IRRP Stakeholder Meeting on Renewable Integration Requirements Jim - - PowerPoint PPT Presentation

IRRP Stakeholder Meeting on Renewable Integration Requirements

Jim Blatchford

• Sr. Policy Issues Rep.

Facilitator IRRP Stakeholder Meeting (Teleconference) October 20, 2009

Slide 2

Overview / Call Objective

• Provide status of ongoing efforts to assess the adequacy of the

existing fleet to manage 20% RPS (and higher RPS)

• Explain updates to the study methodology
• Discuss the draft production simulation results from the “wind only”

case, including overgeneration results

• Discuss alternative over-generation analysis
• Discuss coordinated study effort to evaluate operational and storage

requirements with KEMA/CEC

• Discuss the ISO’s development of market and operational metrics to

inform the ongoing evaluation of renewable integration

Slide 3

Call Agenda

9:00 - 9:10

Introduction Jim Blatchford

10:00 – 10:50

Updates to Production Simulation Methodology Udi Helman

10:50 - 11:15

Overgeneration Analysis Clyde Loutan

9:10 – 9:20

Status of IRRP analyses Grant Rosenblum

9:20 – 10:00

Updates to Integration Requirements Analysis KEMA/ISO Renewable Study Renewable Metrics and Next Steps Clyde Loutan

11:15 - 11:40

David Hawkins

11:40 – 12:00

Grant Rosenblum

Status of ISO Analysis of Generation Fleet Adequacy under a 20% RPS (and higher RPS)

Grant Rosenblum Manager, IRRP David Hawkins Lead Renewables Power Engineer Udi Helman, PhD Principal, Markets and Infrastructure Division Clyde Loutan Senior Advisor, Markets and Infrastructure Division

Slide 5

Status Report

Slide 6

California ISO research and simulation tools to assess integration of variable generation renewables

As system and market operator, CAISO needs accurate

assessments of the operational impacts of variable generation renewables, both to ensure reliability and to support market procurement/design to facilitate integration

CAISO research that began in 2006-7 and continues

today has sought to capture more operational and market detail than most prior studies

Several modeling and analytical efforts are underway

simultaneously

Slide 7

Why the delay in the 20% RPS fleet adequacy study?

Last stakeholder discussion on fleet adequacy study in

January 2009

Most production simulation results were done by May

2009

However, these results assumed incremental wind

resources only; during 2009, calculating the operational requirements of solar technologies became a priority

Also, need to get 33% RPS operational study underway Current presentation explains subsequent changes to

20% RPS fleet adequacy study (and uses in the 33% RPS operational study)

Slide 8

Solar PV plant output variability (partly-cloudy day, 10-second time-step)

Slide 9

April 21 - Concentrated Solar

• 50

50 100 150 200 250 300 350 0:00 0:42 1:24 2:06 2:48 3:30 4:12 4:54 5:36 6:18 7:00 7:42 8:24 9:06 9:48 10:30 11:12 11:54 12:36 13:18 14:00 14:42 15:24 16:06 16:48 17:30 18:12 18:54 19:36 20:18 21:00 21:42 22:24 23:06 23:48

Megawatts

Slide 10

April 12 - Wind + Solar

• 200.0

0.0 200.0 400.0 600.0 800.0 1,000.0 1,200.0 1,400.0 1,600.0 0:00 0:43 1:26 2:09 2:52 3:35 4:18 5:01 5:44 6:27 7:10 7:53 8:36 9:19 10:02 10:45 11:28 12:11 12:54 13:37 14:20 15:03 15:46 16:29 17:12 17:55 18:38 19:21 20:04 20:47 21:30 22:13 22:56 23:39

Megawatts

WIND SOLAR WIND + SOLAR

Solar mitigates the decline of wind generation for morning load ramp Solar shifts the ramp up of wind generation for evening load balancing

Slide 11

The Fleet Adequacy analysis currently has two key components

• 1. Simulation of renewable integration operational

requirements (2007 study methodology and updates)

• Ancillary service requirements (Regulation)
• Generic system requirements – ramp, changes in economic

dispatch

• 2. Production simulation with zonal network model
• Unit commitment and dispatch to evaluate capabilities of

generation (and non-generation) resources to integrate variable renewables

• Ability of existing fleet and additions to meet ramp requirements
• Effect on commitment and dispatch of day-ahead and hour-

ahead forecast error

Slide 12

Step 1: Analysis of Renewable Integration Operational Requirements

Slide 13

Methodology to Assess Intra-Hour Operational Requirements

• Objective is to estimate intra-hour characteristics of Regulation, 5

minute Economic Dispatch (Load Following) and ramp rate magnitude and duration

• Methodology originally used in ISO 2007 study, now updated
• Forecast actual load and renewable output under 20% RPS

Load incremented by 1.5% annually 2012 wind output based on TrueWind simulation Solar profiles under development

• Monte Carlo simulation that generates realistic hour-ahead and 5

minute-ahead load and wind forecast errors, based on statistical properties of the actual 2006 errors

autocorrelation standard deviation truncated normal distribution

Slide 14

Methodology to Assess Intra-Hour Operational Requirements (cont.)

Based on Monte Carlo simulation, the following

quantities are calculated for each interval:

5 minute economic dispatch (load following): the difference between the forecast 5 minute load (net of wind & solar) and the forecast hour-ahead load (net of wind & solar) Regulation: the difference between the actual load (net of wind & solar) and the forecast 5 minute load (net of wind & solar) Ramp rate and duration: estimated ex post using a “swinging door” algorithm (see Makarov, et al. 2009)

Slide 15

Load, MW t Hour Ahead Load Schedule Hour Ahead Load Schedule t+1

Block Hourly Load Schedules

Slide 16

The method approximates actual ISO Hour-Ahead scheduling

• Hour-ahead schedules are hourly block energy schedules

including the 20-minute ramps between hours.

• They are provided 75 minutes before the actual beginning
• f an operating hour.
• The load forecast used for the hour-ahead scheduling

process is provided 2 hours before the beginning of an

• perating hour.
• The forecast error is simulated using a TND random

number generator based on the statistical characteristics of the load forecast error (derived from 2006/2007 data)

• The hour-ahead load schedule:
• The hour-ahead wind generation schedule:
• The hour-ahead solar generation schedule:

( )

{ }

w ha w w a hr w hr ha

CAP G avg G ⋅ − ℜ =

, 1 20 1 ,

ε

( )

{ }

s ha s s a hr s hr ha

CAP G avg G ⋅ − ℜ =

, 1 20 1 ,

ε

( )

{ }

ha L a hr hr ha

L avg L

, 1 20 1 ,

ε − ℜ =

Slide 17

CAISO Scheduling Process

MW t Operating Hour

Hour Ahead Schedule Day Ahead Schedule Hour Ahead Adjustment Load Following Generation Requirement Regulation Hour Ahead Schedule And Load Following

Slide 18

Simulate Forecast Errors – Load, Wind

• The real-time and hour-ahead load and wind forecast errors are

simulated using a random number generator based on the statistical characteristics of the actual real-time and actual hour- ahead load and wind forecast error

• The distribution of forecast errors is an unbiased Truncated

Normal Distribution (TND)

• Same statistical characteristics of the forecast error will be
• bserved in the year 2012.
• A new non-linear optimization-based random number generator

is used to produce forecast errors.

PDF() 1 εmax εmin

• Average

1.1 Minimum

• 349.5

Maximum 349.4

• Std. Dev.

97.8 Autocorrelation 0.6

Table 3 Estimated Hour-Ahead Wind Generation Forecast Characteristics (in Fraction of Capacity)

Seasons Winter Spring Summer Fall Average 0.00012

• 0.0005
• 0.0005

0.0006 Minimum

• 0.3568
• 0.4331
• 0.3219
• 0.3193

Maximum 0.3092 0.3084 0.3074 0.3966

• Std. Dev.

0.0723 0.0899 0.0796 0.0792 Autocorrelation 0.6106 0.7061 0.6519 0.5939

Table 1 Real-time Load Forecast Characteristics Table 2 Hour-Ahead Load Forecast Characteristics of the Yr. 2006 (in MW)

Seasons Winter Spring Summer Fall Average

• 22.49
• 24.05
• 130.43
• 69.21

Min

• 2680.12
• 2101.08
• 3770.73
• 2627.90

Max 1842.06 1930.54 2446.12 2080.98

• Std. Dev.

637.37 601.34 900.13 687.52 Autocorrel ation 0.70 0.73 0.89 0.83

min( ( , , , , ))

e

f a b c d σ

Slide 19

Simulate Forecast Errors – Solar

• The clearness index (CI) for a given period is obtained by dividing the observed global

radiation Rg by the extraterrestrial global irradiation R: k = Rg/R

• where Rg is the horizontal global solar radiation, R is horizontal extraterrestrial solar

radiation.

• If the weather condition of a day is like between a sunny day and a very cloudy day,

the standard deviation of the solar forecast errors will vary. Thus, the standard distribution of the solar forecast errors can be described as a function of a parameter ξ, .

k σ

1 Fig.9. Clearness index v.s. standard deviation of solar forecast errors.

Clearness Index and Std. Dev. Of solar forecast

Slide 20

Simulate Forecast Errors – Solar

Daily pattern of the solar radiation of clearness index. Fig.8. Distribution of solar forecast error in very cloudy day and a very sunny day. Forecast Error

Slide 21

‐Pmax ε Pmax ε

Changes of the solar irradiance error depending the clearness index.

Sunny Day Cloudy Day Probability Sunny Day Probability Cloudy Day Forecast Error Forecast Error

Slide 22

Five Minute Economic Dispatch (Load Following) Requirement shown as blue shaded area

t MW

Economic Dispatch (Load Following) Actual Load Hourly Schedule 5-Minute Schedule

Slide 23

Changes in Five Minute Economic Dispatch/Load Following Capacity -- Results will be similar to the 2007 study shown below (results shown are for incremental wind resources)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 2200 2400 2600 2800 3000 3200 3400 3600 Load Following, Summer, Year 2010 vs.2006 Max Load Following Capacity, Inc, MW 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 −3600 −3400 −3200 −3000 −2800 −2600 −2400 −2200 Hours Ending Max Load Following Capacity, Dec, MW

The maximum upward capacity requirement of 3,500 MW

• ccurs during HE3 and HE11

The maximum increase of 800 MW occurs during HE3 (3,500 – 2,700) The maximum downward capacity requirement of 3,450 MW occurs during HE24 The maximum downward capacity increase of 500 MW (3,050 -2,450) occurred in HE22

Slide 24

Load Following Ramping Requirement -- Results will be similar to the 2007 study shown below (results shown are for incremental wind resources)

It is expected that both the maximum upward and downward load following ramping requirements in 2010 will increase by 40 MW/min.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 50 100 150 200 250 Load Following, Summer, Year 2010 vs.2006 Max Load Following Ramp Rate, Inc, MW/min 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 −180 −160 −140 −120 −100 −80 −60 −40 Hours Ending Max Load Following Ramp Rate, Dec, MW/min

Slide 25

Load Following Ramp Duration

The upward and downward ramp durations are required for approximately 30 and 20 minutes, respectively.

5 10 15 20 25 30 20 40 60 80 100 120 Max Ramp Rates, Inc/Up, MW/min Load Following, Summer, Year 2010 vs. 2006, Hour 8, 9, 10 5 10 15 20 25 30 −200 −150 −100 −50 Duration (min) Max Ramp Rates, Dec/Down, MW/min 2006 Non Wind 2006 Wind 2010 Non Wind 2010 Wind

Slide 26

Regulation Requirement shown as red shaded area

t MW

Economic Dispatch (Load Following) Actual Load Hourly Schedule 5-Minute Schedule Regulation

Slide 27

CAISO Real-Time Scheduling

• The real-time dispatch is automatically conducted by the CAISO’s market

applications using 15-minute intervals for RTUC and 5-minute interval for RTED

• The desired changes of generation are determined in real-time for each 5-minute

dispatch interval 5 minutes before the actual beginning of the interval.

• System information used for dispatch is dated back 7.5 minutes before the

beginning of the interval.

• Units start to move toward the new set point 2.5 minutes before the interval begins.
• They are required to reach the set point in the middle of the interval (2.5 minutes

after its beginning).

• The real-time load forecast:
• Real-time wind forecast (naïve persistence model):
• Real-time solar forecast (naïve persistence model):

{ }

rtf L a rtf

L avg L

, min 5 5 min 5 ,

) ( ε − ℜ =

] 8 [ ] 5 , [

min 5 ,

− = +

−

t G t t G

w a w rtf

] 8 [ ] 5 , [

min 5 ,

− = +

−

t G t t G

s a s rtf

Slide 28

Load, MW t 5 Minute Dispatch Interval Real Time Load Schedule 5 Minute Ramps Actual Load Average Actual Load Forecast Error

5-Minute Dispatch

Slide 29

The Real Time Economic Dispatch software runs every five-minutes starting at approximately 7.5 minutes prior to the start of the next Dispatch Interval and produces Dispatch Instruction for Energy for the next Dispatch Interval and advisory Dispatch Instructions for as many as 13 future Dispatch Intervals.

Dispatch Range MW t t-2.5 t+2.5 t+5 t+7.5 t+10 Minutes Interval 1 Interval 2 Run starts Here for Interval 2 ADS Instruction Sent for Interval 2 Units begin in move to DOT in interval 2 10 minutes A B C D E

Market timelines benefit renewable integration

Slide 30

Regulation capacity requirements evaluated by hour -- Results will be similar to the 2007 study shown below (results shown are for incremental wind resources)

The maximum increase

• f 230 MW occurs

during HE9 (480 MW– 250 MW) The maximum downward increase of 500 MW (750 MW -250 MW) occurred in HE18

Slide 31

Implications of Results for Markets and Need for Further Analysis

Procurement of Regulation (Ancillary Services) should

increase by season and hour in day-ahead market (or after day-ahead market, depending on the ISO)

Increased ramp requirements, particularly in morning

and evening hours

Results are determined independent of particular

commitment or dispatch of generation

Additional studies are needed to verify that actual fleet in

simulated years (e.g., 2012) can provide ramp and load following capabilities (next presentation)

Slide 32

Step 2: Production Simulation

Slide 33

Step 2: Link Results of Step 1 with a Unit Commitment Production Simulation to verify Capabilities of Generation Fleet (and Non-Generation Resources)

Evaluate 2012 CAISO generation resources to

determine their ability to reliably integrate anticipated levels of variable renewable resources

Focus on the ability of CAISO fossil-fired resources to provide sufficient flexibility Evaluate the impact of day-ahead and hour-ahead forecast errors on commitment and dispatch Determine the magnitude and frequency of any system

• perational violations

Test (or extend) the ability of readily available analytical

tools to provide credible integration evaluations

Scalable, repeatable

Slide 34

Production Simulation Vendor For This Phase – PLEXOS

Unit commitment, production cost model Can represent zonal or detailed network representation Hourly and 10-minute simulation time steps Can approximate CAISO market design and procedures

with respect to

Simulated day-ahead, hour-ahead and real-time dispatch solutions Dynamic co-optimization of energy and AS (i.e. simultaneous solution) Locational prices for energy (if needed)

Slide 35

Modeling Assumptions for 2012 Simulations

Only CAISO system modeled

Zonal topology initially (NP15, SP15) CAISO Master File confidential generation data (Pmin, Pmax;

• Min. up- and down time; Ramp rates; AS Ranges)

Hourly hydro generation (2006 and 2007) and AS contribution (2006) is fixed at the station-level based historical records Hourly net interchange for NP15 and SP15 fixed based on 2006

• r 2007 actuals

No AS provision assumed from imports Hourly wind, QF, and geothermal generation is based on the 2006 historical profiles; solar profiles under development 2012 generation resource additions included

Slide 36

Potential Violations Evaluated*

• 1. Regulation-Up
• 2. Regulation-Down
• 3. Spin
• 4. Non-Spin
• 5. Unserved Energy
• 6. Over-generation

* Either insufficient ramping capability or insufficient available capacity results in one of the above violations.

Slide 37

2012 Ancillary Service Requirements reflect Operational Study Results

Parameter Units Incremental Wind Requirement Wind + Solar Requirement Reg-Down MW 350-750 *

TBD *

Reg-Up MW 350-530 *

TBD *

Spin MW 0.5*(3%*L + 3%*G) 0.5*(3%*L + 3%*G) Non-Spin MW 0.5*(3%*L + 3%*G) 0.5*(3%*L + 3%*G)

* Regulation MW vary by TOD and season

Slide 38

Summary of Simulation Methodology for this Study Phase – Three Steps

• 1. 2012 all hours “day-ahead” (DA) unit commitment and

dispatch simulation on hourly time-step with stochastic modeling of load and wind incorporating day-ahead forecast errors

• 2. 2012 all hours “hour-ahead” (HA) unit commitment and

dispatch simulation on hourly time-step with stochastic modeling of load and wind incorporating hour-ahead forecast errors

• 3. Selected days/hours in 2012 subject to sequential DA-HA-

Real-Time unit commitment and dispatch sequence; RTD is conducted on ten-minute time-step

Slide 39

Summary of Simulation Methodology for this Study Phase – Structure of Analysis

[Real Time] “Actual” Wind/Solar Output and Load

• n 10 minute time-step

[Hour-Ahead] “Actual” Wind/Solar Output and Load + HA Forecast Error on hourly time-step [Day-Ahead] “Actual” Wind/Solar Output and Load + DA Forecast Error on hourly time-step

Slide 40

Summary of Inputs and Stochastic Modeling

Inputs

2012 “actual”/real-time load forecasts for IOU’s (based on 2006, 2007) 2012 “actual”/real-time wind and solar generation forecasts for 5 zones 2012 supply = hourly hydro, QF, Geothermal, import and export profiles based 2006 or 2007 historical hourly generation; new resource additions

Simulation mode - 100-iterations Stochastic drivers

Convergent Monte Carlo for generator forced outage modeling 2012 hourly DA/HA load forecasts with forecast deviations modeled with stochastic process with parameters derived from the 2006 and 2007 historical hourly DA/HA load forecast errors by season 2012 hourly DA/HA wind and solar generation forecasts with forecast deviations modeled with stochastic process with parameters derived from the 2006 and 2007 historical hourly DA/HA wind generation forecast errors

Slide 41

Purpose of Different Simulations and Interpretation

• f Results

DA and HA “uncoupled” hourly simulations –

“Screen” for hours that might have RTD violations; so examine selected days/hours with DA/HA violations Suggest frequency/magnitude of violations over year (duration curve) as seen from several hours forward 100 iterations

Caveats –

DA/HA frequency/magnitude results need to be checked through RTD simulations Screen will not reflect RT hours with possible violations that are missed through hourly averaging and forecast error

Slide 42

Purpose of Different Simulations and Interpretation

• f Results (cont.)

DA-HA-RTD “coupled” simulations –

Assess whether fleet can resolve forecast violations in DA and HA simulations in RTD Assess whether fleet will encounter violations in RTD not forecast in DA and HA simulations Correct for effects of hourly averaging on DA-HA unit commitment

Caveats –

Does not reflect forecast uncertainty during dispatch hour

Slide 43

Draft Results of Incremental Wind Resources Only to Meet 20% RPS

The first phase of analysis evaluated additional wind

resources to meet the 20% RPS (consistent with ISO’s 2007 renewable integration study)

Draft results are discussed in next slides

Slide 44

DA Simulation Annual Unserved Energy Duration Curve (2006-based simulation)

Duration Curve of Unserved Energy (MW) 2012 (2006-based)

200 400 600 800 1,000 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Hours MW

DRAFT RESULTS

Slide 45

DA Simulation Annual Over-generation Duration Curve (2006-based simulation)

Duration Curve of Over-Generation (MW) 2012 (2006- based)

500 1,000 1,500 2,000 2,500 0.0 0.6 1.1 1.6 2.2 2.7 3.3 3.8 4.3 4.9 5.4 5.9 6.5 7.0 7.6 8.1 8.6 9.2 9.7 Hours MW

DRAFT RESULTS

Slide 46

DA Simulation Annual Regulation-up Shortfall Duration Curve (2006-based simulation)

Duration Curve of Reg-up Shortfall (MW) 2012 (2006- based) 100 200 300 . 1 . 2 . 3 . 4 . 5 . 6 . 7 . 8 . 9 . 1 . 1 1 . 1 2 . 1 3 Hours MW DRAFT RESULTS

Slide 47

DA-HA-RTD Duration Curves of Over- generation on April 17, 2012 (2006-based simulation)

Duration Curve of Over-Generation (MW)

• n April 17, 2012 (2006-based)

500 1,000 1,500 2,000 2,500 0.0 0.1 0.3 0.4 0.5 0.6 0.7 0.8 1.0 1.1 Hours MW Over-Gen in DA Over-Gen in HA Over-Gen in RTD DRAFT RESULTS

Slide 48

Summary of Violation Occurrences in the DAM-HAM-RTD Simulations

Violation Occurrences from 100-iteration Simulations Service Overgeneration Reg-up shortfall Unserved Energy Date Market DAM HAM RTD DAM HAM RTD DAM HAM RTD February 27, 2012 2006-based April 17, 2012 2006-based 99 49 105 May 7, 2012 2006-based 108 82 8 June 24, 2012 2006-based 1 July 23-24, 2012 2006-based 6 2 5 September 3, 2012 2006-based February 27, 2012 2007-based July 3, 2012 2007-based August 30, 2012 2007-based 5 2 3 2

DRAFT RESULTS

Slide 49

Current phase of production simulation

Evaluate a renewable resource mix more consistent with

current IOU contracts/short-listed projects

Add solar thermal/solar PV production profiles consistent

with CPUC RPS contracts/IOU short-listed projects

Also add a benchmark “all-gas” case for evaluating

integration cost changes and impacts on generators

Changes in # starts/stops Changes in hours at Pmin Changes in cycling and ramping

Slide 50

Next phase of production simulation research

ISO is sponsoring in-house research that would further

include

Full network model (DC power flow) Additional hydro flexibility Additional granularity on commitment decisions (more reflective

• f actual market procedures)

Quantify market benefits of improved wind and solar forecasts

ISO is also evaluating more realistic simulation tools to

evaluate operational impacts of incremental renewable resource additions (e.g., year by year)

Slide 51

Over-Generation Analysis

Slide 52

Over-Generation Analysis

Identify and quantify the frequency, duration and

magnitude of over-generation

Sensitivities

• High Hydro
• Low Hydro

Slide 53

• System frequency is higher than 60 Hz,
• Area Control Error (ACE) is higher than normal and potential control

performance violations can occur,

• Grid operators have difficulties controlling the system due to

insufficient regulating capacity,

• Potential inability to quickly arrest frequency decline following a

disturbance,

• Excess energy flows to neighboring BAs as inadvertent energy, and
• Real-time energy market prices may be negative

Typical concerns during over-generation

Slide 54

Approaches to Over-Generation Analysis

• 1. Production Simulation (already discussed)
• Could underestimate actual frequency and magnitude due to

better optimization of system resources than is possible in actual operations

• 2. Extrapolation from historical experience
• Using statistical analysis, historical generation by technology,

and assumptions about thermal and hydro min gen, imports, etc.

• Could overestimate actual frequency and magnitude by not fully

accounting for dynamic optimization

• Note: neither approach to date considers impact of

dispatch of wind resources, storage or demand response; future simulations may conduct such sensitivities

Slide 55

Average Production by Resource Type (2006 Actual

• vs. Expected 2012 levels)

Nuclear 2,522 4,500 4,526 4,500 3,620 4,500 3,468 4,500 Hydro 3,823 3,500 2,707 2,700 1,009 1,000 2,337 2,000 Thermal 3,822 3,100 4,325 4,400 5,573 4,500 5,263 4,000 Qualifying Facilities 3,339 3,000 4,021 4,000 4,238 4,000 3,651 3,500 Geothermal 783 1,200 789 1,200 794 1,200 800 1,200 Imports 5,149 4,000 5,511 4,500 4,744 4,200 4,630 4,000 Total Generation plus Interchange 19,438 19,300 21,879 21,300 19,978 19,400 20,149 19,200 Average Wind - 2006 711 1,043 430 324 Average Load - 2006 20,149 22,922 20,408 20,473 Minimum Load 19,064 20,800 20,837 22,800 19,189 21,000 18,737 20,500 Maximum Renewable that can be Integrated - 2012 1,500 1,500 1,600 1,300

High Hydro --- Average Generation by Technology

Spring Summer Fall Winter 2006 MW 2012 MW 2006 MW 2006 MW 2006 MW 2012 MW 2012 MW 2012 MW

Slide 56

Production by Resource Type (2007 Actual vs. Expected 2012 levels)

Nuclear 4,279 4,500 4,179 4,500 3,886 4,500 4,196 4,500 Hydro 1,108 1,100 1,392 1,400 879 900 811 800 Thermal 3,107 4,000 4,140 4,400 5,587 4,100 5,667 4,000 Qualifying Facilities 3,744 3,700 4,305 4,200 4,125 4,000 4,402 4,000 Geothermal 801 1,200 827 1,200 816 1,200 798 1,200 Imports 7,023 4,600 6,947 5,400 5,065 4,500 4,598 4,500 Total Generation plus Interchange 20,062 19,100 21,790 21,100 20,358 19,200 20,472 19,000 Average Wind - 2007 823 1,087 308 386 Average Load - 2007 20,885 22,877 20,666 20,858 Minimum Load 19,699 20,800 21,020 22,800 19,630 21,000 19,414 20,500 Maximum Renewable that can be Integrated - 2012 1,700 1,700 1,800 1,500

Low Hydro --- Average Generation by Technology

Spring Summer Fall Winter 2007 MW 2012 MW 2007 MW 2007 MW 2007 MW 2012 MW 2012 MW 2012 MW

Slide 57

Actual Wind Production Distribution

Summer 2006 - Minimum 10% of Load

2 4 6 8 10 12 14 16 18 20 22 24 26 100 200 300 400 500 600 700 800 900 1,000 1,100 1,200 1,300 1,400 1,500 1,600 1,700 1,800

Wind Production (MW) # of Occurrences

Slide 58

2 2 , 9 2 3 , 2 2 3 , 5 2 3 , 8 2 4 , 1 2 4 , 4 2 4 , 7 2 5 , 2 5 , 3 2 5 , 6 2 5 , 9 2 6 , 2 2 6 , 5 2,000 2,800 3,600 4,400 5,200 0.5 1 1.5 2 2.5 3

# of Occurrences

Load (MW)

Wind (MW)

Summer 2012 --- Load vs. Wind

Minimum 10% of Load

Projected Load vs. Wind for summer 2012

Slide 59

Summer 2012 --- Load vs. Wind

Minimum 10% of Load

22,000 22,400 22,800 23,200 23,600 24,000 24,400 24,800 25,200 25,600 26,000 26,400 26,800 27,200 300 600 900 1,200 1,500 1,800 2,100 2,400 2,700 3,000 3,300 3,600 3,900 4,200 4,500 4,800

Wind Production (MW) Load (MW)

Expected Curtailment in 2012 shown in Grey

Slide 60

Estimated Energy Curtailment

Summer 2012 --- High Hydro

10 20 30 40 50 60 70 80 90 100 110 120 100 300 500 700 900 1,100 1,300 1,500 1,700 1,900 2,100 2,300 2,500 2,700 2,900 3,100 3,300 3,500

Curtailment (MW) # of Hours

Min Wind 1,000 MW Min Wind 1,500 MW Min Wind 2,000 MW

Expected Curtailment under High Hydro Conditions

Slide 61

Expected Curtailment High Hydro --- 20% RPS

Spring Summer Fall Winter

Maximum Wind (MW) Expected Curtailed (Hrs) Maximum Wind (MW) Expected Curtailed (Hrs) Maximum Wind (MW) Expected Curtailed (Hrs) Maximum Wind (MW)) Expected Curtailed (Hrs)

Total (Hrs) 1,000 130 1,000 112 1,100 57 800 34 333 1,500 91 1,500 73 1,600 25 1,300 12 201 2,000 46 2,000 46 2,100 6 1,800 4 102

Expected Curtailment High Hydro --- 20% RPS

Spring Summer Fall Winter

Maximum Wind (MW) Expected Curtailed (MWh) Maximum Wind (MW) Expected Curtailed (MWh) Maximum Wind (MW) Expected Curtailed (MWh) Maximum Wind (MW) Expected Curtailed (MWh)

Total (MWh) 1,000 116,600 2.74% 1,000 120,000 2.17% 1,100 30,700 0.83% 800 17,700 0.53% 285,000 1,500 60,300 1.42% 1,500 71,400 1.29% 1,600 9,300 0.25% 1,300 5,400 0.16% 146,400 2,000 24,100 0.57% 2,000 41,400 0.75% 2,100 1,600 0.04% 1,800 1,200 0.04% 68,300

Expected curtailment during high hydro conditions

Slide 62

Estimated Energy Curtailment

Summer 2012 --- Low Hydro

10 20 30 40 50 60 70 80 90 100 100 300 500 700 900 1,100 1,300 1,500 1,700 1,900 2,100 2,300 2,500 2,700 2,900 3,100 3,300 3,500

Curtailment (MW) # of Hours

Min Wind 1,000 MW Min Wind 1,500 MW Min Wind 2,000 MW

Expected Curtailment under Low Hydro Conditions

Slide 63

Expected Curtailment Low Hydro --- 20% RPS

Spring Summer Fall Winter

Minimum Wind (MW) Expected Curtailed (Hrs) Minimum Wind (MW) Expected Curtailed (Hrs) Minimum Wind (MW) Expected Curtailed (Hrs) Minimum Wind (MW) Expected Curtailed (Hrs)

Total (Hrs) 1,200 130 1,200 86 1,300 41 1,000 26 283 1,700 64 1,700 54 1,800 16 1,500 10 144 2,200 30 2,200 36 2,300 2 2,000 3 71

Expected Curtailment Low Hydro --- 20% RPS

Spring Summer Fall Winter

Minimum Wind (MW) Expected Curtailed (MWh) Minimum Wind (MW) Expected Curtailed (MWh) Minimum Wind (MW) Expected Curtailed (MWh) Minimum Wind (MW) Expected Curtailed (MWh)

Total (MWh) 1,200 80,300 1.88% 1,200 88,100 1.66% 1,300 19,800 0.54% 1,000 11,500 0.35% 199,700 1,700 35,700 0.84% 1,700 52,000 0.94% 1,800 5,100 0.14% 1,500 3,100 0.09% 95,900 2,200 12,400 0.29% 2,200 28,600 0.52% 2,300 700 0.02 % 2,000 400 0.01 % 42,100

Expected Curtailment under Low Hydro Conditions

Slide 64

Overview of KEMA Renewables Project

Slide 65

• 1. Simulate and analyze the impact of renewable

generation on system performance with respect to AGC / system regulation and balancing energy / real time dispatch requirements and energy storage requirements.

• Validate the KEMA dynamic simulation model
• 2. Measure the impacts of increasing percentages of

renewables on the California Grid and how storage can be utilized to mitigate those impacts

• Identify potential changes to the market systems and dispatch

algorithms required to accommodate energy storage

Objective of the KEMA Renewables Project

Slide 66

Study process and timeline

R&D/PIER project funded by the CEC

Work started in June 2009

Major Tasks

Task 1 – Calibrating KEMA Simulation model to California ISO Task 2 – Define Simulation Scenarios (June – July 2009) Task 3 – Run Simulation Scenarios (July – Sept. 2009)) Task 4 – Analyze the Results (Aug. - Sept 2009) Task 5 – Prepare Final Report (Oct. 2009)

Slide 67

Scenarios selected

5 “Interesting” days selected for the scenario analysis

June 5, 2008 – Major generator trip – used to calibrate model July 9, 2008 – High summer load day with significant wind and solar generation October 20, 2008 – Fall load day with variable gen. February 9, 2009 – Winter load day with variable gen. April 12, 2009 – Spring load day, post MRTU, with variable gen.

Actual hourly data used for

Generator production schedules and Import schedules

4 second data used for

ACE, Frequency, Load and Regulation

Slide 68

Preliminary results to date

Results correlate with results from other studies Large solar ramps, both up and down, are going to be a

major operating issue

Large storage (2-4 hours) with fast response mitigate the

large ramps and control ACE

Large amounts of regulation by itself does not solve the

ramping problem for 33 % Renewables

Slide 69

Renewable Metrics and Conclusions

Slide 70

ISO is developing daily metrics relevant to the impacts on thermal fleet of renewable integration

Load/Wind, Load/Solar, Solar/Wind correlations Wind/solar ramps (by hour, 10 minutes) Max Upward/Downward ramps ACE/Regulation/Wind profile Several others

These metrics will be monitored over time and compared to simulated results

Slide 71

Sample Day – Frequency of 10-minute Wind/Solar ramps

Frequency of 10-min Wind/Solar Ramps

10 20 30 40 50 60 70 80 90 100

• 100
• 80
• 60
• 40
• 20

20 40 60 80 100 MW # of Occurences Wind Solar

Slide 72

Sample Day – Wind/Solar Hourly Ramps

Wind/Solar Hourly Ramps

• 200
• 150
• 100
• 50

50 100 150 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/hr W in d S

• la

r

Slide 73

Sample Day – ACE/Regulation/Wind profile

ACE/Regulation/Wind Profile

• 700
• 600
• 500
• 400
• 300
• 200
• 100

100 200 300 400 500 600 700 00:00 01:15 02:30 03:45 05:00 06:15 07:30 08:45 10:00 11:15 12:30 13:45 15:00 16:15 17:30 18:45 20:00 21:15 22:30 23:45 ACE/Regulation (MW) 100 200 300 400 500 600 700 800 900 Wind (MW) ACE Regulation Wind

Slide 74

Hourly Wind Statistics

500 550 600 650 700 750 800 850 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3

MW

Hourly wind statistics

Slide 75

Sample Day – Max Upwards and Downwards Ramps

Maximum Upward Ramp 3-Hour 1-Hour 10-min. 5-min 6,255 3,053 826 442 4:20 5:55 5:56 5:59 218 157 54 37 19:41 20:38 22:29 14:37 6,175 3,079 813 440 4:20 5:55 5:56 5:59 Maximum Downward Ramp 3-Hour 1-Hour 10-min. 5-min 5,813 2,224 458 325 20:52 21:58 22:25 3:44 243 162 70 43 16:30 18:34 21:59 14:33 5,933 2,271 495 322 20:52 22:09 22:23 3:44

**** Time indicates the start of respective ramps.

Wind (MW) Load-Wind (MW) Load (MW) Wind (MW) Load-Wind (MW) Load (MW)

Slide 76

Concluding comments and opportunities for feedback

Implications for procurement of Regulation (Ancillary

Services), and how markets and forward procurement processes elicit needed generation and non-generation resource characteristics, are being further assessed

Operational tools utilizing aspects of these Regulation and ramp forecasting methods may be incorporated into market procedures

Comments or questions on this presentation are

welcome

Draft report will offer opportunity for detailed comments

Slide 77

Other CAISO Initiatives Related to Fleet Adequacy

33% RPS operational study Integration of Demand Response Integration of Storage Resources Analysis of Plug-in Hybrid Electric Vehicles (PHEV) Smart Grid

Slide 78

Some References

California ISO, Integration of Renewable Resources,

November 2007 (available at www.caiso.com)

Makarov et al., Operational Impacts of Wind Generation

• n California Power Systems, IEEE Transactions on

Power Systems, 24, 2, May 2009

California ISO, Integration of Renewable Resources

Program: http://www.caiso.com/1c51/1c51c7946a480.html