Pacific Gas and Electric Company Long Term Procurement Plan Proceeding Renewable Integration Model Results and Model Demonstration October 22, 2010 Workshop.

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)