Economic Value of Fossil Fleet Operational Flexibility
Modeling and Insights in US-REGEN
John Bistline, Ph.D.
Senior Technical Leader
36th Annual Fuels Seminar-Washington, DC
November 8, 2017
17cv01906 Sierra Club v. EPA
r--I
L--I
IGI
i/
o ; \ ............ ;
2017 Electric Power Research Institute, Inc, ZUI rights reserved.
ED_001523_00006046-00001
Motivations: Low Gas Prices, High Renewables, and Flexibility
Value of dispatchable assets on the
grid grows as the share of nondispatchable resources increases
Value prospects of minimum load operations for flexible generators
California Independent System Operator net generation, March 11,2017
gigawaBhours
disiriboted solar utilltf-scal solar iBtports tier reBewables
tiennal nuclear
hydroelectric
- Avoid negative operating margins
- Provide grid services (e.g., reserves)
- Avoid frequent startups
17cv01906 Sierra Club v. EPA
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
l--l
i
ED_001523_00006046-00002
Spectrum of Flexible Operations
Operating Mode
Defining Characteristics
Start Reliability Minimum Load
Ramp Rate Operating Reliability
Cost ..i
Start Reliability Startup Speed Minimum Load
Ramp Rate Operating Reliability
'Cost
dinimum Load
Preservation of Equipment
Availability of
Equipment
i (Startup Speed)
A
..Cost
05
Energy Market
Balancing Services Markets
Capacity Market
Credit: Mike Caravaggio
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00003
Research Questions: Economics of Flexible Operations
How can enhanced operational flexibility (in this analysis, lower minimum loads) for coal and gas assets impact operations and profitability under large-scale renewable deployment?
How does the economic value of this flexibility vary across different regions, asset types, and market environments (e.g., level of renewable deployment)?
Goal is to evaluate the economic impact of improving technical parameters like minimum load to increase flexibility
4
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
I--.1 ICI i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00004
Price Duration Curves Drive Net Revenues for Power Plants
Example without Commitment Costs and Constraints
17cv01906 Sierra Club v. EPA
2017 Electric Power Research Institute, Inc, All rights reserved.
l--l
i
ED_001523_00006046-00005
Impacts of Commitment Costs and Constraints
Minimum Down Time
When operating margins for plants (e.g., spark spreads for gas units) turn negative, a plant can:
1. Ramp down: Stay online at minimum stable capacity, which forces the unit to absorb losses for generation but avoids shutdown and startup costs
2. Shut down: Turn off entirely, which avoids operating at a loss but with additional costs
Time (hours)
Output (M W )
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00006
Turn Down for What? Physical Impacte of Flexible Operations
Thermal Fatigue Damage
Corrosion Fatigue
Steam Turbine Impacts
Predominate failure mode in boiler/turbine components subject to: frequent starts, fast ramping, etc.
Caused by temperature mismatches between steam and metal surfaces
Tube Failures
On drum units, corrosion fatigue has been observed on riser tubes
Involves combination of: manufacturing-induced bend stress, water chemistry fluctuations, etc.
Steam Chemistry: Offline Pitting
Reliability areas impacted: differential growth, thermal and speed cycling, creep fatigue, steam chemistry effects, erosion and corrosion, lowload impacts on blade flutter
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00007
Example: NGOC Operations in California
0 GWSolar
$50
fs
1I '
I
lift <40
Seftemfcer $30
Day 1
Day 2
Day 3
Day *
Day 5
NGOC output
(capacity factor)
glgt,g^^
io*
|t||M
9
1SSSSSSSSI m
r <
w f
IMgllBI^^
5
|11B|||^^
|i||jj^^
j|il|W^
^^
Day 6
Day 7
* 3 2
31.
Oli,
! . ` '
SeptBinber Day 1
positive revenue
(from generation during positive spark-spread hours)
Bay 3
Day 4
Day 5
Day 6
Bay 7
; 2017 Electric Power Research Institute, Inc, /Ml rights reserved.
U--1 id i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00008
Example: NGOC Operations in California
40 GWSolar
$50 Begg
-- s
S Si llBII
$ Mb
g jo
1,-sio |Bi[llB[||j|||^^ if5,20 lljlllll|W^ 430 iilillll^^
llll^^ ^1 lagliMM
Septewifcer 530
Day 1
tag
tag
Day 5
Oayf
put (capacity factor)
io soie
Day 7
7
6 S
5 1
3
3i 2
1
0*
E-.t,
tt-520 '
fao
Seftetiler
Day 1
negative revenue (from generation during negative spark-spread hours)
positive revenue
(from generation during positive spark-spread hours)
Bay 2
Day 3
Day 4
Dat 5
DayS
Day 7
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00009
Example: NGOC Operations in California
700 GI/1/S/ar
f -
S *v
spark spread ) (revenue less operating costs)
-
a > '
?
*'
Septensfcer Day 1
DSf 2
Day 1
Days
W O'
. A
f..... o
.$30 Siptimfcer
Day 1
Daya
positive revenue
(from generation during positive spark-spread hours)
_ startup and shutdown costs
Ulf a
Bay 4
Davs
2017 Electric Power Research Institute, Inc. All rights reserved.
put (capacity factor)
I
0w6
Dw 1
io
8 70S 6
5 2
I
30*
s 2
1
0
<
Days
oaf T
U--1 id i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00010
US-REGEN: EPRI's In-House Electric Sector Model
U.S. Regional Economy, GHG, and Energy
Capacity Expansion Customizable State/Regional
Economic Model, Long Resolution for Policy and
Horizon to 2050
Regulatory Analysis
Linked with Hourly Unit Commitment Model
For more information, see our website at
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00011
Scenario Description
Focus on four regions in 2035
Five policy scenarios
Paci1
1. Reference (current policies only)
2. OO2 tax ($16/t-CO2 in 2035)
3. Solar mandate (50% in-region demand)
4. Wind mandate (50% in-region demand) Ca|jfomja
5. 50/50 solar/wind mandate (50%
in-region demand)
Three asset types: coal and NGCC (existing and new)
Two parametric sensitivities
- Lower minimum load
- Startup costs
2017 Electric Power
17cv01906 Sierra Club v. EPA
Mountain-N Mountain-S
NW-Central
NE-Central
NE NY
M-Atlantic
S-Atlantic SW-Central sE^trai
Teas
Florida
ED_001523_00006046-00012
Model Parameters
Startup and Cycling Costs
Minimum Load Levels
Heat Rate
,
Cold Start Cos' lov.er Bojnd$-lnc:odes Odtheis
|< (Maintenance and capital cost pe^ MW capacity.
I
Low Hioh H
Coal
20% 50%
Coal
9.2
NGCC(New)
20% 50%
NGCC (Existing) 20% 50%
NGCC(New)
6.8
NGCC (Existing)
8.2
15% 20% 20%
Using 25th and 75th percentiles for low/high values, but Intertek considers these 'lower bound costsf'
By start type, technology, and cost category (O&M, fuel, and other)
Source: NREL/lntertek (2012), "Power Plant Cycling Costs"
Value of flexibility connected to initial ("High") and final ("Low") parameters for minimum load levels
Using the same values enables apples-to-apples comparisons across asset types
Sources: NREL/lntertek (2012), BNEF
Using representative assumptions for existing supercritical coal, new NGCC, and existing NGCC
Sources: ABB Velocity, EPRI
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00013
Detailed Regional Results
Enhanced Operational Flexibility and Cycling Impacts in Texas
17cv01906 Sierra Club v. EPA
2017 Electric Power Research Institute, Inc. All rights reserved.
ED_001523_00006046-00014
Scenario Assumptions Drive Investments: 2035 Capacity in Texas
250
Observations
New Solar Ex. Solar
New Wind
bex. wind
Hydro/Geo Gas Turbine NewNGCC B.NGCC Coal Other Nuclear
Wind development is extensive in the reference scenario
CO2 tax forces some coal retirements but more significant changes in dispatch
Under 50% RPS, lower capacity needs when portfolio of renewable technologies included
Caveat: Changes in other regions under these scenarios, which are not shown here
Reference
1
50% Wii
Mandate
2017 Electric Power Research Institute, Inc, All rights reserved.
l--l id i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00015
Changes in Operations with Lower Minimums
Texas, NGOC(New)
' '.'Loaf
Mi .imum Generation OStarts
8,000
7,000
5 O 6,000 M E 5,000 Stqu. 4,000 O
3,000
cc
< 2,000
1,000
0
Mir rio
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
Higher capacity factors when NGCC units can avoid starts
and lower losses
I--.1 iCI i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00016
Changes in Operating Revenues with Lower Minimums
Texas, NQOC(New)
Revenue 1 ; ng Cosi ' t..... tartup ONet
Total annual operating margins do not change as
much as dispatch
500
400
o
300
200
100
0
-100
-200
Min.
Scenario
High
Low | High
Low I High
Low I High
Low I High
Low
Reference
|
Tax
50% RPS (Wind)
50% RPS (Solar) I 50% RPS (Both)
2017 Electric Power Research Institute, Inc, All rights reserved.
I--.1 iCI i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00017
Long-Run Cost of Simple Dispatch
Texas
$14
Cost Increase from Startup O&M (m illion
$10 $8 S $4
$0
18
Ofw
lai
50941RS (lotltj
& 2017 Electric Power Research Institute, Inc, All ri
17cv01906 Sierra Club v. EPA
Notes
Vertical axis = long-run O&M costs due to cycling if dispatch is based only on short-run marginal costs o Based on Intertek (2012) o Range shows 25th/75th %ile
Costs track number of starts
Takeaways
CO2 tax scenarios increase cycling and costs for coal units
Startups are lower for coal since in-the-money for many hours (opposite for existing NGCC)
High renewable deployment drives greater system ramping/flexibility needs
I II ;
ED_001523_00006046-00018
Caveats
Not including full range of possible O&M costs
Not including possibility of early retirement
Not including changes in outage rates
Lowering minimum load reduces startups, which reduces E-OR
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
I--.1 ICI ;
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00019
Observations: Lower M inimum Loads in Texas
Expected returns for coal- and gas-fired generators are shaped by future market and policy conditions
- Operational characteristics are a second-order driver of profitability - Flexibility allows assets to adapt to bearish markets
Plant flexibility impacts operations more than annual margins
- Especially true for mid-merit units - Could impact long-run closures if cycling O&M is more costly than
currently anticipated, especially since damage mechanisms give rise to "lumpy" costs
Preparations for flexible operations can help regardless of which state-of-the-world obtains
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00020
Cross-Regional Comparison
How well do insights about flexible operations transfer to other regions?
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00021
Capacity and M inimum Generation Levels Vary across Regions
4*^ Reference Scenario in 2035 200
Observations
Grid sizes and compositions vary significantly without policy
IM
Storage (Battery)
Economics of wind are strong in
Storage (Hydro)
Texas and Mountain-N even
Mew Solar
without policy support
Ex. Solar
Mew Wind
S-Atlantic has more inflexible
Ex. Wind
generation (in absolute and
Hydra / Seo
relative terms) than other regions
Gas TUrblne
MewNGCC
Ex. NQCC
KCoal
. Other
Mudear
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
I--.1 id i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00022
Hourly Dispatch of New NGCC Unit under Reference Conditions
lantic
100%
st -- uU. 50%
0%
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
3,000
4,000
5,000
6,000
7,000
8,000
California
100%
-- 50%
0%
3,000
4,000
5,000
6,000
7,000
8,000
1,000
2,000
3,000
4,000
5,000
6,000
2017 Electric Power Research Institute, Inc, All rights reserved.
7,000
8,000
l--l ICI i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00023
Changes in Operating Hours with Lower Minimum Load
All Regions, Scenarios, and Asset Classes
8,000
7,000
ikn 5 6,000
e
5,000
E
<u O 4,000 G
mt/Cfding
"Baseloacl" --,---------- .--------
u - ,,
1,000
0
0
1,000
2,000
3,000
<0CB
5,000
6,000
7,000
Operating Hours with Higher Minimum load
Mid-merit units benefit most from lower minimum loads
8,000
2017 Electric Power Research Institute, Inc, All rights reserved.
1
I--.1 ICl ;
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00024
Profitability Increases with Lower Minimum Loads
$5.0
$4.5
$4.0 f $35
X $3.0
c
$2.5
a $2.0 z. .... . . $15
J S1
Value for Specific Region/Asset O Median
Takeaways
Flexibility can be valuable even without significant VRE buildout
Lower minimum generation is most valuable (on average) under a CO2 tax (e.g., coal as mid-merit)
Under 50% RPS scenarios, net profitability impacts are similar, even though system costs vary
o Seasonal/diurnal patterns differ, but offsetting effects
i ri.il
$0,5 $0.0
I
-I ZE
-I ZE
-I ZE
-I ZE
-I ZE
o RPS often induces seasonal layups, which lower the value of flexibility
Scenario
Reference
Tax
50% RPS 50% RPS 50% RPS
(Wind)
(Solar)
(Both)
2017 Electric Power Research Institute, Inc, All rights reserved.
l--l ICI i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00025
How Valuable Are Lower MinGen Levels with Reserve Revenues?
$30 j K <
/
( 1
1
($/MW-h)
.
{...
Coal
M6CC (Em.)
Takeaways
NGCC benefits most from reserve revenues due to higher number of hours at minimum generation (i.e., with capacity headroom)
Ancillary services revenues can boost economics of flex, ops., but markets are uncertain and "thin"
System forecast errors become more important to track over time
o Impact reserve needs/prices
o Essential also for determining least-cost commitment schedule, especially if starts are costly
2017 Electric Power Research Institute, Inc, /Ml rights reserved.
I--.1
i
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00026
Takeaways
Flexibility (e.g., through lower minimum loads) lowers costs
- Profitability impacts accrue in the long term, even though the most significant operational impacts are related to short-run dispatch
- Difficulty: Long-run cycling costs are uncertain!
Lower minimums are most valuable when...
- Policies rearrange merit order (e.g., CO2 tax) - Mid-merit units (e.g., new NGCC) are considered - Regions/scenarios do not force seasonal layup - Units are more flexible than other comparable generators
Discrepancy between system value of flexibility and incentivesfor generators to provide?
Ancillary services revenues improve prospects for flexible operations, but their market outlook and depth are unknown
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00027
Together.. Shaping the Future of Electricity
John Bistline
Senior Technical Leader 650-855-8517
LiTidP ^epj.com
17cv01906 Sierra Club v. EPA
ED_001523_00006046-00028