nerctranslate this page modeling powerflow and...º@gpb'p@gè¯aun° x¡*da$t ‡Ê|†j...

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


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


• 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

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

mailto:[email protected]�



Power Plant Modeling

NERC System Modeling WorkshopMinneapolis, October 2012

John Undrill

Sunday, 14 October 12

Procedings of the Royal Society5 March, 1869


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


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


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)


Generator Capability Curve


Generator Dynamic Characteristics


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


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


Dynamics Modeling

Electrical subsystems


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


Generator Data


Generator Magnetization

Curve


Generator Dynamic Characteristics


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


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


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


Excitation System Model


Dynamics Modeling

Mechanical/control subsystems


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


Relative effect of generation and load frequency sensitivity


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



Dips go down to -60 mHzPeaks to up to +40mHz

For reference - GE programmed deadband = 0.025% = 5 mHz percent


Response with 100 percent participation in primary control


Sensitivity of frequency dip to net system inertia constant


Sensitivity of frequency dip to net fraction of capacity contributing primary response


Sensitivity of frequency dip to principal time constant of primary response


Frequency dip response along the chain of subsystems


Turbine power response along the chain of subsystems


Measured frequency of Eastern Interconnection following a loss of generation


Basic Steam Turbine Control Model


Legacy IEEE Type 1 Steam Turbine Control Model


Turbine control - Westinghouse ancestry


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│ ║ ║ └───────────┘ ║ ╚═════════════════════════════════════════════════════════╝


Steam plant control elements


ggov1 Turbine Control Model


Current Technology Hydro Turbine Control Model


Action of Turbine Load Controller


Simple Turbine Load Controller Model


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


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


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


Frequency Response & Modeling

NERC Modeling Workshop – Bloomington, MN October 1-3, 2012

2 RELIABILITY | ACCOUNTABILITY

Modeling Eastern Interconnection Frequency Response


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


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


2010 Governor Survey Results


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


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)


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


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


Online, No Data on

Response, 3.4, 4%



Texas


Online, No Data on

Response, 8.6, 14%



Response by Capacity On-Line


Gen. Response by Prime Mover

Eastern Interconnection



Western Interconnection



ERCOT Interconnection


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


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


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


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%


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%


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%


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


ERCOT Experience w ith Deadbands


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


ERCOT 2008 Frequency Profiles

23

September and March 2008 in 5 mHz Bins


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


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


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


±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


Frequency Response Withdrawal


• 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


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


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


Frequency Response Initiative Implications


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


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


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


Questions?