pnm preliminary reliability analysis

1

PNM Preliminary Reliability Analysis

Astrape Consulting

Nick Wintermantel

4/18/2017

Preliminary Draft Results

2

Topics

Resource Adequacy Overview

SERVM Model Overview

Traditional Reserve Margin Target

Flexibility Metrics

RPS Scenarios and Rules of Thumb


3

Resource Adequacy Overview


4

Resource Adequacy

Resource Adequacy Definition: The ability of supply-side and demand-side resources

to meet the aggregate electrical demand (NERC Definition)

Resource Adequacy Studies

Reserve Margin Study

Goal: Calculate generating capacity deficiencies and determine the amount of capacity needed to

maintain resource adequacy during peak conditions

Purpose: Input or validation of expansion planning processes

Flexibility Study

Goal: Determine reliability deficiencies including both firm load shed events and renewable resource

curtailment due to system ramping/startup constraints (not capacity deficiencies)

Purpose: Provides assistance in setting appropriate parameters for resource additions and to

determine system operating reserve requirements


5

Resource Adequacy Metrics

Loss of Load Expectation (LOLECAP): Expected number of firm load shed events in a

given year due to capacity shortfalls

Loss of Load Expectation (LOLEFLEX): Expected number of firm load shed events in a

given year due to not having enough ramping capability

Loss of Load Hours (LOLHCAP): Expected number of hours of firm load shed in a given

year due to capacity shortfalls

Loss of Load Hours (LOLHFLEX): Expected number of hours of firm load shed in a given

year due to not having enough ramping capability

Expected Unserved Energy (EUECAP): Expected amount of firm load shed in MWh for a

given year due to capacity shortfalls

Expected Unserved Energy (EUEFLEX): Expected amount of firm load shed in MWh for a

given year due to not having enough ramping capability


6

Traditional "Generic Capacity" Metrics New "Flexible Capacity" Metrics

Definitions of Existing and New Reliability Metrics

LOLEGENERIC-CAPACITY

Traditional metric to capture events that occur due to

capacity shortfalls in peak conditions

LOLEMULTI-HOUR

New metric to capture events due to system ramping

deficiencies of longer than one hour in duration

LOLEINTRA-HOUR


deficiencies inside a single hour

20,000

30,000

40,000

50,000

1 5 9 13 17 21

41,500

42,500

43,500

44,500

45,500

10:00 10:10 10:20 10:30 10:40 10:50 11:00

0

10,000

20,000

30,000

40,000

50,000

60,000

1 3 5 7 9 11 13 15 17 19 21 23

MW

Hours

Load Generation

LOLECap LOLEFLEX


7

SERVM Model Overview


8

Strategic Energy Risk Valuation Model (SERVM)

SERVM has over 30 years of use and development

Probabilistic hourly and intra-hour chronological production cost model designed

specifically for resource adequacy and system flexibility studies

SERVM calculates both resource adequacy metrics and costs

SERVM used in a variety of applications for the following entities:

• Southern Company

• TVA

• Louisville Gas & Electric

• Kentucky Utilities

• Duke Energy

• Progress Energy

• FERC

• NARUC

• PNM

• TNB (Malaysia)

• EPRI

• Santee Cooper

• CLECO

• California Public Utilities Commission

• Pacific Gas & Electric

• ERCOT

• MISO

• PJM

• Terna (Italian Transmission Operator)

• NCEMC

• Oglethorpe Power


9

SERVM Uses Resource Adequacy

Loss of Load Expectation Studies

Optimal Reserve Margin

Operational Intermittent Integration Studies

Penetration Studies

System Flexibility Studies

Effective Load Carrying Capability of Energy Limited Resources

Wind/Solar

Demand Response

Storage

Fuel Reliability Studies

Gas/Electric Interdependency Questions

Fuel Backup/Fixed Gas Transportation Questions

Transmission Interface Studies

Resource Planning Studies

Market Price Forecasts

Energy Margins for Any Resource

System Production Cost Studies

Evaluate Environmental/Retirement Decisions

Evaluate Expansion Plans


10

Resource Commitment and Dispatch

8760 Hourly Chronological Commitment and Dispatch Model

Simulates 1 year in approximately 1 minute allowing for thousands of

scenarios to be simulated which vary weather, load, unit performance, and

fuel price

Capability to dispatch to 1 minute interval

Respects all unit constraints

Capacity maximums and minimums

Heat rates

Startup times and costs

Variable O&M

Emissions

Minimum up times, minimum down times

Must run designations

Ramp rates

Simulations are split across multiple processors linked up to the SQL

Server


11

Resource Commitment and Dispatch

Commitment Decisions on

the Following Time Intervals

allowing for recourse

Week Ahead

Day Ahead

4 Hour Ahead, 3 Hour

Ahead, 2 Hour Ahead, 1

Hour Ahead, and Intra-Hour

Load, Wind, and Solar

Uncertainties at each time

interval (decreasing as the

prompt hour approaches)

Benchmarked against other

production models

47,000

48,000

49,000

50,000

51,000

52,000

53,000

0 0.5 1 1.5 2 2.5 3 3.5 4

Net

Lo

ad

MW

Hour

1 - 4 Hour Ahead Forecast Error

Actual Net Load Forecast Error Range from Hour 0

At hour 0, SERVM draws from correlated load, wind,

and solar forecast error distributions for intra-hour, 1 hr

ahead, 2 hrs ahead, 3 hrs ahead, and 4 hrs ahead

uncertainty. SERVM then makes commitment &

dispatch adjustments based on the uncertain forecast,

but ultimately must meet the net load shape that

materializes.

Current Position: t = 0


12

Ancillary Service Modeling

Ancillary Services Captured

Regulation Up Reserves

Regulation Down Reserves

Spinning Reserves

Non Spinning Reserves

Load Following Reserves

Co-Optimization of Energy and Ancillary Services

Each committed resource is designated as serving energy or energy plus one of the

ancillary services for each period


13

SERVM Framework

Base Case Study Years (2021 and 2024)

Weather (36 years of weather history)

Impact on Load

Impact on Intermittent Resources

Economic Load Forecast Error (distribution of 5 points)

Unit Outage Modeling (thousands of iterations)

Multi-State Monte Carlo

Frequency and Duration

Base Case Total Scenario Breakdown: 36 weather years x 7 LFE points = 252 scenarios

Base Case Total Iteration Breakdown: 252 scenarios * 10 unit outage iterations = 2,520 iterations

Intra Hour Simulations at 5-minute Intervals


14

Reserve Margin Study


15

Load Modeling: Summer Peak Weather Uncertainty Includes Entire Balancing Area

-6.0%

-4.0%

-2.0%

0.0%

2.0%

4.0%

6.0%

8.0%20

04

19

87

20

01

19

88

19

84

19

97

19

86

20

00

19

99

20

08

20

09

19

82

20

02

20

06

19

83

19

93

20

12

20

14

19

96

20

03

19

85

20

07

20

05

19

92

20

11

19

91

20

15

20

10

19

90

19

89

19

98

19

81

19

95

19

80

20

13

19

94

% F

rom

No

rmal

Weath

er

Fo

recast

Weather Year


16

Economic Load Forecast Error

Using CBO GDP approach and assuming 30% multiplier for electric

load growth compared to GDP growth

Load Forecast Error Multipliers Probability %

0.95 2.7%

0.97 14%

0.99 23.8%

1.00 19.1%

1.01 23.8%

1.03 14%

1.05 2.7%


17

0%

10%

20%

30%

40%

50%

60%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Av

era

ge

Ca

pa

city

Fa

cto

r

Hour of Day

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sept

Oct

Nov

Dec

Wind

Solar (Albuquerque)

Renewable Shapes: 36 Years

• Wind Shapes Based on Historical Data

• New Wind given 40+% Capacity Factor

• Solar Shapes Based on NREL Data

• Represented by 6 Locations Across the

State

• Both Fixed and Tracking 0%

10%

20%

30%

40%

50%

60%

70%

80%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Norm

aliz

ed O

utp

ut

(% o

f N

am

epla

te)

Hour of Day

Fixed Tracking


18

Unit Outage Modeling

Full Outages

Time to Repair

Time to Failure

Partial Outages

Time to Repair

Time to Failure

Derate Percentage

Startup Failures

Maintenance Outages

Planned Outages

Based on Historical Operation

Unit Name Capacity (MW) EFOR P. VERDE_1 134 2.15 P. VERDE_2 134 0.73 P. VERDE_3 134 3.11

FOURCORN_4 100 20.75 FOURCORN_5 100 17.61 SAN JUAN_1 170 18.29 SAN JUAN_4 327 16.06

LUNA_1 185 5.42 AFTONCC_1 230 5.77

DELTA_1 138 9.32 La_Luz 40 6.64

REEVES_1 44 5.05 REEVES_2 44 0.71 REEVES_3 66 7.17

VLNCPPA_1 145 1.43 LORDSBRG_1 40 7.08 LORDSBRG_2 40 6.45


19

System Unplanned Outages

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 200 400 600 800 1000 1200

Perc

en

t o

f T

ime

System MWs offline

10% of the time, the system has

more than 15% of its fleet

capacity offline due to unplanned

outages


20

BA = PNM + Tri-State

Regions

Committed and

Dispatched as a single

region

100

130 50

100

107

50

610

1200

100

610

1300

230

60 Bi-directional

204

64 25

141

50

0

50

200 0

100

80 80

0

300

Arizona

Arizona Entities

(APS, AEPCO, TEP,

Salt River Project,

Gila River Power

Station)

El Paso Electric

Southwestern Public Service

Company

PNM - Four Corners

PNM ownership of

PV 1 - 3, FC 4 - 5,

SJ 1 - 4

PNM - North

Reeves 1 - 3,

Rio Bravo, Valencia,

Renewables

PNM - South

Afton CC,

Lordsburg1,

Lordsburg 2, PNM

portion of LUNA 1

Public Service Company of

Colorado

Tri - State North

Tri - State South

Study Topology and Market Assistance 2013 Reserve Margin Study


21

Emergency Operating Procedures

Demand Response

Firm load shed to maintain reserves equal to 4% of load

Power Saver

Program

Peak Saver

Program

Capacity (MW) 33.75 20

Season June-Sept June-Sept

Hours Per Year 100 100

Hours Per Day 4 6


22

LOLECAP Results 2021 Study Year

Month LOLECap

1 0%

2 0%

3 0%

4 0%

5 0%

6 22%

7 45%

8 32%

9 1%

10 0%

11 0%

12 0%

• Recommend minimum reserve margin at no higher than 0.2 LOLECAP: 17% reserve margin

• Traditional 1 day in 10 year standard: 0.1 LOLECAP = 21% reserve margin

• Assuming no external regions, a 16.5% reserve margin results in over 8 LOLECAP events per year.

Neighbor assistance during peak hours can range from 0 MW to 300 MW depending on neighbor

conditions.


-

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

8% 9% 10% 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% 21% 22%

BA

LO

LE

CA

P

Reserve Margin

23

Flexibility Metrics


24

Traditional "Generic Capacity" Metrics New "Flexible Capacity" Metrics

Definitions of Existing and New Reliability Metrics

LOLEGENERIC-CAPACITY

Traditional metric to capture events that occur due to

capacity shortfalls in peak conditions

LOLEMULTI-HOUR


deficiencies of longer than one hour in duration

LOLEINTRA-HOUR


deficiencies inside a single hour

20,000

30,000

40,000

50,000

1 5 9 13 17 21

41,500

42,500

43,500

44,500

45,500

10:00 10:10 10:20 10:30 10:40 10:50 11:00

0

10,000

20,000

30,000

40,000

50,000

60,000

1 3 5 7 9 11 13 15 17 19 21 23

MW

Hours

Load Generation

LOLECap LOLEFLEX


25

Renewable Curtailment Example

-

500

1,000

1,500

2,000

2,500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

MW

Hour of Day

Load

Renewable + Nuclear

Renewable

Curtailment

Represents a day in the 2021 50% RPS scenario


26

Flexibility Study Approach

Identify LOLEFLEX events and renewable curtailment (overgen) events

Solve the deficiencies using the following approaches and calculate

costs:

Change operating procedures (i.e. raise load following requirement)

Add flexible capacity and/or replace existing capacity

Production Costs = Fuel Costs + Variable O&M + Emission

Costs + Cost of EUE

PPA prices assumed for wind and solar


27

Base Case Physical Reliability Results Varying Operating Reserve Levels

Study Years: 2021

Reg Requirement: 4% of Load

Spin + Load Following Requirement =

Simulated at 5% and 7%

Quick Start Target: 4% of Load

Renewable Penetration

LF Target

Renewable Curtailment


LOLE CAP

LOLE FLEX

Production Costs

% of Load % of Load

% of Renewable MWh

Events Per Year

Events Per Year M$

2021 Base Case 17.0% 5.0% 1.8%

43,131

0.18

0.18 369.9

2021 Base Case 17.0% 7.0% 1.9%

45,019

0.18

0.08 373.3

System reliability is

reasonable with 7%

LF target

Cost

Increase

due to

LF

increase

Curtailment Comparison 1x20 MW solar plant = annual output of 44,000 MWh


28

Base Case Physical Reliability Results Varying Operating Reserve Levels

Study Year: 2024

Reg Requirement: 4% of Load

Spin + Load Following Requirement =

Simulated at 5% and 7%

Quick Start Target: 4% of Load

Renewable Penetration LF Target


Renewable Curtailment LOLE CAP

LOLE FLEX

Production Costs

% of Load % of Load % of

Renewable MWh Events

Per Year Events

Per Year M$

2024 Base Case 21.1% 5.0% 1.1%

34,213

0.19

0.44 514.2

2024 Base Case 21.1% 7.0% 1.1%

34,747

0.19

0.11 519.6

System reliability is

reasonable with 7%

LF target

Cost

Increase

due to

LF

increase



29

2024 Base Case (Monthly Basis)

2024 LOLE Cap LOLE Flex Curtailment Month Events Per Year Events Per Year MWh

1 - 0.007 2,667 2 - 0.008 3,455 3 - 0.026 3,718 4 - 0.016 3,413 5 - 0.007 4,747 6 0.036 0.002 1,918 7 0.090 0.009 931 8 0.064 0.007 1,001 9 0.003 0.000 1,920

10 - 0.016 2,440 11 - 0.011 4,494 12 - 0.005 3,044

Total 0.193 0.114 33,747


30

RPS Scenarios and Rules of Thumb


31

2024 RPS Scenarios

To build RPS portfolios, additional solar and wind was added to the

system: Either 66% of the incremental additions were designated as wind

or solar

Technology LF Target

Base Case 5%, 7%

Base Case 40% RPS (66.7% Solar) 7%, 13%, 15%

Base Case 40% RPS (66.7% Wind) 7%, 13%, 15%






32

Net Load Intra Hour Uncertainty

-300

-200

-100

0

100

200

3000

%

4%

8%

12%

16%

20%

24%

28%

32%

36%

40%

44%

48%

52%

56%

60%

64%

68%

72%

76%

80%

84%

88%

92%

96%

100

%

Intr

a H

ou

r U

nce

rtain

ty (

5 m

in M

W D

ive

rge

nce

)

Cumulative Probability

2024 Base Case

2024 40% RPS

2024 50% RPS

2024 80% RPS


33

2024 RPS Scenarios @ 7% LF

Renewable Penetration LF Target Curtailment Curtailment LOLEFLEX

Production Costs

% of Load % of Load % of

Renewable MWh Events Per

Year M$

Base Case 21.1% 7.0% 1.1%

33,747

0.11 519.6

Base Case 40% RPS (66.7% Solar) 40.9% 7.0% 10.5%

613,496

6.83 527.5

Base Case 40% RPS (66.7% Wind) 40.6% 7.0% 7.8%

454,097

5.04 523.5


1,333,517

16.05 553.8


948,907

16.39 541.4


4,457,962

55.62 686.2


3,423,730

68.60 650.3


1x500 MW solar plant = annual output of 1,100,000 MWh


34


0

10

20

30

40

50

60

70

80

0% 20% 40% 60% 80% 100%

LO

LE

Fle

x @

7%

LF

RPS


35



LF Target Renewable Curtailment


LOLEFLEX

Production Costs

% of Load % of Load

% of Renewable MWh

Events Per Year M$

Base Case 21.1% 7.0% 1.1% 33,747

0.11

519.6

Base Case 40% RPS

(66.7% Solar) 40.9% 15.0% 12.9%

751,179

0.44

554.0

Base Case 40% RPS

(66.7% Wind) 40.6% 15.0% 10.0%

579,932

0.28

549.7

Base Case 50% RPS

(66.7% Solar) 51.4% 15.0% 19.8% 1,450,165

1.22

575.1

Base Case 50% RPS

(66.7% Wind) 50.8% 15.0% 15.3% 1,111,892

0.85

564.2

Base Case 80% RPS

(66.7% Solar) 82.8% 15.0% 38.8% 4,578,472

7.12

694.7

Base Case 80% RPS

(66.7% Wind) 81.7% 15.0% 31.0% 3,610,875

10.15

656.2


36


0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

0% 20% 40% 60% 80% 100%

Re

ne

wa

ble

Cu

rta

ilm

en

t (%

of

Re

ne

wa

ble

)

RPS

Average Curtailment

Marginal Curtailment

Average = % of total renewable fleet curtailed at each RPS level

Marginal = % of next MWh that will be curtailed at each RPS level


37

Cost by RPS Level

Assumes $39 PPA pricing for new solar and $40 PPA pricing for new wind.

0%

10%

20%

30%

40%

50%

60%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

% C

os

t In

cre

as

e C

om

pa

red

to

Ba

se

C

as

e

RPS


38

2024 RPS Scenarios add Flexible Generation or Battery


LF Target Curtailment Curtail-ment

LOLECAP LOLEFLEX

ProductionCosts

% of Load % of Load % MWh Events Per

Year Events Per

Year M$

Base Case 40% RPS (66.7%

Wind) 40.6% 13% 9.4%

541,689

0.10

0.48

543.0


Wind) 40.6% 15% 10.0%

579,932

0.10

0.28

549.7


Wind) and 2 LM6000 (80 MW) 40.6% 13% 9.2%

534,093

0.04

0.50

539.0


Wind) and 100 MW 2 hour storage 40.6% 13% 8.9%

514,306

0.04

0.31

536.7



495,383

0.03

0.27

535.7



483,445

0.02

0.27

535.5


Represents First Pass at Results; Production Costs do not include capital costs

39

Load Following Requirements General Takeaways


RPS Required Load Following Renewable Curtailment

20% 7% of Load <50,000 MWh

30% 13% of Load 100,000 - 200,000 MWh

40% >15% of Load 400,000 - 800,000 MWh

50% >15% of Load + Significant Additional

Flexible Resources 900,000 - 1,500,000 MWh

80% >15% of Load + Significant Additional

Flexible Resources 3,500,000 - 4,500,000 MWh

pnm preliminary reliability analysis

Documents