power systems engineering research center (pserc) - reduced … · for research, simpler model can...
TRANSCRIPT
Reduced Network Modeling of the Western Electricity Coordinating Council (WECC) as a Market Design Prototype
Jim PriceLead Engineering SpecialistCAISO, Market Analysis & Development
PSERC WebinarSeptember 6, 2011
2
Agenda
Overview of California ISO and its markets
Development of simplified WECC model through PSERC projects
Illustration of upcoming issues
Illustration of applying model to operational issues: How effective will various forms of storage & demand response be?
Other topics where models have been applied to current issues
3
California ISO provides open, non-discriminatory access to transmission grid
Responsibilities:• Reliability, grid planning, outage coordination• Market development, operations, monitoring
CAISO manages approximately 80% of California’s electricity load– 55,183 MW in-state power plant capacity– 10,000 MW import capacity– 50,270 MW record peak demand (7/24/2006)– 25,526 circuit-miles of transmission lines – 30 million people served– 286 million annual megawatt-hours of electricity
delivered annually– 38,000 generation & transmission outages per
year– Over 30,000 day-ahead market transactions per
day, similar volume for real-time market
1.8% coal12.2% large hydro13.9% hydro, geothermal,
biomass, wind, solar15.3% nuclear56.7% natural gasSource: 2009 Total ElectricitySystem Power, CaliforniaEnergy Commission
4
CAISO markets match supply & demand for reliability in day-ahead through real-time
Real TimeMarket
Manage energy flows on transmission grid with telemetry and 1-minute state estimator solutions
Update FNM for RT conditions
Dispatch balancing energy/ ancillary service
Hourly market for 24 hours of next day
Establish energy and ancillary service schedules
Manage congestion (transmission access) using Full Network Model (FNM)
Determine residual unit commitment requirements
Day AheadMarket
Hour AheadScheduling
Prior to real-time (RT) market, schedule energy and ancillary services for static interchange for 24 individual hours
Manage congestion using FNM
As one of 4 RT pre-dispatch processes, establish unit commitment & advisory schedules for internal & dynamic resources
5
A theme across CAISO markets is foundation on Full Network Model
In production, full network model provides needed detail for accurate market and dispatch
For research, simpler model can be adequate, while detail in results can hide trends.
PSERC projects using simplified WECC models: M-13: Agent Modeling for Integrated Power Systems M-21: Technical & Economic Implications of
Greenhouse Gas Regulation in a Transmission Constrained Restructured Electricity Market
M-24: Interactions of Multiple Market-based Energy and Environmental Policies in a Transmission-Constrained Competitive National Electricity Market
6
Contributing to PSERC & own research, CAISO assembled 240-bus WECC model
7
Model characteristics
Hourly loads for 11 areas within CAISO and aggregated sub-regions outside CAISO
Hourly profiles for wind & solar resources: 3 wind areas & 1 solar area in CAISO, 13 aggregated wind & 4 solar areas outside CAISO
Hourly geothermal & biomass aggregations in CAISO
Hydro optimization profiles using PLEXOS software Necessarily simplified, but focus is on differences between cases
23 aggregated gas-fired generators in CAISO (associated with owners) and 27 in WECC modeled as dispatchable, 17 coal plants & 4 nuclear as base-loaded
8
Changes over the next 10 years present operational (and research) challenges
Over 20,000 MW of wind and solar capacity is expected to be interconnected by 2020 – Increased supply volatility
Approximately 18,000 MW of thermal generation will be repowered or retired in next 10 years – Uncertainty surrounding thermal resources
Potential changes to load patterns as a result distributed generation and electric vehicles – Changing less predictable load patterns
9
20% and 33% Renewable Portfolio Standards lead to significant new resources
Biogas/Biomass
Geo-thermal
Small Hydro
Solar Thermal
Solar PV Wind
2006 Reference Case
701 1,101 614 420 2,648
Reference Case with 20% Renewables
701 2,341 614 2,246 6,688
Reference Case with 33% Renewables
1409 2,598 680 6,902 5,432 11,291
33 % solar PV includes 2,262 MW of customer side PV
10
Dispatch of conventional resources would not follow the typical pattern – emerging trend
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
0:00
0:39
1:18
1:57
2:36
3:15
3:54
4:33
5:12
5:51
6:30
7:09
7:48
8:27
9:06
9:45
10:2
411
:03
11:4
212
:21
13:0
013
:39
14:1
814
:57
15:3
616
:15
16:5
417
:33
18:1
218
:51
19:3
020
:09
20:4
821
:27
22:0
622
:45
23:2
4
MW
MW
Load, Wind, Solar Profiles April 2020
Load Net Load Wind Solar
Dispatch of conventional resources follow the red curve
Source: CAISO Renewable Resource Integration Studies
11
Average
Total Wind Generation
0
200
400
600
800
1000
1200
1400
1600
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
Wind production varies greatly from day to day and intra-day
April 2009 Daily Profiles
12
Wind and solar resources can vary significantly within days
0.0
5.0
10.0
15.0
20.0
25.0
0
20
40
60
80
100
120
140
160
0:00
0:41
1:22
2:03
2:44
3:26
4:07
4:48
5:29
6:10
6:52
7:33
8:14
8:55
9:36
10:1
810
:59
11:4
012
:21
13:0
213
:44
14:2
515
:06
15:4
716
:28
17:1
017
:51
18:3
219
:13
19:5
420
:36
21:1
721
:58
22:3
923
:20
Sola
r Gen
erat
ion,
MW
Win
d G
ener
atio
n, M
W
Time
Variability in Wind and Solar
Wind Solar PV
A 150 MW wind plant and a 24 MW solar resource
June 24, 2010
13
CAISO will need operational and market enhancements to support renewable integration
Operational Enhancements Wind & solar forecasting tools (output, ramping requirements) More sophisticated grid monitoring systems Over-generation mitigation procedures Coordination with neighboring balancing areas Generation interconnection standards Pilot projects (includes storage, demand response)
Market Enhancements New market products & changes to market rules Increased regulation and reserve requirements More sophisticated day-ahead unit commitment algorithms
1414
Requirements for integration of renewables
GenerationPortfolio Storage Demand
Response
Resources Required for Renewables Integration
Quick Start UnitsFast Ramping
Wider Operating Range (lower Pmin)
Regulation capability
Shift Energy from off-peak to on-peak
Mitigate Over Generation
Voltage SupportRegulation capability
Price sensitive loadResponsive to ISO dispatchesFrequency ResponsiveResponsive to Wind Generation Production
“Partners in Success”
Wind Generation
Solar Generation
Hydro Generation
Geo-thermal
Generation
15
Past CAISO research using simplified WECC model identified regional flow variability
Variability of Unscheduled Loop Flowat California-Oregon Intertie
-1200
-900
-600
-300
0
300
600
900
1200
5/1/
2006
5/16
/200
6
5/31
/200
6
6/15
/200
6
6/30
/200
6
7/15
/200
6
7/30
/200
6
8/14
/200
6
8/29
/200
6
9/13
/200
6
9/28
/200
6
MW
Loop Flow (Southw est to Malin)
Impact of Increased Intermittent Resources on Flow at COI (North to South)
Findings in paper 2010GM0783 (2010 IEEE General Meeting): existing issues of unscheduled flow at California-Oregon Intertie may increase as renewable resources develop throughout WECC region
16
Issues to explore: How effective will various forms of storage & demand response be?
Adding storage may reduce variability & reduce congestion
Modeling for 2011 IEEE PES paper (2011GM0942) provides similar preliminary results as previous paper:
• Comparison of existing vs. future resources shows similar increases in congestion
Adding storage may stabilize flows & reduce congestionImpact of Increased Intermittent Resources on Flow at COI (North to South)
17
More issues remain to be explored in modeling
Disaggregating 4 of 166 generators (to 24 units) may have significant difference in results
• Demand of electric vehicles may be added with or without coordination of charging with system conditions. Coordination may allow higher transmission utilization.
• Aggregating generators simplifies model but impacts results.Coordinating electric vehicle charging may improve transmission utilization.
18
Simple model as prototype for market shows potential for focusing research
Experience with model confirms increased ease of testing market alternatives compared to fully detailed full network model.
As PSERC project M-24 progresses, model access will be identified through PSERC project report (to be available at www.pserc.org).
Modeling results illustrate impacts of attention to model details, e.g., level of aggregation.
19
Next example for modeling market issues: supply variability requires flexibility in dispatch of controllable resources
MW
tOperating Hour
Hour Ahead Schedule
Day Ahead Schedule
Hour AheadAdjustment
Load Following
Generation Requirement
Regulation
Hour Ahead ScheduleAnd Load Following
20
Real-Time Unit Commitment (RTUC) prepares for Real-Time Economic Dispatch
HASP = Hour Ahead Scheduling ProcessSTUC = Short Term Unit Commitment
21
Real-Time Economic Dispatch (RTED) meets energy balance in 5-minute intervals
22
Existing processes: Extremely Long-Start Commitment (ELC) provides longer horizon
Residual Unit Commitment (RUC) considers up to 18-hour start-up
ELC follows RUC, with 48-hour commitment Allows operators to consider start-up of a subset of generators
ELC still has shortcomings Adds to operator workload, may not be optimal Manual steps address cycling issues but may exacerbate over-
generation
23
Enhancements now in progress to improve process
72-hour RUC process Extends RUC to multi-day optimization, to determine whether
generators remain on-line Binding commitment & financial settlement consistent with
existing ELC rules
Flexible ramping constraint Proposal in stakeholder process: RTUC will ensure sufficient
upward ramping capability for RTED Compensation based on opportunity costs
24
CAISO’s annual prioritization of future initiatives considers future enhancements
Potential future enhancements in prioritization process include: Simultaneous RUC & market dispatch Multi-day commitment in day-ahead market Bid cost recovery changes for units running over multiple
operating days Consideration of limited run time or start-ups Unit commitment and price formation improvements
25
Illustration of analysis leading to consideration of potential changes
Example: unit commitment improvements Model from competitive path assessment, using stochastic unit
commitment in PLEXOS software
Initial results suggest analysis needed before reaching conclusions: Average cost of $54.82/MWh with standard optimization is
comparable to to $54.51 with day-ahead stochastic optimization Preliminary result shows further examination warranted: increase
in average cost to $55.32 in stochastic analysis if real-time resources restricted to those committed in standard day-ahead commitment
26
Issues of congestion management and other reliability issues in ISO markets: Data and forecasting requirements Dispatchability requirements Curtailment rules Transmission reservations and utilization Locational modeling and pricing
Supporting analyses: Meaningful compliance, transmission utilization
Additional example: dynamic transfer of renewable resources to meet Renewable Portfolio Standards
27
Transmission & congestion management proposals maximize transmission use
Energy schedule
Additional reservation
Recallable transmission
Dynamic scheduling allows (but does not require) transmission reservation exceeding expected energy – room for variable delivery.
Options for 5-minute real-time scheduling of recallable transmission allow filling unused capacity while managing congestion.
After economic dispatch, operating orders ensure compliance for reliability.
28
Source: NERC’s “Available Transfer Capability Definitions and Determination”
Transmission reservations are part of interchange scheduling
• Non-recallable reservations cannot exceed capacity
• Non-recallable schedules + recallable schedules cannot exceed capacity
• Reserved capacity may be used by recallable schedules
• Non-recallable reservations + recallable use can exceed capacity
If capacity is available, dynamic transfer can exceed initial (non-recallable) reservation, as recallable schedule.
29
Dynamic Transfer Capability study examined limits on variability of resources
• Study reviewed potential limits on transfer of variable resources while shaping and firming energy within ISO:• Impacts to existing path limits that accommodate planned hourly
variations from 20 minute ramping period• Voltage control issues• Risk to stability or excitation of low frequency modes of oscillation
ISO concluded no limits need to be applied within the ISO’s BAA at this time but will monitor operational issues and coordinate with other BAAs on regional issues affecting dynamic transfer capability. Having studied reliability issues, remaining question was
potential for market impacts
30
Market impact study: PLEXOS model establishes benchmarks and simulates market
Potential for market impacts compared four scenarios:1. Reference case, internal intermittent resources: adding wind &
solar generation within ISO isolates impact of dynamics from simply having new generation
2. Add renewable dynamic transfers without transmission reservations: adding low cost MWh without new market design features establishes additional benchmark
3. Dynamic transfers bid for transmission reservations beyond their energy schedules: represents pre-scheduling in DAM & HASP, before RT dispatch
4. RT dispatch, with hourly intertie schedules fixed from Case 3: in RT dispatch, hourly intertie schedules are limited to their pre-scheduled amounts
31
Results show very limited impact on market prices in ISO’s load aggregation areas Market prices are higher in cases 2 & 4 than in case 1 simply due to
existing limits on importing new generation into ISO
Prices in cases 2 & 4 are indistinguishable in most hours: price impacts of adding reservations are limited in frequency and size
32
New resources add to congestion, but transmission reservation has limited impact
Comparing cases 1 & 2 at Palo Verde intertie shows impact of simply adding renewable resources scheduled as imports.
Case 4 adds impact of transmission reservations, but usage in case 4 has little change in most hours from case 2.
33
Conclusions
Market models have provided insights into a variety of market design issues: Simplified WECC market model used in areas including agent
modeling of market power mitigation, technical & economic implications of environmental policies, regional power flow variations, market design for demand response & storage
Optimization for unit commitment Market impacts of changing resource portfolios
In some areas, results are preliminary and will be pursued further in future research