MISO MOD-033-1 Model Validation

Update NERC System Analysis and Modeling

Subcommittee Meeting April 17, 2017

Presentation Outline

MISO’s journey for MOD-033-1

• Status update

MISO MOD-033-1 process document highlights

Results of model validation per MOD-033

• Suggestions on improving governor modeling in Eastern Interconnection (EI) for better frequency response

2

MISO MOD-033-1 Journey

MISO developed R1 process document in collaboration with NATF

• 30+ utilities from WECC, EI and ERCOT involved • MISO process document aligned with NATF document guidelines, sections customized

and details added • Discussed in MISO Modeling User Group (public). Will be finalized in May, 2017. Will

be posted on our website at: https://www.misoenergy.org/Planning/Models/Pages/MOD-033-1.aspx

MISO successfully performed a model validation using large plant trip event from Sep 13, 2015

• Performed steady state validation on a part of the system • Differences in frequency response found during dynamic model validation

3

MOD-033 R1 Time Line

• Steady State

Validation process

• Includes performing a test validation

Fall, 2016

• Dynamic Validation process

• Includes performing a test validation

• Developing R1.3

Late, 2016

• Finalize Process and write Document

• Audit Readiness

• Feedback Process

Early 2017

• Compliance Effective Date

July, 2017

4

MOD-033 Process Document Structure

Section-1

• Describes the process for performing R1.1 (power flow model validation)

Section-2

• Describes the process for performing R1.2 (dynamic model validation process)

Section-3

• Describes the guidelines MISO shall use to determine unacceptable differences in model performance per part R1.3

Section-4

• Describes process which MISO shall use to resolve the unacceptable differences to fulfill requirement R1.4

Section-5

• Describe the process of obtaining data from RC and TOP

5

.

Section 1: Data Sources

Planning model: starting point is Outage Coordination daily case

• Planning case, from Model On Demand (database) with all monthly projects applied • Has all known outages for MISO + external modeled • Load and Gen profiles from NERC SDX system, values much closer to State Estimator

model compared to monthly planning model (MOD) base case • Internally vetted through multiple reviews

Real Time data: S.E. case, PI historian, SCADA data, PMU data

Details on preparing model, mapping sources

• How load and gen values will be matched, sanity checks will be formed • EMS to planning mapping, MISO commercial node mapping for non-standard buses

6

Section 2: R1.2 Dynamic Model Validation

Dynamic model high level process

• Chose an event , prepare and validate power flow case, simulate the fault, compare the simulation

Data sources

• Planning data : power flow model->Outage Coordination daily base case • Dynamic model: latest MTEP dynamic package • Real time : PMU data, Pi historian data , MISO Frequency response scorecard

Details on preparing models, performing simulation and comparison • Sanity checks and model limitations which need to be considered will be

documented • Focus at system level response, e.g. monitoring frequency and voltage at 345 kV 7

Section 3: R1.3 guidelines to determine unacceptable differences • Inputs from NATF, experienced utilities, NERC • Engineering judgment must be exercised, guidelines to be applied with

discretion • Guidelines will evolve with time as more validations are performed

8

Excerpts from MISO document

Quantity Acceptable Differences Bus voltage magnitude ±2% (>346 kV)

±3% (200>kV>345kV) ±4% (100>kV>199kV)

Generating Bus voltage magnitude ±2% MVA Current flow ±10% or ±100 MVA Difference in % normal loading ±10% on branch normal continuous rating

In accordance with NERC MOD-033 application

guidelines, MISO will plot the simulation result on the same graph as the actual system response, and the two plots will be given a visual inspection to see if they look similar or not and will determine if the model performance is acceptable.

Steady State Guidelines

Dynamic model performance guidelines

Section 4 : R1.4 process to resolve unacceptable differences

9

Per R1.3 evaluations, MISO will contact data owner citing the issue

Data owner to review the information,

determine corrections

Data owner responds to MISO

MOD-033 Model Validation Results

• Event Details: – On the night of Sep 13, 2015, a 1100 MW plant tripped – Event produced a frequency dip good for model validation

• MISO prepared a base case for Sep 2013 – Used planning case of Sep 13, 2015 with outages applied – 6 pm S.E. case used to map gen and load values for select areas – Procedure laid out in process document followed

• Gen for disturbed areas matched on each unit level, Gen voltage schedules matched

• Load matched on area level, and mapped buses in S.E. and planning case • Transformer taps, shunt devices matched • Area interchange matched with first tier companies

10

Steady State Validation Results for Branch Flows

11

R² = 0.9843

-10.0

10.0

30.0

50.0

70.0

90.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0

S.E.

Flo

w (M

VA/R

atin

g)

Planning Flow (MVA/Rating)

Branch Flow Comparison Graph (170 Circuits)

All circuits pass the criteria

Steady State Validation Results for Bus Voltages

R² = 0.9481

0.50

0.60

0.70

0.80

0.90

1.00

1.10

1.20

0.99 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08

S.E.

Vol

tage

(pu)

Planning Voltage (pu)

Voltage Comparison for 230 and 220kV buses

230-220KV

All buses pass the criteria

12

59.8600

59.8800

59.9000

59.9200

59.9400

59.9600

59.9800

60.0000

60.0200

0.00

0.28

0.55

0.83

1.09

1.37

1.64

1.92

2.19

2.47

2.74

3.02

3.29

3.57

3.84

4.12

4.39

4.67

4.94

5.22

5.49

5.77

6.04

6.32

6.59

6.87

7.14

7.42

7.69

7.97

8.24

8.52

8.79

9.07

9.34

9.62

9.89

10.1

710

.44

10.7

210

.99

11.2

711

.54

11.8

212

.09

12.3

712

.65

12.9

213

.20

13.4

713

.75

14.0

214

.30

14.5

714

.85

15.1

215

40

Freq

uenc

y in

Hz

TIme in sec

Frequency with Current Models

PMU

13

System Frequency comparison on 345 kV Bus

frequency nadir differs

Settling frequency is optimistic.

Gen trip simulated at 0.5 sec

Conclusions: • Models are predicting system frequency settling at a higher value. • This difference is a known issue, due to governor modeling (dead band, non-responsive

governor). • What can be done?

What needs to be done for EI Governor Modeling Improvement ? Governor responsiveness and dead-band modeling should be improved in EI

• Pointed in many studies Eto 2010, NREL 2013, ORLN 2014

• Focus on top 3-4 type of models can be a good start • Entire EI needs to address issue

MISO performed model improvement based on ORLN 2014 study recommendation

• Focused on top 3 governor models • if unit has IEEEG1, TGOV1 or IEEGO type governor

model • If unit loaded >80% , then governor removed • else model converted to WSEIG1 with dead-band of

36 mHz • WSEIG1-WECC Modified IEEE Type 1 Speed-

Governing Model • Included additional dynamic data file from ERAG

MMWG for frequency response

Model performance compared against similar event in EI

• ERAG MMWG 2024 Summer Peak Model

0

50000

100000

150000

200000

250000

300000

350000

MVA

Governor Types in EI (MVA)

Turbine Governor

Sum of Mbase (MVA)

14

EI Frequency Response recorded by DFR (FNET)

~13 sec to Freq Nadir for 980 MW loss

~30mHz drop

15

59.93

59.94

59.95

59.96

59.97

59.98

59.99

60

0.01

6666

80.

5291

7039

41.

0416

7723

71.

5541

8407

92.

0500

2260

22.

5625

1478

23.

0750

0696

23.

5874

9914

24.

0999

9132

24.

6124

8350

15.

1249

7568

15.

6374

6786

16.

1499

6004

16.

6624

5222

17.

1749

4440

17.

6874

3658

18.

1999

2923

78.

7124

2141

79.

2249

1359

79.

7374

0577

710

.229

0649

410

.741

5571

211

.254

0493

11.7

6654

148

12.2

7903

366

12.7

9152

584

13.3

0401

802

13.8

1651

0214

.329

0023

814

.841

4945

615

.353

9867

415

.866

4789

216

.379

0569

316

.891

6664

117

.404

2758

917

.916

8853

818

.429

4948

618

.942

1043

419

.454

7138

219

.967

3233

Simulation Results for loss of generator (1100+ MW) with 24% governor (on MVA basis) modified in EI

~8 sec to System Freq Nadir

Remote Frequencies in EI; ~25mHz drop; Lazy “L” Shape

Overall performance is similar. Some difference expected as model has different dispatch and load levels as EI 2024 Summer Peak model was used.

Local Frequency swing

Seconds 16

What if improvements were done only in MISO?

Improvement in the entire EI is the key!

Local frequency swing match improved after EI wide correction

Gen trip simulated at 0.5 sec

Settling frequency is matched after 15 sec, EI wide modification

Marginal improvement if governors only in MISO are improved

17

Summary MISO on track for MOD-033-1 R1 effective date

MISO successfully performed test model validation

Some modeling issues identified

Governor modeling needs improvement

Entire EI needs to focus

Process document will be posted shortly (R1)

18

Contact Info • Nihal Mohan (nmohan@misoenergy.org)

19

EI Frequency ResponseChanging Resource Mix AssessmentScenario Analysis of Changing Resource Mix

Olushola J. Lutalo, MS, P.E., PMP, Senior Engineer of System AnalysisSystem Analysis and Modeling Subcommittee UpdateApril 20, 2017, Atlanta, GA

RELIABILITY | ACCOUNTABILITY2

• Additional CRM analysis was performed using the 2016 Series MMWG Cases to create four CRM Frequency Responsive (FR) 2021 LL Study Cases. Business-As-Usual (BAU) FR Case CPP Case with 9250 MW of NTRs Reduced Synchronous Generation Case Replace Synchronous Generation Cases with FR NTRs Case

• Case validation was performed for each CRM FR Study by evaluating the case initialization and the No-Disturbance simulations.

CRM Study Cases Development Using2016 Series MMWG 2021 LL Base Cases

RELIABILITY | ACCOUNTABILITY3

(Median) Millstone May 25, 2014 Event Comparison to 2016 Series CPP Case

RELIABILITY | ACCOUNTABILITY4

(Median) Millstone May 25, 2014 Event Comparison to 2016 Series CPP Case

RELIABILITY | ACCOUNTABILITY5

(Median) Calvert Cliff April 7, 2015 Event Comparison to 2016 Series CPP Case

RELIABILITY | ACCOUNTABILITY6

(Median) Calvert Cliff April 7, 2015 Event Comparison to 2016 Series CPP Case

RELIABILITY | ACCOUNTABILITY7

• 2016 Series CPP FR Case with 9250 MW of added NTRs initializes very well and has a valid No-Disturbance simulation.

• The BAU Case was successfully benchmarked against the Millstone May 25, 2014 and the Calvert Cliff April 7, 2015 Events.

• The plots demonstrate that the 2016 Series FR CPP Case with 9250 MW of added NTRs compares very well to the Millstone May 25, 2014 Event and the Calvert Cliff April 7, 2015 Event with realistic settling frequencies and no oscillations.

2016 Series CPP FR Case with 9250 MW of added NTRs Conclusions

RELIABILITY | ACCOUNTABILITY8

Reduce Synchronous Generation by 9250 MW in Add NTR Case

Reduced Synchronous Generation Case Assumptions• Beginning with the CPP Case containing 9250 MW of Non-FR

NTRs, reduce EI FR Synchronous Generation by 9250 MW by disabling Governors on 9250 MW of GGOV and HYGOV generation.

• The purpose of this case is to determine the Impact of Non FR NTRS on the EI when FR dispatched generation equal to the NTRs are made Non FR by bypassing the governors.

RELIABILITY | ACCOUNTABILITY9

(Median) Millstone May 25, 2014 Event 2016 Series Reduce FR Sync. Gen. Case

RELIABILITY | ACCOUNTABILITY10

(Median) Millstone May 25, 2014 Event 2016 Series Reduce FR Sync. Gen. Case

RELIABILITY | ACCOUNTABILITY11

(Median) Calvert Cliff April 7, 2015 Event 2016 Series Reduce FR Sync. Gen. Case

RELIABILITY | ACCOUNTABILITY12

(Median) Calvert Cliff April 7, 2015 Event 2016 Series Reduce FR Sync. Gen. Case

RELIABILITY | ACCOUNTABILITY13

• The 2016 Series Reduce Synchronous Generation Case initializes very well and has a valid No-Disturbance simulation.

• The plots demonstrate that the 2016 Series FR Series Reduce Synchronous Generation Case comparisons very well to the Millstone May 25, 2014 Event and the Calvert Cliff April 7, 2015 Event with realistic settling frequencies and no oscillations.

• The Reduced Synchronous Generation case is less frequency responsive for each comparison Event.

Reduce Synchronous Generation Case Conclusions

RELIABILITY | ACCOUNTABILITY14

Replace Synchronous Generation with 9250 MW of Non-FR NTRs Case Assumptions

Replace Synchronous Generation Case Assumptions• Beginning with the Reduced Synchronous Generation Case

containing 9250 MW of Non-FR NTRs, enable FR on the approximately 9000 MW of NTRs dispatched at 90% of PMAX.

• The purpose of this case is to determine the Impact of FR NTRS on the EI when FR NTRS are dispatched to replace generation made Non FR by bypassing the governors.

RELIABILITY | ACCOUNTABILITY15

Overview of ColumbiaGrid NERC MOD 33 Process

Bo Gong

NERC SAMS meeting, Atlanta, GA, April 20, 2017

1

ColumbiaGrid – Grid Planning

2

� Predominantly hydro generation¡ ~36,000 MW generation capacity, 70%

are hydro power, Columbia river basin contributes to ~40% total hydro capacity national wide.

¡ 7,900 MW combined capacity for COI and PDCI to export power to California

¡ 2,200 MW capacity to import power from Montana

¡ 3,000 MW capacity import/export to Canada

¡ RAS were implemented includes generation run back or tripping.

¡ BPA PMU and PDC installed system wide, with Oscillation Detection Monitor to provide real-time monitoring of stability issues.

Pacific NW Electric Power System

3

� NERC MOD-033-1 will become effective in 7/1/2017. Each PC will have 24 months to perform Power Flow and Dynamic validation.

� ColumbiaGrid developed a process document, includes a guideline for unacceptable discrepancy, and a guideline to resolve unacceptable discrepancy

� Evidence of model validation will be retained by ColumbiaGrid for audit purpose

Overview of NERC MOD-033 Requirement

4

� Peak RC and WECC will develop several event scenario each year that can be used for MOD-033 purpose¡ Peak RC will prepare a system snapshot in west-side system model

(WSM) format (node breaker) and perform some preliminary validation¡ WECC will convert the WSM model into planning model and provide the

related dynamic data¡ Event sequence and recorded measurement can also be requested from

Peak RC

Peak RC and WECC

5

� CG MOD 33 will focus on major system events and main grid validation. ¡ Members can perform their own validation of local events on local

system

� A MOD 33 work group will be formed for selected events¡ Utilities will report in CG planning meeting for any significant events

that can be used for MOD 33 validation.¡ CG will call for participation after an event has been selected for MOD 33

purpose by at least one member. ¡ Once an event has been selected, a workgroup will be formed to hold

regular meetings to work on the MOD 33 validation for this event.

Event Selection

6

� MOD 33 will be primarily performed for event scenario. Power Flow and Dynamic validation can be performed using the same event together.

� On top of the base cases developed for MOD 33 events, CG will facilitate to develop certain peak hour cases (heavy summer or heavy winter) for power flow validation purpose (no event necessary).

Base case preparation

7

� Challenges come from:¡ 1. How good we understand the system wide dynamic modeling vs real-

time performance ¡ 2. How accurate and precise we can use some metrics to describe a

dynamic process

� We believe a guideline is aim to translate relatively complicated dynamic behavior into some relatively simple metrics that can sufficiently capture the similarity and distinguish the difference.¡ Sufficient to recognize the similarity¡ Allow reasonable variation

Unacceptable Discrepancy of Dynamic Comparison

8

Some Existing Criterions

9

� WECC: Model Validation Report for May 16, 2014 RAS Event

Example 2014/5/16 Event

10

59.7

59.75

59.8

59.85

59.9

59.95

60

60.05

0 50 100 150 200

MALIN_500KV Frequency

-0.3500%

-0.3000%

-0.2500%

-0.2000%

-0.1500%

-0.1000%

-0.0500%

0.0000%

0.0500%

0.1000%

-20 0 20 40 60 80 100 120

MALIN_500_FREQ Errors

Example 2014/5/16 Event

11

1.05

1.06

1.07

1.08

1.09

1.1

1.11

0 50 100 150 200

BIG EDDY 500 VMAG

-3.0000%

-2.5000%

-2.0000%

-1.5000%

-1.0000%

-0.5000%

0.0000%

0.5000%

1.0000%

1.5000%

2.0000%Voltage Errors

Example 2014/5/16 Event

12

-12000

-10000

-8000

-6000

-4000

-2000

0

2000

0 50 100 150 200

PMU Voltage Angle Simulated Voltage Angle

-200

-100

0

100

200

300

400

500

116

332

548

764

981

197

311

3512

9714

5916

2117

8319

4521

0722

6924

3125

9327

5529

1730

7932

4134

0335

6537

2738

8940

5142

1343

7545

3746

99

Angle Error %

Example 2014/5/16 Event

13

-60

-40

-20

0

20

40

60

80

100

0 50 100 150 200

-20000.0000%

-15000.0000%

-10000.0000%

-5000.0000%

0.0000%

5000.0000%

10000.0000%

0 20 40 60 80 100 120

Line Flow Error

� A metrics for dynamics difference is no longer a steady state one¡ A band all along the trajectory, at final time, or after oscillation damped out?¡ Can we allow exception at the time of switching?

� Percentage vs Absolute metrics?¡ Percentage value are depends on the base value, what if the base value

may become 0, or come close to 0?¡ Using an absolute error, on the other hand, may fail to disclose similarity

and neglecting the severity of the disturbance

Problem of Using Certain Metrics

14-2500

-2000

-1500

-1000

-500

0

500

0 50 100 150 200

-500.00

0.00

500.00

1000.00

1500.00

2000.00

2500.00

0 20 40 60 80 100 120

Line Flow MW Error

� Type of Events ¡ e.g., for a voltage event, frequency may be a less important metrics and

vice versa

� Dominant Trends to Capture¡ e.g., In a longer term frequency event, post fault generator power

ramping rate may be more important than their absolute values

� Metrics related to certain Dynamic Behaviors¡ e.g., with oscillations, frequency and magnitude can be more important

than point wise difference along trajectories.

Criterion relation to Events and Variables

15-200

-150

-100

-50

0

50

100

150

200

0 50 100 150 200

-60

-40

-20

0

20

40

60

80

100

0 50 100 150 200

� NERC MOD-033 allowed the guideline for dynamic comparison can be less precise. ColumbiaGrid will primarily adopt visual inspection to compare the real-time measurement and simulation.

� We also acknowledge that, during the process of gaining more understanding on how dynamic behavior can be evaluated using various metrics under different situations, we may adopt some metrics in the future.

CG Guideline on Unacceptable Discrepancy

16

Question:

Bo Gong, gong@columbiagrid.org

17

Essential Reliability Services Working GroupThomas ColemanSAMS MeetingApril 19, 2017

RELIABILITY | ACCOUNTABILITY2

• “Building blocks” of physical capabilities • Stressed by resource changes• Not all MWs are equal• Some partly covered through ancillary services• Accommodate local/regional needs

Essential Reliability Services (ERS) Fundamentals

Resource Adequacy

Essential Reliability Services

Reliability

Reliability Assessments URL

RELIABILITY | ACCOUNTABILITY3

Load & Resource Balance (Ramping & Balancing)o Track and project the maximum one-hour and three-hour

ramps for each balancing area

Voltage (V) Supporto Track and project the static and dynamic reactive power

reserve capabilities to regulate V at points in the systemo Review the short circuit current at each transmission bus

in the network, and calculate short circuit ratios

Frequency Support (restoring after major unit loss)o Track minimum frequency & its response post N-1 evento Track & project level of conventional synchronous inertiao Track & project the initial frequency deviation in the 1st

1/2 second following the largest N-1 event

Essential Reliability Services (ERS) Some Measures

ERSTF Final Report - URL

RELIABILITY | ACCOUNTABILITY4

• 2016 Deliverables

Perform process development for approved ERS Measures

A whitepaper on methodology for ERS Measures Sufficiency Guideline

• 2017 Deliverables

Develop papers on ERS Measures data collection and reporting

Length of each brief paper is anticipated to be 3 to 5 pages

Goal is to clearly document how each respective measure’s data is used:

o Historical reporting → State of Reliability report, NERC PAS reports

o Future reporting → Long Term Reliability Assessment report

ERSWG 2016 and 2017 Deliverables

RELIABILITY | ACCOUNTABILITY5

How do we plan for a loss of a major resource in order to equip the system with sufficient frequency (f) response?

• Answer: Track frequency w.r.t. under frequency load shedding criteria Post a contingency event track the min f and the f response Post the largest contingency event project the initial f and the f deviation

• M1 : Interconnection Level Synchronous Inertial Response (SIR)

• M2 : Initial Frequency Deviation Following Largest Contingency

• M3 : Balancing Authority Level Synchronous Inertial Response (SIR)

• M4 : Frequency Response at Interconnection Level (a measure set)

Frequency Support Measures (M1 - M4)

MEASURES

RELIABILITY | ACCOUNTABILITY6

• Propose to use future cases as created by the Mod-032 designee ERAG, HQ, WECC, ERCOT planning cases

• Future looking M1 - Select for each interconnection case:

3 Future light loading cases (Anticipate 1 year, 3 year, and 5 year)o Assumption: light loading will provide a min SIR

Apply dispatch in the selected case (∑H in existing case)o Off the shelf dispatch is the planning area’s economic dispatcho Case to contain high level of renewables thus min SIR conditions

Goal: Calculate min SIR and trend future year cases for LTRAo No comparison of Historic H and Future H values (uncoupled)

M1 Interconnection Level SIR: Forward Looking Approach

RELIABILITY | ACCOUNTABILITY7

• Currently No Historic BA Trended Data Available (RS is now collecting)

• Propose to use future cases as created by the Mod-032 designee ERAG, HQ, WECC, ERCOT planning cases

• M3: will follow all M1 Interconnection Assumptions:

Approach: Use Assessment Areas in M1 cases to report H by area

o Not a 1:1 match with historic M3 but able to trend

Apply dispatch in the selected case (∑H in existing case by area)

Goal: Calculate min SIR and trend future year cases for LTRAo No comparison of Historic H and Future H values (uncoupled)

M3 Balancing Authority Level SIR: Forward Looking Approach

RELIABILITY | ACCOUNTABILITY8

• If f point Nadir (C or C’) > highest UFLS set point than: Primary Frequency Response (PFR) sufficiently arrested and stabilized f

• If f point Nadir (C or C’) ≤ highest UFLS set point than: firm loads are dropped as a precaution to arrest f decline

Measure 4: Interconnection Level Frequency Response Measure Set

RELIABILITY | ACCOUNTABILITY9

• Select specific contingencies from M4 historical event data • Pick a base case that results in least amount of SIR A light load case, scenario cases with resource mix varied , M1 base cases

ERS M4 Future Looking Approach

ERS M4 Calculates 7 measures

A:B PFRA:C load damping &

initial gov. responseC:B gov. responseC’:C min f gov. w/dr to NadirtC-t0 Δt in f Nadir & initial eventtC’-tC Δt gov. w/dr min & initial f NadirtC’-t0 Δt gov. w/dr min & initial event

Abbreviations : Governor (gov.), withdrawal (w/dr)

Primary Frequency Response (PFR)

RELIABILITY | ACCOUNTABILITY10

Table 1: Resource Contingency Criteria (RCC) for each Interconnection

ERCOT EI WECC HQ

2750 MW 4500 MW 2740 MW 1700 MW

Measure 2’s Interconnection Level :Rate of Change of Frequency (RoCoF)

No Load Damping 𝑹𝑹𝑹𝑹𝑹𝑹𝑹𝑹𝑹𝑹 = ∆𝑷𝑷𝑴𝑴𝑴𝑴(𝟐𝟐∗ 𝑲𝑲𝑲𝑲𝒎𝒎𝒎𝒎𝒎𝒎−𝑲𝑲𝑲𝑲𝑹𝑹𝑹𝑹𝑹𝑹 )

∗ 𝟔𝟔𝟔𝟔 [𝑯𝑯𝑯𝑯/s]

No Load Damping Calculations because estimating future RoCoF

Apply resource mix scenarios to cases to observe RoCoF changes

• Calculate RoCoF for ERS M4 future events with Historic M2 method: At minimum SIR conditions for ERS M4 events, Apply Interconnection RCC MW loss, For 0.5 second time frame calculate RoCoF:

RELIABILITY | ACCOUNTABILITY11

• Intended to provided insight on reactive strength of the system

• Reactive capability calculated and tracked at BA sub-area level At Peak, Shoulder, and light load levels

1. Static and dynamic reactive capability / total MW load2. Static and dynamic load power factor for distribution at the low

side of transmission buses

• Data collected by System Analysis and Modeling Subcommittee

• SAMS results recommendation is to discontinue M7 and use: ERSWG White paper Sufficiency Guidelines (BA sub-areas) Reliability Guideline for Reactive Power Planning

M7: Reactive Capability of the System

RELIABILITY | ACCOUNTABILITY12

PJM©2017

Steady State and Dynamic Model Validation

Wenzheng QiuByoungkon Choi

PJM©20172

MOD-033-1 Highlights

• Purpose: To establish consistent validation of steady state and dynamic system models

• Effective date: July 1, 2017• Each Planning Coordinator shall implement a documented data validation

process that includes the following attributes:• Comparison between planning power flow model with real time date sources

through simulation at least every 24 calendar months• Comparison between planning dynamic model with actual system response

through simulation of a dynamic local event at least once every 24 calendarmonths

• Guidelines to determine unacceptable difference in performance• Guidelines to resolve the unacceptable difference in performance

PJM©20173

Outline

www.pjm.com

• Steady State Model Validation– Process– Challenges– Results

• Dynamic Model Validation– Process– PMU Data Resource– Tool Discussion– Results

PJM©20174

Steady State Model Validation Process

EMS PSS/E format Snapshot (summer peak)

RTEP Planning model

Replace Planning External model with EMS outside model

Bus Mapping

Confirm with TO /Feedback

to EMS

Update EMS model

EMS model correction

Planning modeling correction

Project file for Model On Demand

Adjust Planning case to match EMS snapshot

Compare flow and voltage for major

branches and buses

Root analysis Yes

All differences are

acceptable?No End

PJM©20175

Bus mapping

www.pjm.com

• Totally different bus # in EMS snap shot case and planning case

• Related bus name, but bus names are not unique and some bus names show no clue

• Not one to one mapping

KV level EMS Planning

765 31 31

500 132 122

345 296 322

230 909 1219

161 41 48

138 3460 3576

115 862 899

Number of Gens EMS Number of Gens Planning

1706 1429

PJM©20176

Branch mapping

www.pjm.com

• Mapping branches between two cases with the bus mapping information

• Direction matters• Manually cleaning to remove the zero impedance

branches and generator/SVC/shunts leads• Compare branch flows for mapped branches

PJM©20177

Shall We Compare?

www.pjm.com

PJM©20178

Planning Case Adjustment

• All significant outages in EMS case• Generators are dispatched with the same output and voltage schedule as

in EMS case.• Loads in each PJM LDA are scaled to the same level as in the EMS case.• HVDC, VFT and PARs are set with the same setting points as in EMS

case.• Line reactors, bus shunts, switching shunts and other reactive device are

set matching with EMS status and settings.

PJM©20179

Initial Comparison Results-1

There are 161 500kV and above lines are compared and 9 of these have more than 100MW flow difference.

PJM©201710

Initial Comparison Results-2

There are 345kV lines are compared and 35 of these have more than 50MW flow difference.

PJM©201711

Initial Comparison Results-3

There are 985 230kV lines are compared and 93 of these have more than 50MW flow difference.

PJM©201712

Reconciliation Process—iterative process

• Missed local transmission outages • Local generator dispatch • Open/close bus tie during operations• Line impedance/Topology difference• Local loads

PJM©201713

Final Comparison Results

www.pjm.com

PJM©201714

Dynamic Model Validation Process

Select disturbance

scenario

Prepare models, run simulations

Evaluate model

performance

Resolve identified

issues

Area level model validation

• Validate area-wide model performance

• Full or larger planning model than plant level model

• Could be time consuming in model preparation

Plant/Component levelmodel validation

• PMU/DFR data should be available at the plant/component

• Reduced network model• Less time consuming

PJM©201715

PJM’s Implementation Approach

PJM©201716

PMU Locations in PJM footprint

• Around 400 PMUs are installed at 106 stations.

• Each station may have different number of PMUs depending on the topology (lines, generators, etc.).

• Normally PMUs are required to send voltage, current phasor data, frequency.

PJM©201717

PMU Installed Generators

• Currently total 27 existing generators have PMUs installed at the high side of GSU.

• PMUs are required at new generators (100MW or larger) –PJM Manual 14D.

• Several generators' dynamic models have been validated using multiple event data.

PJM©201718

PMU Database and ePDC

Database Type: MS SQL Server

Database Name: RTDMS2012

Host Name: SQL36DWP

Port: 1433

Long-term archive (3 years)

Database Type: MS SQL Server

Database Name: RTDMS2012

Host Name: SQL39VWP\SQL39VWPPort: 51433

Short-term archive (90 days)

PJM PDC

PI RTDMS

ePDC User

PJM Internal Network

Calculate MW, MVAr locally

Calculate MW, MVArlocally

PJM©201719

Dynamic Model Validation – Simulation Tools

www.pjm.com

Three tools are utilized:

• PPPD* developed by EPRI

• PSS/E Playback function

• PJM developed tool (using phase shifter method and PSS/E dynamic simulation)

*PPPD: Power Plant Parameter Derivation

PJM©201720

PPPD

• Generator Model• Exciter Model• Governor Model• PSS Model

HVLV

PPPD requires measured data at LV side

PJM©201721

PJM Generator Model Validation Tool

PJM©201722

Playback-Phase Shifter Method

G G

GExternalSystem

Study PlantGSU

LVHV

PMU

BoundaryBus

Study PlantGSU

206Classical

generator with large inertia

(e.g. 10,000MVA)

IdealPhase Shifter 205

P,Q203

Full System

Reduced System

V,I,f, P,Q

Measured voltage and angle are played back at bus 205 and P,Q are calculated by PSS/E

PJM©201723

Playback-PSS/E 'PLBVFU1' model

G G

GExternalSystem

Study PlantGSU

LVHV

PMU

BoundaryBus

Study PlantGSU

206Classical

generator with large inertia & zero Zsource

ZeroImpedance Line

205P,Q

203

Full System

Reduced System

V,I,f, P,Q

‘PLBVFU1’ model plays back measured voltage and frequency signals to bus 203.

PJM©201724

Dynamic Model Validation - Examples

• Voltage comparison

0 1 2 3 4 5 6 70.94

0.95

0.96

0.97

0.98

0.99

1

1.01

1.02

1.03

Time (seconds)

Vol

tage

(pu)

MeasuredOptimizedOriginal

Numerical issue with 30Hz sampling rate

Resampling(150Hz)

PPPD

PJM©201725

Dynamic Model Validation - Examples

Phase Shifter Method PSS/E Playback

• Real power comparison

PJM©201726

Dynamic Model Validation - Examples

Phase Shifter Method PSS/E Playback

• Reactive power comparison

PJM©201727

Dynamic Model Validation - Examples

PSS/E Playback

Line Trip Event 1 Line Trip Event 2

PJM©201728

Dynamic Model Validation - Observations

• Several dynamic performances can be reviewed including oscillation magnitude and frequency, damping ratio, overshoot, recovery time and steady state, etc.

• Engineering judgement and domain knowledge are crucial to validation

• Collaboration with generation owners who have more information and experiences about plant dynamic models can be very useful.

• MOD-026/027/032 process are useful tools to communicate with GOs.

PJM©201729

2017 IEEE PES General Meeting Paper

www.pjm.com

PJM©201730www.pjm.com

Thanks!

CONFIDENTIAL – Limited Distribution

NATF Modeling Practices GroupUpdate

April 19, 2017NERC SAMS Meeting

Ed Ernst- NATF Program Manager

CONFIDENTIAL – Limited Distribution (NERC)Copyright © 2017 North American Transmission Forum. Not for sale or commercial use. Limited Distribution documents are confidential and proprietary. Limited Distribution documents may be used by employees of North American Transmission Forum (“NATF”) member companies who have a need to know the information in the document, by NATF staff, and by entities who have permission to receive Limited Distribution documents pursuant to a written agreement with the NATF, for purposes consistent with the NATF’s mission. All rights reserved.

CONFIDENTIAL – Limited Distribution

Outline

• NATF Practices Groups• NATF Modeling Practices Group Activities• Coordination between NATF and NERC

2

CONFIDENTIAL – Limited Distribution

NATF Practice Groups• Compliance• Human Performance• Modeling• Operator Training• Security• System Operations• System Protection• Transmission-Nuclear Interface• Vegetation Management• Equipment Performance and Maintenance

3

CONFIDENTIAL – Limited Distribution

NATF Modeling Practices Group Current Activities

• Recent Public Posting of NATF Documents• On-going monthly work of Modeling Practices

Group and its various working groups• June 20-21, 2017 NATF-NERC Modeling

Workshop

4

CONFIDENTIAL – Limited Distribution

Public Posting of NATF Modeling and Planning documents

• MOD-033-1 Reference Document • NATF CIP-014-2 Reference Document

– Both documents submitted to NERC for Compliance Implementation Guidance approval

• Located at www.natf.net/documents

5

CONFIDENTIAL – Limited Distribution

On-going work of Modeling Practices Group and its various working groups• Dynamic Load Modeling Working Group

– Sharing experiences– Following work of other groups: NERC Load Modeling Task Force, etc.– No documents under development

• Transmission Planning Working Group– Sharing experiences on TPL-001-4, TPL-007, MOD-033 model validation and transmission/sub-

transmission connected renewables– Following work of other groups: NERC GMD Task Force, etc.– No documents under development

• Distributed Energy Resources Working Group– Sharing experiences – Following work of other groups: NERC Distributed Energy Resources Task Force, etc.– Awaiting NERC Distributed Energy Resources Document – Plan to develop a Distributed Energy Resources Reference Document during 2017

• EMS Modeling Working Group– Current focus is on the building of external models– Working on EMS External Model Reference Document – target completion mid 2017

CONFIDENTIAL – Limited Distribution

June 20-21 Modeling Practices Group Workshop• Where: Exelon/ComEd in Oak Brook, Illinois (Chicago)• Co-hosted by NATF and NERC• Planned topics

– Dynamic Load Modeling– Power Plant Modeling – MOD-033 – Integrating Renewables at the Transmission Level – Modeling DER (renewables at the distribution level)– General Session: NERC modeling updates, Emerging modeling issues

• We are at workshop attendance cap. – If you are interested in attending and have not registered, please contact Ed

Ernst at eernst@natf.net– If you have registered and will not be attending, please contact Ed Ernst at

eernst@naf.net so we can register persons on the waiting list

John Pearson)

7

CONFIDENTIAL – Limited Distribution

Coordination between NATF and NERC

• Document development• Jointly Sponsored June 20-21 Modeling

Workshop hosted by Exelon(Com Ed) in Chicago • Regular NATF-NERC meetings at Gerry

Cauley/Tom Galloway level to coordinate efforts• Ryan Quint of NERC staff has standing slot on

Monthly NATF MPG and its working group calls to cover topics as needed

• Ed Ernst has standing slots on SAMS and MWG calls to cover topics as needed

8

CONFIDENTIAL – Limited Distribution

NATF Modeling Practices Group Update

Questions?

9

Coordinated Substation Topology

SAMS April 2017

1

Agenda• History & Methodology

• Topology Strategy

• TREND Subs Overview

• Timeline

• Challenges & Questions

2CONFIDENTIAL – Transmission Information – This Data Should Not Be Shared With The Merchant Function

History• Historical focus on line-centric facility ratings‒ Focused on “most limiting element” – sometimes terminal equipment‒ All operational and planning datasets reflect this focus

• Separate, independently maintained sources of duplicate data‒ Differing focus and perspective can lead data-drift‒ Results in a EMS real-time model with multiple varying data sources

• “Annual Facility Ratings True-up”‒ Manual labor intensive process‒ Consumes significant resources in Transmission Planning,

Transmission Operations, and EMS

Existing TO Data

Line ElementsDB

Line ElementsDB

Line Elements Spreadsheet

LineElementsSpreadsheet

Task #1:Determine Common Equipment

Data Structure

CommonDeviceData

Existing TO Data

CommonDeviceData

Company-Specific Parser

Company-Specific Parser

Company-Specific Parser

Company-Specific Parser

Line ElementsDB

Line ElementsDB

Line Elements Spreadsheet

LineElementsSpreadsheet

Task #2:Parse existing data into new

common format.

Existing TO Data

CommonDeviceData

Line ElementsDB

Substation Topology

Node-BreakerModel

Task #3:Incorporate substation

topology to create a fully-functional, multi-

purpose model

Existing TO Data

CommonDeviceData

Line ElementsDB

Node-BreakerModel

TopologicalSubstation

Model

TREND SubsSubstationTopology

Data

New TO Data

InternalSubstation

Devices

21

3

Device Importer

PSSE 34 Case

Topology Strategy

8

Current Bus-Branch Model

No visibility of station

configuration … could be straight bus, ring bus, etc

9

Enhanced Node-Breaker Model

Total Equipment

Nodes

3 4

512

PrimaryComponent

TREND Subs Overview

12

TREND Subs Overview

Navigation

Device Section

EquipmentSection

Total Equipment

Nodes

12

3 4

5

PrimaryComponent

Nodes

12

3 4

5

Nodes

12

3 4

5

NERC & SOCO Timelines

17

NERC Node-Breaker TimelineQ4 ‘15 Q4 ‘16 Q4 ‘17 Q4 ‘18 Q4 ‘19 Q4 ‘20

NERC

SOCO

Common Data Structure Design and Processing

Data Structure Initial Software Development

Construct Node-Breaker model from Operations model

TransitionSmall-scale Pilot

Planning Tool Development

PSSeTesting

Operations Tool Development

Data Gathering, Topology Creation, Data Mapping, Data Entry

Cross-Functional Model Sync &

Reporting

Model Intelligence Development

Bus-Branch to Node-Breaker Comparison

Line-Centric Data Retirement

To-Date Detail

Selective Node-Breaker Case

Creation

Q2 ‘16 Q2 ‘17 Q2 ‘18 Q2 ‘19 Q2 ‘20

18

SOCO Detailed Node-Breaker TimelineQ3 ‘15 Q4 ‘15 Q1 ‘16 Q2 ‘16 Q3 ‘16 Q4 ‘16 Q1 ‘17

DEVELOPMENT

DATA GATHERING & INPUT

DESIGN

Determine data scope Parser for existing data

Substation Data Schema

TREND Subs UI

PSSE 34 Nodal Raw Output

TREND Subs Data Importer

230kV+ Topology

Gathering total substation device data

Topological mapping of existing dataEntry of additional device data

OpsDB data structure, cross-functional mapping

TREND Subs Usability and feature improvements

TREND Subs -> OpsDBDataloader

Model Interpolator

19

Challenges and Questions

CONFIDENTIAL – Transmission Information – This Data Should Not Be Shared With The Merchant Function

• Bus-Load Modelo Expanding load model to fit node-breaker topology

• Operations & Planning Data Structure Mappingo Design in conjunction, maintain flexibility

o Model intelligence performance

• Industry Model Exchangeo Incorporating node-breaker into larger models

• PSSE Analysis Automation Tools

The 2nd Generation RES Models

RES Model Combination

Type 1 WTG wt1g, wt1t, wt1p_b

Type 2 WTG wt2g, wt2e, wt2t, wt1p_b

Type 3 WTG regc_a, reec_a, repc_a, wtgt_a, wtgar_a, wtgpt_a, wtgtrq_a

Type 4 WTG regc_a, reec_a, repc_a (optional: wtgt_a)

PV plant regc_a, reec_b (or reec_a), repc_a

BESS regc_a, reec_c (optional: repc_a)

1

Three-Phase Modeling of

Transmission Networks

Reynaldo Ramos, Ph.D., P.E.

Principal Engineer

SCS Transmission Planning

Wayne Dias, BSc. EE, MBA.

U.S. Product Line Manager

PSS®SINCAL – Siemens PTI

RELIABILITY | ACCOUNTABILITY 2

Outline

• Imbalance Background

• On The Need To Develop Three Phase Models

• Three Phase Modeling Summary

• Readily Available Software Package

Siemens Network Calculation (PSS®SINCAL) Tool and Modules

• Three-Phase Power Flow Analysis Using PSS®SINCAL

Modeling of Network Elements - Low-Medium-High Voltage

• Importance of Three-Phase Modeling

DER Integration

• T&D Demo – T&D Large Interconnected Networks

• Q&A

RELIABILITY | ACCOUNTABILITY 3

Imbalance Background

• In a perfectly balanced three-phase power system, voltages and currents are balanced and sum to zero; however, in a typical bulk-power system imbalance is common

• The primary cause of system imbalance is power flow through an unbalanced impedance network (e.g. untransposed transmission lines)

• In a typical transmission line configuration, e.g., flat spaced T-line, the distances between phases are different, and; thus, so are the mutual impedances

RELIABILITY | ACCOUNTABILITY 4

Imbalance Background Cont.

• When balanced voltages are applied to an unbalanced impedance network, the line currents no longer sum to zero

• Both I2 and I0 currents are created

• Imbalance can lead to:

Excessive negative sequence currents flowing into the generators (limits* are established to maintain the integrity of the generator)

Operation of ground overcurrent relays (zero sequence currents add together and mimic flow of ground currents)

* IEEE Std. C50.13 and C50.12

RELIABILITY | ACCOUNTABILITY 5

Imbalance Background Cont.

• Imbalance is further exacerbated when multiple transmission lines share the same right of way (ROW) and during N-1, N-2 conditions

• Example highlights the value of studying the effect of unbalanced current flows

RELIABILITY | ACCOUNTABILITY 6

On The Need To Develop 3-Ph Models

• An Unbalanced Power Flow (UPF) study of the system showed that the imbalance resulting from the lack of line transposition could cause an open phase alarm at Sub-B and a ground overcurrent relay operation at Sub-A

• T-line impedance matrices showed a significant amount of zero-sequence and negative-sequence unbalance

RELIABILITY | ACCOUNTABILITY 7

• The UPF analysis also revealed that with the addition of future generation in the area, the resulting 3I0 flows would double, and thus, could potentially exceed the pick-up level of ground overcurrent relays located at Sub-A and Sub-B

On The Need To Develop 3-Ph Models Cont.

RELIABILITY | ACCOUNTABILITY 8

• In this particular example the transmission line imbalance also causes a considerable amount of negative-sequence current

• Therefore, the effects of negative-sequence currents on nearby synchronous generators were analyzed

On The Need To Develop 3-Ph Models Cont.

• Analysis showed that nearby generators had sufficient capability

RELIABILITY | ACCOUNTABILITY 9

Three-Phase Modeling Summary

• Unbalanced power flow studies can reveal system issues that cannot be identified with normal power flow studies (positive sequence based models)

• Electromagnetic imbalance can lead to ground overcurrent relay misoperations (3I0); and cause I2 currents to flow in generators

• Whenever situations occur where two or more untransposed T-lines are mutually coupled particular attention should be given to the effects of these voltage and current imbalances, mainly with respect to:

Protective relay settings, and

Thermal capability of synchronous generators

RELIABILITY | ACCOUNTABILITY 10

Three-Phase Modeling Summary Cont.

• Incorporate unbalanced power flow studies into the Planning process to help avoid the need for post-event analysis

i.e., move this type of analysis from post-event analysis domain to the planning domain

• The potential impacts of distributed generation resources on unbalanced transmission networks can be analyzed

• Need robust software tools (handle large three-phase models)

RELIABILITY | ACCOUNTABILITY 11

RELIABILITY | ACCOUNTABILITY 12

What is PSS®SINCAL?

RELIABILITY | ACCOUNTABILITY 13

Highlights - PSS®SINCAL

• PSS®SINCAL is a “one stop shop” application with a customizable solution

• PSS®SINCAL facilitates the modeling and analysis of all types of electrical networks – balanced as well as unbalanced – ranging from high-voltage to low voltage

• PSS®SINCAL compliments PSS®E’s strong transmission planning capability broadening into an extensive integrated transmission and distribution (T&D) world where users can study the combined impacts on the low-medium-high voltage as a single network

RELIABILITY | ACCOUNTABILITY 14

• PSS®SINCAL is equipped with numerous interfaces to allow simple integration

Geographical Information Systems (GIS)

Supervisory Control and Data Acquisition (SCADA)

Enterprise Resource Planning (ERP)

Meter Data Management Systems (MDMS)

• The tool comes with a complete set of advanced algorithms, including economic and strategic planning, and dynamics

• The tool has an interactive visualization of network models in schematic, geographic, or multilayer diagrams

Highlights - PSS®SINCAL

RELIABILITY | ACCOUNTABILITY 15

• In addition, the tool combines planning and analysis for electrical as well as gas, water, and district heating/cooling networks, making it a perfect tool for handling future challenges

• Programming can be done extensively and with standard scripting languages such as VBA, VBS, C++, .net, Python and Java.

No special programming language has to be learned

• PSS®SINCAL also provides special simulation applications for protection device management and dynamic network calculation

Highlights - PSS®SINCAL

RELIABILITY | ACCOUNTABILITY 16

Real Time Applications - PSS®SINCAL

• PSS®PDMS

Protection Device Management System that allows for the storage and management of protection data such as settings, documents, and files

It also enables users to connect protection data from parameterization software (e.g., DIGSI®) and PSS®SINCAL’s protection simulations

• PSS®NETOMAC

Application optimized for dynamic network calculations

It provides real-time capability and interactive diagrams as well as a structured code and model management system – developed for dynamic simulation

RELIABILITY | ACCOUNTABILITY 17

RELIABILITY | ACCOUNTABILITY 18

Large Interconnected 3 Phase Network

• Importance of 3 phase modeling of the power system

• Unbalanced power flow analysis

• Large interconnected power network – low-medium-high voltage

• T&D Integration

Demo software's capability to handle balanced as well as unbalanced networks – ranging from high-voltage to low voltage.

RELIABILITY | ACCOUNTABILITY 19

Advantages at a Glance

User-Friendly Software…

…with high level performance…

…and a wide range of interfaces…

…for accurate and reliable technical and economical results…

…to provide users with measurable benefits.

RELIABILITY | ACCOUNTABILITY 20