outline
DESCRIPTION
Pacific Gas and Electric Company Long Term Procurement Plan Proceeding Renewable Integration Model Results and Model Demonstration October 22, 2010 Workshop. Outline. Part 1 – Review of RIM Methodology and Inputs Part 2 – Results With PG&E’s October 22, 2010 Assumptions - PowerPoint PPT PresentationTRANSCRIPT
Pacific Gas and Electric CompanyLong Term Procurement Plan Proceeding
Renewable Integration ModelResults and Model Demonstration
October 22, 2010 Workshop
2
Outline
• Part 1 – Review of RIM Methodology and Inputs
• Part 2 – Results With PG&E’s October 22, 2010 Assumptions
• Part 3 - Results With Assumptions from Different Parties
• Part 4 - Closing Thoughts
3
RIM objectives
• Understand and quantify the integration requirements and cost of higher levels of intermittent resources
• Study integration impacts under different scenarios quickly
• Transparent, user friendly model
4
RIM uses a variety of inputs to determine renewable integration requirements and costs
InputsInputs ModelModel OutputsOutputs
RenewableIntegration
Model (RIM)
Detailed profiles and variability for load &
generation
Forecast errors for load& generation
Operating Flexibility Requirements
(Reg, Load Following, Day-Ahead, Ramp)
Resources required to integrate
Intermittent renewables
Fixed and variable cost of integrationCost of conventional
resources
Installed intermittent renewable generation
To the extent possible, RIM uses the same inputs as CAISO’s studyTo the extent possible, RIM uses the same inputs as CAISO’s study
5
Intra 5-min variability
5-min forecast error
Intra-hour variability
Hour-ahead forecast error
Day-ahead forecast error
Regulation Load-following DA Commitment
RIM is a statistical model that accounts for variability and unpredictability
Minute-by-minute actual 5-minute forecast Hour-ahead forecast Day-ahead forecast
• Regulation
– RIM uses parameters that describe deviations from relevant scheduling
– Two primary parameters: intra 5-min volatility and average 5-minute forecast error
• Load following
– RIM uses parameter that describe deviations between the 5-minute and the hour-ahead schedules
– Two primary parameters: intra-hour variability and average hour-ahead forecast error
• Day-ahead commitment
– Deviation between day-ahead and hour-ahead schedule
• The model uses all 5 statistical parameters shown in diagram
6
Steps in Estimating Resource Requirements
Reliability Requirement
Operating Flexibility Requirement
Forecast Peak Load
Renewable Reliability
Contribution(NQC)
Planning Reserve Margin
Renewable Hourly
Generation
Operating Flexibility
Hourly Requirement
Projected Hourly Load
Residual Reliability
Requirement
MW
MW
Additional Capacity Required
for integration
Residual Operating Flexibility
Requirement
Forecast Peak Load+ Planning Reserve Margin– Reliability Contribution of Renewables (NQC)
Reliability Requirement
Forecast Peak Load+ Planning Reserve Margin– Reliability Contribution of Renewables (NQC)
Reliability Requirement
Hourly Load+ Hourly Operating Flexibility Services– Hourly renewable generation
Operating Flexibility Requirement
Hourly Load+ Hourly Operating Flexibility Services– Hourly renewable generation
Operating Flexibility Requirement
7
Integration costs
• Fixed Costs
• Fixed cost of resources in excess of reliability requirement
• Variable Costs
• Fuel and operating costs of resources providing flexibility services
• Emission Costs
• Emission costs based on the incremental fuel use by resources providing integration services
8
RIM’s results vary depending on inputs
• RIM is a flexible tool
• RIM’s results vary depending on inputs and assumptions used
• Range of results is illustrated by:
– PG&E’s October 22, 2010 Assumptions
– Other Parties’ assumptions and sensitivities
9
Model improvements and inputs changes implemented since Aug 25 workshop
Intra 5-min variability
5-min forecast error
Intra-hour variability
Hour-ahead forecast error
Day-ahead forecast error
Regulation Load-following DA Commitment
Minute-by-minute actual 5-minute forecast Hour-ahead forecast Day-ahead forecast
• Modified forecast errors
– Day-ahead forecast errors for load, wind, and solar from other studies*
– Hour-ahead forecast errors for load and wind from CAISO improved error data set
– 5 minute load forecast errors from CAISO improved forecast error
– 5 minute wind forecast error corrected
• Decreased service level standard deviations
• Added capability for user to exclude day-ahead commitment if desired
* Sources: “SPP WITF Wind Integration Study” by Charles River Associates, and DOE’s “Solar Vision Study” Draft May 28, 2010.* Sources: “SPP WITF Wind Integration Study” by Charles River Associates, and DOE’s “Solar Vision Study” Draft May 28, 2010.
10
Outline
• Part 1 – Review of RIM Methodology and Inputs
• Part 2 - Results With PG&E’s October 22, 2010 Assumptions
• Part 3 - Results With Other Parties’ Assumptions
• Part 4 - Closing Thoughts
11
Renewable resource scenarios
The RPS scenarios will be updated to reflect the LTPP Scoping MemoThe current scenarios are:
1. 20% Reference Case in 20202. 27.5% Reference Case in 20203. 33% Reference Case in 2020
All scenarios include additional self-gen PV treated as PV supply to capture the integration requirementAll scenarios include additional self-gen PV treated as PV supply to capture the integration requirement
Intermittent Renewable Generation Scenarios
0
5,000
10,000
15,000
20,000
25,000
2009 20% RPS 27.5% 33% RPS
PV
Solar Thermal(CST)
Wind - New
Wind - Existing
ScenariosScenarios
MW
MW
12
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
20% 27.5% 33%
Input changes decrease operating flexibility requirements
MW
MW
October 22, 2010 Operating Flexibility Requirements
(Summer Season, 2020)
August 25, 2010 Operating Flexibility Requirements
(Summer Season, 2020)
MW
MW
Operating flexibility requirements decrease by about 1,000 MW (Step 1 results) Operating flexibility requirements decrease by about 1,000 MW (Step 1 results)
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
20% 33%
DAY-AHEADCOMMITMENT
LOADFOLLOWING
REGULATION
13
Additional flexible resources are needed for integration above PRM
MW
MW
• 20% RPS Scenario need about 1,000 MW
• 27.5% RPS and 33% RPS need about 4,600 MW and 4,800 MW, respectively
* The “All Gas” Scenario does not need additional flexible resources above PRM requirement.* The “All Gas” Scenario does not need additional flexible resources above PRM requirement.
Resource Requirements for Integration (MW) by Scenario in 2020*
-
1,000
2,000
3,000
4,000
5,000
6,000
20% 27.5% 33%
14
The contribution of wind/solar to reliability affects renewable integration need
Wind/solar reliability value (NQC) is so large that the system has:
• Surplus reliability resources (i.e., it has NQC surplus), but
• Unmet operating needs to cover net load and increased flexibility
Resource Requirements for (MW) by Scenario in 2020*
(4,000)
(3,000)
(2,000)
(1,000)
-
1,000
2,000
3,000
4,000
5,000
6,000
20% 23.5% 27.5% 33%
ReliabilityRequirement
Additionalcapacity needed to meet load and flexibility need
NQC surplus
15
Shift in critical hour drives results
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Total Load
33% Generation
33% Net Load
27% Generation
27% Net Load
Renewable additions shift critical hour to hours when there is low renewable production
Net Load in 27.5% and 33% RPS Scenarios in 2020(Aug 16, 2020)
16
The bulk of the integration cost is fixed costs
• On a $/MWh basis, 27.5% RPS has higher integration costs than 33% RPS because fixed costs are divided by smaller amount of wind/solar generation
Integration Costs by Scenario, $/MWh in 2020
$/M
Wh
$/M
Wh
**$-
$5.0
$10.0
$15.0
$20.0
$25.0
20% 27.5% 33%
Emissions Cost
Variable Cost
Fixed Cost
17
Outline
• Part 1 - Review of RIM Methodology and Inputs
• Part 2 - Results With PG&E’s October 22, 2010 Assumptions
• Part 3 - Results With Other Parties’ Assumptions
• Part 4 - Closing Thoughts
18
Parties’ questions and sensitivities
• Is there a need for day-ahead commitment requirement?
• What’s the sensitivity of results with 90% vs. 95% coverage of forecast errors?
• What’s the combined effect of sensitivities TURN explored?
• What if the system can integrate 20% RPS with 15%-17% PRM in 2020?
19
-
1,000
2,000
3,000
4,000
5,000
6,000
Without Day Ahead With Day Ahead
Is there a need for day-ahead commitment requirement?
Day-ahead or multi-hour commitment is needed because more than 50% of the existing fleet requires 5 hours or more to start
If day-ahead commitment is not considered, resource need decreases by less than 1,000 MW in 33% RPS Reference Scenario
33% RPS Scenario’s Resource Requirement for Integration (MW) in 2020
20
-
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
90% 95% 99%
-
1,000
2,000
3,000
4,000
5,000
6,000
90% 95% 99%
What’s the sensitivity of results with 90% vs. 95% coverage of forecast errors?
Assuming forecast errors are normally distributed*,
• Maximum operating flexibility requirements decrease ~ 1,500 MW from 95% to 90% (Step 1 results)
• Capacity need for integration decreases by less than 500 MW from 95% to 90% (Step 2 results)
* Calculated by RIM using 2 standard deviations (for 95% coverage), and 1.65 standard deviations (90% coverage), assuming normal distribution of forecast deviations.
* Calculated by RIM using 2 standard deviations (for 95% coverage), and 1.65 standard deviations (90% coverage), assuming normal distribution of forecast deviations.
Regulation Regulation
Load following Load following
Day-ahead commitmentDay-ahead commitment
Maximum operating flexibility requirements (MW)(Step 1 Results)
Maximum operating flexibility requirements (MW)(Step 1 Results)
Capacity need for integration (MW)(Step 2 Results)
Capacity need for integration (MW)(Step 2 Results)
21
What’s the combined effect of sensitivities TURN explored?
RIM’s flexibility allows the user to test sensitivity of results
Operating Flexibility Requirements (Step 1 Results)(Summer Season, 33% RPS in 2020)
Resource Need for Integration (MW) 33% RPS Scenario in 2020 (Step 2 Results)
Oct. 22 Assumptions TURN Suggestions
Forecast error coverage 2.0 Std Deviation 1.65 Std Deviation
Day Ahead Yes No
Load DA & HA Correlation 0.5 1
Resource need for integration (Step 2 results)
4,800 MW 3,800 MW
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
Oct 22 Assumptions TURN Suggestions
DAY-AHEADCOMMITMENT
LOAD FOLLOWING
REGULATION
0
1,000
2,000
3,000
4,000
5,000
6,000
Oct 22 Assumptions TURN Suggestions
22
What if the system can integrate 20% RPS with 15%-17% PRM in 2020?
Assuming the system with 15%-17% PRM can integrate 20% RPS in 2020, resource need for 33% RPS is reduced by ~1,000 MW
4,800 MW RIM estimate for 33% RPS in 2020
- 1,100 MW RIM estimate for 20% RPS in 2020
3,700 MW RIM estimate for 33% RPS in 2020, assuming system can integrate 20% RPS with 15%-17% PRM in 2020
23
Outline
• Part 1 - Review of RIM Methodology and Inputs
• Part 2 - Results With PG&E’s October 22, 2010 Assumptions
• Part 3 - Results With Other Parties’ Assumptions
• Part 4 - Closing Thoughts
24
Insights from analysis
• Critical need hours shift from afternoon to evening
• Increased forecast uncertainty and variability also contribute to integration need/cost
• There is a substantial amount of intermittent renewable NQC that does not reduce resource need
25
Next steps
• Update RPS scenarios based on scoping memo
• Continue work with CAISO and other parties to improve model inputs and model functionality
– Calibrate balance year assumption or find simplified ways to represent existing system integration capability
– Calibrate variable integration cost inputs
• Welcome suggestions for improvements to the model
26
Appendix
27
Forecast errors and variability
Season 5-min Forecast Error INTRA 5-min Volatility HA Forecast Error INTRA-Hour Volatility Day-ahead Forecast Error
CAISO Spring 126 831Summer 126 1,151Fall 126 835Winter 126 873
RIM Spring 126 55 831 472 1,465= Summer 126 65 1,151 618 1,828
Fall 126 56 835 512 1,593Winter 126 62 873 519 1,486
CAISO Spring 5.0%Summer 4.5%Fall 4.4%Winter 4.1%
RIM Spring 1.2% 0.22% 5.0% 1.26% 9.3%Summer 0.6% 0.20% 4.5% 1.14% 9.3%Fall 0.5% 0.28% 4.4% 1.04% 9.3%Winter 0.8% 0.16% 4.1% 0.90% 9.3%
CAISO Spring 5.0%Summer 10.0%Fall 7.5%Winter 5.0%
RIM Spring 1.6% 1.0% 5.6% 7.8% 13.00%Summer 0.7% 0.6% 4.1% 6.3% 13.00%Fall 1.2% 0.8% 4.7% 7.4% 13.00%Winter 1.3% 0.8% 5.4% 6.9% 13.00%
(Standard deviation errors and variability expressed in MW)
(Standard deviation errors and variability expressed as % of installed capacity)
(Standard deviation errors and variability expressed as % of installed capacity)Clearness index (CI)
2020 Load
Wind
Solar Thermal and PV
28
Regulation-up requirements comparison
Summer - Regulation Up
0
200
400
600
800
1000
1200
1400
2009 20% Reference 33% Reference
Scenarios
MW
Fall - Regulation Up
0
200
400
600
800
1000
1200
1400
2009 20% Reference 33% Reference
Scenarios
MW
Winter - Regulation Up
0
200
400
600
800
1000
1200
1400
2009 20% Reference 33% Reference
Scenarios
MW
CAISO Max of 95% high CAISO Average of Max 95% high RIM Max
Spring - Regulation Up
0
200
400
600
800
1000
1200
1400
2009 20% Reference 33% Reference
Scenarios
MW
CAISO Max of 95% high CAISO Average of Max 95% high RIM Max
29
Load Following-up requirements comparison
Summer - Load Following Up
0
1000
2000
3000
4000
5000
6000
7000
8000
2009 20% Reference 33% Reference
Scenarios
MW
Fall - Load Following Up
0
1000
2000
3000
4000
5000
6000
7000
8000
2009 20% Reference 33% Reference
Scenarios
MW
Winter - Load Following Up
0
1000
2000
3000
4000
5000
6000
7000
8000
2009 20% Reference 33% Reference
Scenarios
MW
CAISO Max of 95% high CAISO Average of Max 95% high RIM Max
Spring - Load Following Up
0
1000
2000
3000
4000
5000
6000
7000
8000
2009 20% Reference 33% Reference
Scenarios
MW
CAISO Max of 95% high CAISO Average of Max 95% high RIM Max
30
Operating need
NQC Surplus
Incremental Net load
-6,000
-5,000
-4,000
-3,000
-2,000
-1,000
0
1,000
2,000
3,000
4,000
5,000
6,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Integration need in 27.5% vs. 33% RPS scenarios
MW
MW
August 16, 2020 Peak day for operating need - 27.5% RPS ScenarioAugust 16, 2020 Peak day for operating need - 27.5% RPS Scenario
Relative to 33% RPS Scenario, 27.5% RPS Scenario has:- Lower operating flexibility requirement, but - Higher net load (due to lower RPS generation)
Operating need
NQC Surplus
Incremental Net load
-6,000
-5,000
-4,000
-3,000
-2,000
-1,000
0
1,000
2,000
3,000
4,000
5,000
6,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24MW
MW
August 16, 2020 Peak day for operating need - 33% RPS ScenarioAugust 16, 2020 Peak day for operating need - 33% RPS Scenario
* NQC Surplus is also referred to as Residual Reliability Requirement* NQC Surplus is also referred to as Residual Reliability Requirement
Operating flexibility requirement
Operating flexibility requirement
Operating flexibility requirement
Operating flexibility requirement
Net effect is 200 MW reduction in integration need in 27.5% RPS Scenario compared to 33% RPS Scenario
4,600 MW Integration need4,600 MW Integration need 4,800 MW Integration need4,800 MW Integration need
31
Resource additions by scenario
Wind/solar additions reduce conventional gas-fired resources, primarily combined cycles
NQC of Resources by Scenario, NQC MW
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
All gas 20% 27.50% 33%
Total Wind/solar
CT
CC
32
The normal distribution (public domain image)