Locational Capacity Demand Curves in ISONE Samuel A Newell Kathleen Spees Ben Housman October 7 2014 ISO New England Markets Committee Contents Introduction Impact of Model Updates Stakeholder Questions ID: 812414
Download The PPT/PDF document "4 th Stakeholder Update:" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
4th Stakeholder Update: Locational Capacity Demand Curves in ISO-NE
Samuel A. NewellKathleen SpeesBen Housman
October 7, 2014
ISO New England Markets Committee
Slide2Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide3IntroductionObjectives for Today
Benchmark the impact of model updates (described in ISO-NE’s prior presentation)Provide
analytical results illustrating performance of
ISO-NE’s proposed locational
demand curve and market clearing
rules
Summarize analysis of stakeholder proposed curves and respond to stakeholder
questions
Slide4IntroductionIndex to Stakeholder Questions
Importing Zone Questions
Slides
NU &UI: FCA
8 Back-Cast, ISO-NE Proposed Curve, FCA 8 Cleared Supply as Price-Takers
11
NU: FCA
8 Back-Cast, ISO-NE Proposed Curve, FCA 8 Cleared Supply, Plus 600 MW plant in CT at $7/kW-m
11
NU: Cap at 1-in-5, Foot at 1-in-87
12
NESCOE: NESCOE Curve,
70% shock size, more elastic supply curve
13
CT PURA & DEEP:
Proposed
Curve
, LOLE capped at TSA
14-16
CT PURA & DEEP:
Proposed
Curve
, No Demand Shock
14-16
CT PURA & DEEP:
Proposed
Curve
Shifted
to LOLE at 0.105, 70% Shock Size
14-16
CT PURA & DEEP:
Proposed
Curve
Shifted
to LOLE at 0.105, LOLE capped at TSA
14-16
CT PURA & DEEP:
Proposed
Curve
Shifted
to LOLE at 0.105, No Demand Shock
14-16
CT PURA & DEEP: Cap at 1-in-5, LOLE at 0.105
14-16
CT PURA & DEEP: Cap at 1-in-5, LOLE at 0.105, LOLE Capped at TSA
14-16
CT PURA & DEEP
: Cap at 1-in-5, LOLE at 0.105, LOLE Capped at TSA, No Demand Shock
14-16
UI: What Are Local Prices When LOLE is Similar to System?
17-18
Exporting Zone Questions
Slides
Emera
Energy:
1.5x Width
20
NESCOE:
Cap at MCL, Foot at 1x Foot
21
NESCOE: Cap at MCL, Vertical to
System Net CONE, Foot at 1x Foot
21
Slide5Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide6Model UpdatesDescription of Model Updates
We describe and show the impacts of two simulation model changes that we have implemented to reflect ISO-NE’s requested updates (see Appendix and written stakeholder responses for more detail):
Updated LOLE Calculation and Reporting
FCM Auction Clearing Rules
The updated LOLE calculation results in minor changes
Adjustment to the auction clearing has more impact on
results
Differences occur in
approximately
1/3 of draws, when one or both import zones price separate
Results are the same as the prior clearing model if there is no price separation (if assuming the same supply and demand in a particular draw, i.e. with no “smart block” adjustment)
Comparison of results for a range of zonal curves is included in appendix
Slide7Model UpdatesImpact on System Results
Updated LOLE calculation shows minimal impact (but new reported metric provides additional information)Revised clearing algorithm
slightly increases price volatility and reduces reliability system-wideEffect becomes larger with increasing width of local curves and number of importing zones
System Performance Impacts of Model Updates
ISO-NE Proposed System and Local Curves (1x No TTC)
Slide8Revised clearing algorithm shows slightly lower price volatility in zones
Reliability below
target
in zones,
associated
primarily with lower reliability
on
a system basis (see prior slide)
But more supply is locally-sourced (reducing the contribution of LOLE from locally-driven events)
Model Updates
Impact on Importing Zone Results
Importing Zone Impacts of Model Updates
ISO-NE Proposed System and Local Curves (1x No TTC)
Slide9ISO-NE Proposal
Impact on Exporting Zone Results
Minimal changes to realized performance in Maine
Primary performance change is associated with impact of revised clearing mechanics on system price and reliability results
Exporting Zone Impacts of Model Updates
ISO-NE Proposed System and Local Curves (1x Curve in Exporting Zone)
Slide10Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide11Back-Cast of FCA 8
Results
Import-Constrained Zones
NU & UI: FCA 8 Back-Cast with Proposed Curves
Request to re-run FCA 8 auction results with system and local demand curves treating cleared supply as price takers
No price separation occurs across zones
Back-cast prices would have dropped from $15.0 to $13.9/kW-m in the presence of the system demand curve
Adding a 600 MW plant would have reduced the prices to $11.9/kW-m
Assumptions
FCA 8 demand
parameters
System demand curve with Net CONE at 11.08/kW-m
ISO-NE proposed curve in zones
Case 1: FCA 8
cleared
supply as price-takers
Case 2: Add a 600 MW plant at $7/kW-m in Connecticut
Note:
Cleared quantity excludes TTC.
Slide12Performance
Import-Constrained Zones
NU Curve: Cap at 1-in-5, Foot at 1-in-87
Concept is to define local demand curves according to the same LOLE-based definition as in system
Results in flatter local curves and lower price volatility
However, reliability is degraded in both zones:
Result may appear counter-intuitive, but is a consequence of implementing a wide curve in combination with updated clearing rules
Wide local curves result in less rest-of-system procurement during price separation, therefore reduced rest-of-system reliability and increased rest-of-system price volatility
ISO-NE clearing
rules will prohibit use
of a
zonal curve
this
wide with three importing zones (demand quantity in
zones
could exceed system curve width)
Connecticut
Note:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all
cases.
Slide13Import-Constrained Zones
NESCOE Curve w/ 70% Shocks, Supply Elasticity
Requested analysis of NESCOE proposed curve (1x system ratio, above and below LSR; see figure on slide
23)
With three changes to modeling assumptions that each reduce price volatility and improve reliability: (1) reduce shock sizes to 70% of base; (2) increase supply elasticity (see right); (3) remove supply lumpiness
Combined effect under sensitivity assumptions is a substantial improvement in all price volatility and reliability metrics
Performance
Supply Curves with More Elasticity
Slide14Import-Constrained Zones CT PURA & DEEP: Simulation Requests
CT PURA & DEEP requested simulation results under different combinations of sensitivity assumptions, and using two different LOLE tuning approachesSensitivity assumptions would reduce price volatility and increase reliability, including:
70% shock sizes No demand shocks
LOLE capped at the value when local MW are at TSA (i.e. consistent with an assumption of ISO-NE intervention to restore local MW to TSA)
Tuning approaches
are conceptually similar to
approaches we used system-wide, with local LOLE tuned to 0.105 LRA target (note: this request contemplates allowing for the violation of the “minimum acceptable” at TSA)
However, because system LOLE is above 0.105 under the revised clearing
mechanics,
no amount of tuning can restore the zones to this level. We therefore re-interpret this question as tuning the local curves to reach a local adder of
0.005 LOLE (consistent with the LRA delta above NICR)
Re-interpreted tuned curves reflect:
ISO-NE proposed curve, shifted to a local LOLE adder of 0.005 events/year
Cap at 1-in-5 LOLE, adjust foot to meet a local LOLE adder of 0.005 events/year
Slide15Import-Constrained Zones
CT PURA & DEEP: Tuned Curves
Candidate Curve, Left/Right Shifted to Tune LOLE*
Cap at 1-in-5, Foot Adjusted to Tune LOLE*
NEMA
Connecticut
NEMA
Connecticut
Notes:
The “truncated LOLE” curves were not possible to develop in NEMA because the LOLE adder at TSA was too low (i.e. the LOLE adder cannot be increased to 0.005 because the LOLE adder at TSA is close to
zero).
See appendix slide 34 for system simulation results under the same
curves.
Slide16Import-Constrained Zones
CT PURA & DEEP: Simulation Results
Note: Tuned curves could not be drawn in NEMA in the cases where local LOLE was capped at TSA. Because the local LOLE “adder” becomes close to zero at TSA, the local curve could never be 0.005 events/year above Unconstrained System LOLE.
See appendix slide 34 for system simulation results.
Slide17Import-Constrained ZonesUI: What Are Local Prices When LOLE is Similar to System?
Question:
“How often would we expect customers in import constrained zones to pay higher capacity prices than customers in the rest of pool for the same level of reliability (as measured by the LOLE metric)?”
We focus on the subset of draws in which: (a) local price separates above system (b) local LOLE is within 0.005 events/year above unconstrained system LOLE
16% and 2% of draws fall into this subset in NEMA and CT respectively (greater in NEMA because TSA & LSR are right-shifted compared to LRA, and so local LOLE adder is usually very small)
The results in this
subset
of draws may seem counter-intuitive when taken individually (i.e. why pay a price premium in the zones when local reliability is no worse than rest of pool?)
However, these results are consistent with the overall reliability and price volatility objectives of the demand curve and FCM, because the price premium paid in this subset of draws helps to support sufficient local supply to prevent larger price separation or lower-reliability events from occurring in other years
Further, focusing only on LOLE does not account for the transmission security objective TSA
Slide18Import-Constrained Zones
UI
:
Comparison of
Statistics Across
All
Curves
Comparing all candidate curves in instances where:
The zone price-separates, and
LOLE adder in zone is < 0.005
Anywhere from 0-22% of draws depending on the curve
NEMA always shows greater
frequency from
these draws
for
any
curve drawn to the right of TSA Average price differential above system is the same across all curves (even though the price differential associated with this subset of draws varies substantially)
Slide19Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide20Export-Constrained Zone
Emera
Energy: 1.5x Width Curve
Emera
Energy proposed a curve with width 1.5x of system ratio (flatter than proposed)
Results in modest improvement in price volatility in Maine (modest degradation in System LOLE) compared to ISO-NE proposal
Performance
Note:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all
cases.
Slide21Export-Constrained Zone
NESCOE: Curves at or Above MCL
NESCOE proposed looking at two curves that would be right-shifted compared to MCL
Both curves perform similarly to the ISO-NE proposal (very small increase in price volatility in both cases; and degradation in system reliability in one case)
Both options would
prevent procuring less than MCL supply in
Maine
in
the event of system-wide
shortage
Notes:
Runs modeled with ISO-NE Proposed curve in NEMA/Boston and Connecticut.
Curves
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all cases
Performance
Slide22Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide23Curve Comparison
Alternative Importing Zone Curves (Connecticut)
Note:
Curves
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all
cases.
Slide24Curve Comparison
Performance in Import-Constrained Zones
Notes:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all cases.
Base case assumes true Net CONE in NEMA/Boston and Connecticut is 10% higher than system.
Zonal load costs reflect capacity procurement costs paid by customers in each zone, assuming all zonal CTRs are awarded to local customers.
Curve Comparison
System-Wide Performance Impacts
Notes:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all cases.
All curves were run with the ISO-NE Proposed 1.0x System Ratio (No TTC) curves
for NEMA/Boston and
Connecticut and the ISO-NE Proposed 1.0x System Ratio curve in Maine.
Slide26Curve Comparison
Alternative
Exporting Zone
Curves
Note:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s
recommended approach
in all cases
Slide27Curve Comparison
Performance in
Export-Constrained
Zones
Notes:
Curve
names are labeled based on the stakeholder suggesting those curves for analysis, but do not necessarily reflect that stakeholder’s recommended approach in all cases.
All curves were run with the ISO-NE Proposed 1.0x System Ratio (No TTC) curves
for NEMA/Boston and
Connecticut.
Base case assumes true Net CONE in NEMA/Boston and Connecticut is 10% higher than system, Maine Net CONE is 10% lower than system .
Zonal load costs reflect capacity procurement costs paid by customers in each zone, accounting for CTRs that are awarded to local customers.
Slide28Contents
IntroductionImpact of Model Updates
Stakeholder QuestionsImport-Constrained Zonal Demand Curves
Export-Constrained Zonal Demand Curves
Summary Comparison of Zonal Demand Curves
Appendix
Slide29AppendixUpdated LOLE Calculation and Reporting
System:Previously reported only “System LOLE,” which reflected the LOLE that ISO-NE would calculate if there were no internal constraints (i.e. “copper sheet” assumption used when estimating NICR)
That metric will now be labeled as “Unconstrained System LOLE” and used as a primary metric for evaluating the system demand curveAlso reporting a new metric “Constrained System LOLE,” which reflects the max of the LOLE of any zone in any one draw (consistent with NPCC definition of system LOLE in the presence of zonal constraints)
Import Zones:
Zonal LOLE is affected by MW of supply both within the zone (determines local events) and outside the zone (determines system events)
Previously calculated zonal LOLE as the maximum of local and system LOLE
Updated approach calculating local LOLE “adder” on top of system events (results very similar to prior approach)
Revised
Local
LOLE Approach
Local
LOLE
Unconstrained System LOLE
“Local LOLE Adder”
Slide30Appendix
Updated Local Clearing Approach
System-Wide Clearing Price
and Quantity
Updated clearing mechanics will produce the same prices as the prior model if there is no price separation (clear on system demand curve)
If importing zones
do
price-separate, then system-wide prices and quantities will clear
below
and to
the
left compared
to
prior modeled clearing mechanics (which assumed system + all zones supply would clear on
the
aggregate
system
demand curve)
Affects approximately 1/3 of draws (any time one or both import zones price separate
)
Clearing in Example Draw
Example Draw
(See Below)
Example
Draw
Individual Draws
(Each w/ unique “Residual” System Demand Curve)
Slide31Appendix
Range of Import Zone Curves Compared
We tested a
range of import zone demand curves under the updated simulation
model
Updated LOLE Calculation
FCM Auction Clearing Rules
Intended to provide stakeholders more information about
results
that can be expected with varying widths of local curves
General observations:
System reliability and price volatility worsen with wider zonal curves
Zonal metrics move in the opposite directions (i.e. local reliability and price volatility improve with wider curves)
Connecticut
Curves Tested Under Updated Model
Note:
“1x System (No TTC)” curve applies the system ratio multiplier to NEMA and CT local curves without including the TTC MW in the calculation.
Slide32Appendix
System Simulation Results
Note: All runs modeled with a 1x system ratio curve in Maine.
Slide33Appendix
Importing Zones
Simulation Results
Note: All runs modeled with ISO-NE Proposed 1x system ratio curve in Maine
.
Slide34Appendix
CT PURA & DEEP System Simulation Results
Note: All runs modeled with a 1x system ratio curve in Maine.