NWV – Northwest valley NWI – Northwest inland RMN – Rocky mountain NCC – N. Calif. coast NCV – N. Calif. Valley NCI – N. Calif. Inland HID – High desert SCC – S. Calif. coast SCV – S. Calif. Valley SCI – S. Calif. Inland DSW – Desert southwest
49
WECC Climate Areas ID Climate Zone Representative City NWC Northwest Coast Seattle, Vancouver BC NWV Northwest Valley Portland OR NWI Northwest Inland Boise, Tri-Cities, Spokane RMN Rocky Mountain North Calgary, Montana, Wyoming NCC Northern California Coast Bay Area NCV Northern California Valley Sacramento NCI Northern California Inland Fresno SCC Southern California Coast LA, San Diego SCV Southern California Valley LA, San Diego SCI Southern California Inland LA, San Diego DSW Desert Southwest Phoenix, Riverside, Las Vegas
HID High Desert Salt Lake City, Albuquerque, Denver, Reno
50
• Residential – typical of your suburban neighborhood
• Commercial – typical of downtown load • Mixed (default) – mix of residential and
commercial loads • Rural / agricultural areas
Substation / Feeder Types
51
• A base case includes “Climate Zone IDs” populated
• Default load composition data sets exist for 12 climate zones X 4 feeder types + 10 industrial loads o Can be generated for 4 seasons, 24 hours
• Load Model matches “Climate Zone IDs” with the load composition data sets and creates DYD records for PSLF o Will be a program build-in feature in the future
• The process is simple and effective
Load Model Data Tool
Load Model Validation Studies
Reproducing Delayed Voltage Recovery Events with CMPLDW
Simulations of delayed voltage recovery event due to air-conditioner stalling Done by Alex Borden and Bernard Lesieutre at University of Wisconsin
August 4, 2000 Oscillation – Interim Model
525
527
529
531
533
535
537
539
541
543
545
0 5 10 15 20 25 30 35 40
Volta
ge (k
V)
Time (sec)
Malin Voltage Actual Malin Voltage - Simulated MOTORW
August 4, 2000 Oscillation - CMPLDW
524
526
528
530
532
534
536
538
540
542
544
546
0 5 10 15 20 25 30 35 40
Malin Voltage Actual Malin Voltage Simulated
56
• We can now achieve the great accuracy with generator models: o We model physical equipment that is well defined and under
our control
• We will never be able to achieve a comparable level of accuracy with load models o Yes, we can tune load models to accurately reproduce and
explain past events
o But, Load models is only capable of predicting the future load response only in principle, and not in detail
Load Modeling – Setting Expectations
Where we are now … • WECC Composite load model version 1 is
implemented in GE PSLF and Siemens PTI PSS®E, similar models exist in Power World, Power Tech TSAT
• Default sets are developed:
– 12 climate zones in WECC,
– four types of feeders
– Summer, winter and shoulder conditions
• Tools are developed for load model data management
… Where we are now … • WECC is taking phased approach for approving the
composite load model for TPL compliance studies
– Phase 1: air-conditioner stalling is disabled by setting Tstall parameter to a large number
– Phase 2: better understand the reliability implications of delayed voltage recovery due to air-conditioner stalling, develop appropriate reliability metrics
• WECC members are conducting system impact studies for Phase 1
… Where we are now • Tens of thousands runs have been done with
the composite load model up to date
• Data sets are revised based on the validation studies
• It is a very good idea to test your plans of service against the composite load model even it is not currently approved for TPL compliance studies
60
61
62
Contacts
• Dmitry Kosterev, BPA, WECC LMTF Chair, [email protected]
• Donald Davies, WECC Staff, [email protected]
• Jun Wen, SCE, [email protected]
• Stephanie Lu, WECC MVWG Chair
Power Plant Modeling
NERC System Modeling WorkshopMinneapolis, October 2012
John Undrill
Sunday, 14 October 12
Sunday, 14 October 12
Procedings of the Royal Society5 March, 1869
Sunday, 14 October 12
Modeling Requirements
Grid simulations require Load Flow and Dynamic Simulation models that are:
A practical and reasonably accurate representation of how the power system should be expected to behave
Consistent as to level of detail and sphere of applicability across the entire power system
Well behaved and free of ‘quirks’ associated with any one given simulation program
Stable in the sense that a model setup that worked properly for one engineer today will work properly for a different engineer tomorrow
Sunday, 14 October 12
Common Load Flow data Errors
It is instructive to look at the more common load flow data errors:
Wrong generators on lineCombined cycle steam turbines ON when GTs are OFF
Incorrect area swing power assignments
Overloaded/underloaded generatorsIncorrect Combined cycle loadingsAmbient temperature issues (gas turbines, generator cooling)
Overloaded transformersIncorrect impedance / MVA base / tap
Sunday, 14 October 12
Generator Capability Curve
The capability curve describes the generator.
It DOES NOT describe the real or reactive power capability of the generating unit.
The electrical capability of the unit depends on- the generator- the exciter- the step up transformer- generator and transformer protection settings- protections and limits in controls- the system and auxiliary load (MW and MVAR)
Sunday, 14 October 12
Generator Capability Curve
Sunday, 14 October 12
Generator Capability Curve
Sunday, 14 October 12
Generator Dynamic Characteristics
Sunday, 14 October 12
Dynamics Modeling
Initial loads and generator terminal conditions are established by a load flow solution
Generator control, turbine, and protection models are initialized to match generator terminal conditions
Similar process will evolve for solar, battery, flywheel, etc
Data is required to describe what each unit would do - in normal conditions- in emergency conditions that might last for a up to about 30 seconds
Dynamic simulations need:Generator MVA ratings and dynamic characteristicsExcitation system dynamic models, ratings, limiting detailsTurbine control dynamic modelsDynamic models of electronic power coupling systems
Sunday, 14 October 12
Dynamics Modeling
Initial loads and generator terminal conditions are established by a load flow solution
Generator control, turbine, and protection models are initialized to match generator terminal conditions
Similar process will evolve for solar, battery, flywheel, etc
Data is required to describe what each unit would do - in normal conditions- in emergency conditions that might last for a up to about 30 seconds- but we have just experienced a major grid event that evolved over 30 minutes
Dynamic simulations need:Generator MVA ratings and dynamic characteristicsExcitation system dynamic models, ratings, limiting detailsTurbine control dynamic modelsDynamic models of electronic power coupling systems
Sunday, 14 October 12
Dynamics Modeling
Electrical subsystems
Sunday, 14 October 12
Sunday, 14 October 12
Sunday, 14 October 12
Sunday, 14 October 12
Generator Data
Generator magnetic parameters are criticalMagnetization curveSynchronous reactance
Excitation sizing/rating/limits must be properly coordinated with generator excitation requirement
Generator inertia constant and transient/subtransient reactance are important when transient stability is a main concern
Generator rotor time constants and excitation system transient gain are important when oscillatory behavior is the main concern
Current-technology excitation systems incorporate MANY limits and secondary controls
Overexcitation/underexcitation limitsStator current limitsPower factor control
Sunday, 14 October 12
Generator Data
Sunday, 14 October 12
Generator Magnetization
Curve
Sunday, 14 October 12
Generator Dynamic Characteristics
Sunday, 14 October 12
Common Dynamics Modeling Errors
It is instructive to look at the more common modeling errors:
Incorrect generator magnetic parameters eg L” > L’
Incorrect sizing/ratings of excitation system equipment
Incorrect damping factors in generators / turbines
Incorrect turbine power capability (data handling problem in PSS/E)
Too many plants operating in ‘droop governing’ mode (cf recent ERAG work)
Inadequate recognition of plant secondary controls
Sunday, 14 October 12
Generator Modeling
Gensal model is obsolete (reflects computer limitations of 1970s)
Preferred generator model is gentpf or gentpj (PSLF) or genrou (PSS/E)
Excitation system modeling in present programs is out of date and not adequate for representation of current technology generator/excitation controls
- OEL/UEL modeling is either oversimplified or too elaborate- Stator current limiting is not properly represented- Data management for excitation system elements is inadequate
- No intelligent / safe defaults- Inadequate error checking
Sunday, 14 October 12
How to get Excitation System Modeling Data
Need to determine
Type - DC, Brushless, Transformer Fed (PPT, SCPT)Manufacturer - GE, W, ABB, Basler, etcModel - WMA, EX2100, Alterex, DECS200,
If Transformer fed -Excitation transformer rating, voltages, impedance
If brushless -Control power source, PMG, aux bus
If new/digital -Form of transfer function, P, PI, PID
Sunday, 14 October 12
Excitation System Model
Sunday, 14 October 12
Excitation System Model
Sunday, 14 October 12
Excitation System Model
Sunday, 14 October 12
Dynamics Modeling
Mechanical/control subsystems
Sunday, 14 October 12
Sunday, 14 October 12
Frequency and Power Flow Control
Dips go down to -60 mHzPeaks to up to +40mHz
For reference - GE programmed deadband = 0.025% = 5 mHz percent
Sunday, 14 October 12
Frequency and Power Flow Control
Sunday, 14 October 12
Frequency and Power Flow Control
Sunday, 14 October 12
Sunday, 14 October 12
Sunday, 14 October 12
Relative effect of generation and load frequency sensitivity
Sunday, 14 October 12
Deadband
190
195
200
205
210
215
220
3 3.05 3.1 3.15 3.2 3.25 3.3 3.35 3.4
TNR-TNH
DWATT
63
64
65
66
67
68
69
3 3.05 3.1 3.15 3.2 3.25 3.3 3.35 3.4
TNR-TNH
FSR
Programmed deadband = 0.025 percent
6.25 mHz in 60 Hz system5.00 mHz in 50 Hz system
Sunday, 14 October 12
Frequency and Power Flow Control
Dips go down to -60 mHzPeaks to up to +40mHz
For reference - GE programmed deadband = 0.025% = 5 mHz percent
Sunday, 14 October 12
Sunday, 14 October 12
Response with 100 percent participation in primary control
Sunday, 14 October 12
Response with 30 percent participation in primary control
Sunday, 14 October 12
Response with 100 percent participation in primary control
Response with 30 percent participation in primary control
Sunday, 14 October 12
Sensitivity of frequency dip to net system inertia constant
Sunday, 14 October 12
Sensitivity of frequency dip to net fraction of capacity contributing primary response
Sunday, 14 October 12
Sensitivity of frequency dip to principal time constant of primary response
Sunday, 14 October 12
Sunday, 14 October 12
Frequency dip response along the chain of subsystems
Sunday, 14 October 12
Turbine power response along the chain of subsystems
Sunday, 14 October 12
Sunday, 14 October 12
Measured frequency of Eastern Interconnection following a loss of generation
Sunday, 14 October 12
Basic Steam Turbine Control Model
Sunday, 14 October 12
Legacy IEEE Type 1 Steam Turbine Control Model
Sunday, 14 October 12
Turbine control - Westinghouse ancestry
Sunday, 14 October 12
Turbine Control - GE Ancestry <<< Rung Number 39 >>> ╔═════════════════════════════════════════════════════════╗ ║ FSRNV3 - SPEED CONTROL FSR ║ ║ ║ ║FSRMAX max┌─────┐ ║ >──╫──────────────────────────────────┤ │ ║ ║FSRMIN min│CLAMP│ ║ >──╫──────────────────────────────────┤ _│ ║ ║FSKNH │ / │ FSRN║ >──╫─────────────┐-7 │ / ├──────┬─────────╫──< ║TNKRNR + ┌┴┐ ┌─┤_/ │ │ ║ >──╫─────────O──┤x├─────┤ ├─┐ │ └─────┘ +│- FSRNDIF║ ║TNH -│ └─┘ · │ └───────────┬──O─────────╫──< >──╫─────────┘ 0x7FFF · │ ┌───┐ │ ║ ║ ────────┤/├─┴──────────┤ │ │ ║ LFALSE 2║L_hpsor · │MIN│ + │ ║ ────────────────────╫─────────────────────┘ -7 │ ├──O─┘ ║ ║FSKNG ┌─┐ │SEL│ +│ ║ >──╫──────────────────────────────┤x├──┤ │ │ ║ ║L83SCDB └┬┘ └───┘ │ ║ >──╫────────────────────────┐ │ │ ║ TNRL 1║Speed_sp + · │ │ TN_ERR║ ────────────────────╫────────────O─┬────────┤/├┬────┴──────────(──────────────╫──< TNHF 0║Speed -│ │ · │ │ ║ ────────────────────╫────────┤/├─┤ │ ┌─┤ ├┘ │ ║ ║TNH · │ │ └────────────────┐ │ ║ >──╫────────┤ ├─┘ ├────────────────┐ │ │ ║ ║L83HOST · │ ┌────┐ │ │ │ ║ >──╫─────────┘ ├─────┤A │ │ │ │ ║ ║ │ 0 │ A>B├─┐ │ │ │ ║ ║ │ ───┤B │ · │ │ │ ║ ║TNKEDB │ └────┘ · -│+ │ │ ║ >──╫────────────┬─(─────────┬─┤ ├──O──┤ ├─┤ │ ║ ║ │ │ │ · +│ · │ │ ║ ║ │ │ ┌──────┐└─┤/├──┘ · │ │ ║ ║ │ └─┤A │ · │ │ ║ ║ │ │ │A│>B├───────────┤ │ │ ║ ║ └───┤B │ 0 · │ │ ║ ║ └──────┘ ──┤/├─┘ │ ║ ║FSR ┌───────────┐ │ ║ >──╫───────────────────┤V V │ │ 0x7FFF FSRNH║ ║FSKNTC │ ────=OUT├──────────┘ ─────────────╫──< >──╫───────────────────┤T 1+Ts │ 0x7FFF FSRNL║ ║ ┌┐ init ├───────────┤ ─────────────╫──< ║ ┘└ ──────┤RESET:OUT=V│ ║ ║ └───────────┘ ║ ╚═════════════════════════════════════════════════════════╝
Sunday, 14 October 12
Steam plant control elements
Sunday, 14 October 12
Sunday, 14 October 12
Sunday, 14 October 12
Sunday, 14 October 12
ggov1 Turbine Control Model
Sunday, 14 October 12
Current Technology Hydro Turbine Control Model
Sunday, 14 October 12
Action of Turbine Load Controller
Sunday, 14 October 12
Simple Turbine Load Controller Model
Sunday, 14 October 12
Everything else
Generator protection
Minimal models in PSLF and PSS/E
Do not assume that generator protection relay settings are fully coordinated with limits and protections in excitation and turbine controls
Plant auxiliaries
Include in the simulation setup as appropriate
Need to ensure that auxiliary loads are properly coordinated with turbine output in load flow base cases
Sunday, 14 October 12
Wind PlantsWECC, NERC, IEEE working groups continue to struggle with model development.
Model development is evolutionary and reactive to instances of wind plant behavior.
Period = 80 seconds Period ~= 0.1 seconds
Sunday, 14 October 12
Detail in Simulations is often Misleading
Increasing the detail of a model with regard to the equipment components that it represents SHOULD NOT BE ASSUMED TO MAKE IT MORE ACCURATE
It is entirely possible for simulations made with very detailed equipment models to produce inaccurate results
The introduction of detailed modeling can give a false impression as to the quality of the simulation
A ‘perfect’ model is valid only if the equipment it describes is in service and in the operating mode considered by the model
Sunday, 14 October 12
Frequency Response & Modeling
NERC Modeling Workshop – Bloomington, MN October 1-3, 2012
2 RELIABILITY | ACCOUNTABILITY
Modeling Eastern Interconnection Frequency Response
3 RELIABILITY | ACCOUNTABILITY
Actual (DFR)
Simulation with original modeled
governor response
Simulation with 20% governor response
Actual (DFR)
Simulation with original modeled
governor response
Simulation with 20% governor response
Sept. 18, 2007 Event Forensics
4 RELIABILITY | ACCOUNTABILITY RELIABILITY | ACCOUNTABILITY
EI FR Modeling
• Based on 4,500 MW loss event
• ~5,400 units above 20 MW
5 RELIABILITY | ACCOUNTABILITY
ERAG/NERC Modeling Findings
Best match performance characteristics:
• 30 % of units on line provide primary frequency response
• 2/3 of those units exhibit withdrawal
• 10 % of units on line sustain primary frequency response
Worldwide comparison (per John Undrill)
• 35 % response is typical
6 RELIABILITY | ACCOUNTABILITY
2010 Governor Survey Results
7 RELIABILITY | ACCOUNTABILITY
2010 Governor Survey
• Performed in September 2010
• Requested information and settings for turbine governors 20 MVA & Higher
Plants aggregate of 75 MVA or hither
• Three types of information requested Policies on installation, maintenance and testing procedures
Unit-specific characteristics and governor settings
Unit-specific performance information for a recent, single event
8 RELIABILITY | ACCOUNTABILITY
Generators as Reported
• Not all generators were reported
Interconnection Total Generators
Reported
Generators Reported as Having
Governors
Generators Not Having Governors
Eastern 4,372 (648.7 GW) 4,217 (630.2 GW) 152 (18.5 GW)
Western 1,560 (171.6 GW) 1,445 (162.9 GW) 114 (8.7 GW)
ERCOT 503 (95.6 GW) 446 (85.6 GW) 53 (9.0 GW)
Totals 6,435 (915.9 GW) 6,110 (878.7 GW) 319 (36.2 GW)
9 RELIABILITY | ACCOUNTABILITY
Survey Findings
• Widely varying understanding of role of turbine governors in frequency response Need for educational outreach
• Units with governors – 95% to 99% are operational
• Sustainable response – 80% to 85% are capable
• Unit-Level or Plant-Level Control Schemes that Override or Limit Governor Performance – roughly 50% FR withdrawal problem
• Governor operating philosophy as important as data
10 RELIABILITY | ACCOUNTABILITY
East
No Response, 159.9, 38%
Online, No Data on
Response, 53.2, 13%
Expected Response, 124.7, 30%
Opposite of Expected Response, 77.6, 19%
West
No Response, 34.6, 35%
Online, No Data on
Response, 3.4, 4%
Opposite of Expected Response, 16.9, 17%
Expected Response, 42.7, 44%
Texas
No Response, 7.8, 13%
Online, No Data on
Response, 8.6, 14%
Expected Response, 31.6, 53%
Opposite of Expected Response, 11.8, 20%
Response by Capacity On-Line
11 RELIABILITY | ACCOUNTABILITY
Gen. Response by Prime Mover
Eastern Interconnection
12 RELIABILITY | ACCOUNTABILITY
Gen. Response by Prime Mover
Western Interconnection
13 RELIABILITY | ACCOUNTABILITY
Gen. Response by Prime Mover
ERCOT Interconnection
14 RELIABILITY | ACCOUNTABILITY
Usability of Deadband Data
51%
63%
79%
53%
65%
77%
49%
37%
21%
47%
35%
23%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
East West Texas East West Texas
No. of Units Capacity
Unusable
Usable
15 RELIABILITY | ACCOUNTABILITY
Governor Deadband Settings
5402000700
0
50
100
150
200
250
300
350
400
<500 MW 500-1000MW
>1000 MW <500 MW 500-1000MW
>1000 MW <500 MW 500-1000MW
>1000 MW
East West Texas
Dea
dban
d Se
tting
(mH
z)
700
16 RELIABILITY | ACCOUNTABILITY
0
1
2
3
4
5
6
7
8
9
10
<500 MW 500-1000MW
>1000 MW <500 MW 500-1000MW
>1000 MW <500 MW 500-1000MW
>1000 MW
East West Texas
Dro
op S
ettin
g (%
)Droop Setting by Unit Size
17 RELIABILITY | ACCOUNTABILITY
Operational Status of Governors
394, 99%
4, 1%
0, 0%
1, 0%
Yes No N/A Unknown
Eastern Western
ERCOT
4015, 95%
39, 1%
26, 1%
128, 3%
1378, 97%
30, 2%
0, 0%
21, 1%
18 RELIABILITY | ACCOUNTABILITY
Response Sustainable > 1 Min.
333, 83%
27, 7%
2, 1%
37, 9%
Yes No N/A Unknown
Eastern Western
ERCOT
3359, 80%
360, 9%
9, 0%
480, 11%
1213, 85%
99, 7%
0, 0%
117, 8%
19 RELIABILITY | ACCOUNTABILITY
Over-Riding Outer-Loop Controls
197, 49%
168, 42%
2, 1%
32, 8%
Yes No N/A Unknown
Eastern Western
ERCOT
2026, 48%
1818, 43%
27, 1%
337, 8%
700, 49%
664, 46%
0, 0%
65, 5%
20 RELIABILITY | ACCOUNTABILITY
Use of Survey Data
• Provided to all NERC Regions
• Comparison to models underway
• Goal – improve models and highlight the importance of correct governor modeling
Long-Term Goal
• Unit & Plant pedigree and operational information database for modeling
21 RELIABILITY | ACCOUNTABILITY
ERCOT Experience w ith Deadbands
22 RELIABILITY | ACCOUNTABILITY
Deadbands in ERCOT
• Initially specified ±36 mHz deadbands (prior to 2010)
• Allowed stepped response at deadband
• Resulted in a flat frequency response for small disturbances
• Resulted in generators trying to respond by larger amounts when deadband was crossed
• Resulted in less stable operation when near boundary conditions of deadbands
23 RELIABILITY | ACCOUNTABILITY
ERCOT 2008 Frequency Profiles
23
September and March 2008 in 5 mHz Bins
24 RELIABILITY | ACCOUNTABILITY
Frequency Response
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
59.50 59.55 59.60 59.65 59.70 59.75 59.80 59.85 59.90 59.95 60.00 60.05 60.10 60.15 60.20 60.25 60.30 60.35 60.40 60.45 60.50
Hz
MW
Cha
nge
Deadband Setting
Hz600.000Capability (MW)
0.036
Step response at dead-band.
± 36 mHz Deadband – Step Response
25 RELIABILITY | ACCOUNTABILITY
Frequency Response
-150.00
-100.00
-50.00
0.00
50.00
100.00
150.00
59.50 59.55 59.60 59.65 59.70 59.75 59.80 59.85 59.90 59.95 60.00 60.05 60.10 60.15 60.20 60.25 60.30 60.35 60.40 60.45 60.50
Hz
MW
Cha
nge
Deadband Setting
0.0166 Hz600.000Capability (MW)
No Step response at dead-band.
± 16.6 mHz Deadband – No Step Response
26 RELIABILITY | ACCOUNTABILITY
ERCOT Frequency Profile
26
0
5000
10000
15000
20000
25000
30000
35000
40000
59.9
59.91
59.92
59.93
59.94
59.95
59.96
59.97
59.98
59.99 60
60.01
60.02
60.03
60.04
60.05
60.06
60.07
60.08
60.09 60
.1
One
Min
ute
Occ
uran
ces
2010 2008
January through September of each Year
27 RELIABILITY | ACCOUNTABILITY
±0.036 Hz Vs ±0.016 Hz Deadband
27
0
20000
40000
60000
80000
100000
120000
140000
59.9
59.91
59.92
59.93
59.94
59.95
59.96
59.97
59.98
59.99 60
60.01
60.02
60.03
60.04
60.05
60.06
60.07
60.08
60.09 60
.1
MW
2008 MW Response of 0.036 db 2010 MW Response of 0.0166 db
545670.0
404989.0 2010 MW Response of 0.0166 db 25.78% Decrease in MW movement with lower deadband.
2008 MW Response of 0.036 db
MW Minute Movement of a 600 MW Unit @ 5% Droop
28 RELIABILITY | ACCOUNTABILITY
Frequency Response Withdrawal
29 RELIABILITY | ACCOUNTABILITY
• Function of dispatch – what types of units are on line and responding
• Typical causes: Plant outer-loop control systems – driving the units to MW
set points
Unit characteristics o Plant incapable of sustaining
o Governor controls overridden by other turbine/steam cycle controls
Operating philosophies – operating characteristic choices made by plant operators o Desire to maintain highest efficiencies for the plant
Frequency Response Withdrawal
30 RELIABILITY | ACCOUNTABILITY
1,711 MW Loss – Sat 3:30 pm EDT
ΔF = 0.0722 Hz FR = -2,369 MW/0.1 HZ
Value A 60.021 HZ
Value B 59.948 Hz
31 RELIABILITY | ACCOUNTABILITY
1,049 MW Trip – Sun 11:20 pm EDT
ΔF = 0.0799 Hz FR = -1,312 MW/0.1 HZ
Value A 60.026 HZ
Value B 59.946 Hz
32 RELIABILITY | ACCOUNTABILITY
Frequency Response Initiative Implications
33 RELIABILITY | ACCOUNTABILITY
FRI Report Recommendation 1
Frequency Response Resource Guideline
• Define the expected performance characteristics
Existing Generator Fleet
• Deadbands of ±16.67 mHz
• Droop settings of 3%-5% depending on turbine type,
• Continuous, proportional (non-step) implementation
• Appropriate operating modes to provide FR
• Appropriate outer-loop controls modifications to avoid plant level primary FR withdrawal
34 RELIABILITY | ACCOUNTABILITY
FRI Report Recommendation 1
Frequency Response Resource Guideline
• Define the expected performance characteristics
Other Frequency-Responsive Resources
• Contractual high-speed demand-side response,
• Wind and photo-voltaic – particularly for over-frequency
• Storage – automatic high-speed energy injection
• Variable Speed Drives – non-critical, short time load reduction
35 RELIABILITY | ACCOUNTABILITY
FRI Report Modeling Implication
Needs:
• To improve FR modeling of existing plants
• To be able to model electronically-coupled resources and loads – model their FR characteristics
• Studies to ensure control parameters, gains, bandwidths, and their interactions don’t cause instability
Pacific Southwest Disturbance Warning Signs
• Address generator tripping during PSW disturbance
• UFLS design implications
36 RELIABILITY | ACCOUNTABILITY
Questions?