pnm preliminary reliability analysis
TRANSCRIPT
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
New metric to capture events due to system ramping
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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%
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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.
Preliminary Draft Results
-
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
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
New metric to capture events due to system ramping
deficiencies of longer than one hour in duration
LOLEINTRA-HOUR
New metric to capture events due to system ramping
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
Preliminary Draft Results
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
Preliminary Draft Results
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
Preliminary Draft Results
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
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
Preliminary Draft Results
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
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
Curtailment Comparison 1x20 MW solar plant = annual output of 44,000 MWh
Preliminary Draft Results
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
Preliminary Draft Results
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%
Base Case 50% RPS (66.7% Solar) 7%, 13%, 15%
Base Case 50% RPS (66.7% Wind) 7%, 13%, 15%
Base Case 80% RPS (66.7% Solar) 7%, 13%, 15%
Base Case 80% RPS (66.7% Wind) 7%, 13%, 15%
Preliminary Draft Results
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
Preliminary Draft Results
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
Base Case 50% RPS (66.7% Solar) 51.4% 7.0% 18.2%
1,333,517
16.05 553.8
Base Case 50% RPS (66.7% Wind) 50.8% 7.0% 13.1%
948,907
16.39 541.4
Base Case 80% RPS (66.7% Solar) 82.8% 7.0% 37.7%
4,457,962
55.62 686.2
Base Case 80% RPS (66.7% Wind) 81.7% 7.0% 29.4%
3,423,730
68.60 650.3
Curtailment Comparison 1x20 MW solar plant = annual output of 44,000 MWh
1x500 MW solar plant = annual output of 1,100,000 MWh
Preliminary Draft Results
34
2024 RPS Scenarios @ 7% LF
0
10
20
30
40
50
60
70
80
0% 20% 40% 60% 80% 100%
LO
LE
Fle
x @
7%
LF
RPS
Preliminary Draft Results
35
2024 RPS Scenarios @ 15% LF
Renewable Penetration
LF Target Renewable Curtailment
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
Preliminary Draft Results
36
Renewable Curtailment
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
Preliminary Draft Results
37
Cost by RPS Level
Assumes $39 PPA pricing for new solar and $40 PPA pricing for new wind.
0%
10%
20%
30%
40%
50%
60%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
% C
os
t In
cre
as
e C
om
pa
red
to
Ba
se
C
as
e
RPS
Preliminary Draft Results
38
2024 RPS Scenarios add Flexible Generation or Battery
Renewable Penetration
LF Target Curtailment Curtail-ment
LOLECAP LOLEFLEX
ProductionCosts
% of Load % of Load % MWh Events Per
Year Events Per
Year M$
Base Case 40% RPS (66.7%
Wind) 40.6% 13% 9.4%
541,689
0.10
0.48
543.0
Base Case 40% RPS (66.7%
Wind) 40.6% 15% 10.0%
579,932
0.10
0.28
549.7
Base Case 40% RPS (66.7%
Wind) and 2 LM6000 (80 MW) 40.6% 13% 9.2%
534,093
0.04
0.50
539.0
Base Case 40% RPS (66.7%
Wind) and 100 MW 2 hour storage 40.6% 13% 8.9%
514,306
0.04
0.31
536.7
Base Case 40% RPS (66.7%
Wind) and 100 MW 4 hour storage 40.6% 13% 8.6%
495,383
0.03
0.27
535.7
Base Case 40% RPS (66.7%
Wind) and 100 MW 6 hour storage 40.6% 13% 8.4%
483,445
0.02
0.27
535.5
Preliminary Draft Results
Represents First Pass at Results; Production Costs do not include capital costs
39
Load Following Requirements General Takeaways
Preliminary Draft Results
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