Lessons Learned in Wind Integration 2009 Southwest Renewable Energy Conference Flagstaff, AZ September 11, 2009 J. Charles Smith Executive Director UWIG
What is UWIG? Non-profit corporation established by 6 utilities in 1989 with support from EPRI and DOE/NREL Over 150 members, including utilities, developers, manufacturers, consultants, government organizations Focus on technical issues Mission: To accelerate the development and application of good engineering and operational practices supporting the appropriate integration of wind power into the electric system SWREC AZ 09 -- 2
Outline of Topics Lessons from Recent Studies Wind Forecasting Capacity Value Energy Storage System Stability and Grid Codes Conclusions and Recommendations SWREC AZ 09 -- 3
It s All About Dealing with Variability and Uncertainty Variability Load varies by seconds, minutes, hours, by day type, and with weather Supply resources may not be available or limited in capacity due to partial outages Prices for power purchases or sales exhibit fluctuations Uncertainty Operational plans are made on basis of best available forecasts of needs; some error is inherent Supply side resource available with some probability (usually high) Key questions How does wind generation affect existing variability and uncertainty What are the costs associated with the changes What does the future hold SWREC AZ 09 -- 4
Flexibility Supply Curve High Cost Study needed to determine shape of Flexibility Supply Curve and Quantify Costs Low Cost High Flexibility Low Flexibility SWREC AZ 09 -- 5
Impacts Within the Hour QUESTIONS Is control area demand more variable within the hour with wind? What are the consequences? ANSWERS Effect is confined mostly to tail Impact on CPS2 Increase of 1-2 MW/min in loadfollowing capability SWREC AZ 09 -- 6
Interestingly Generators Do Not Appear To Command A Premium For Sub-Hourly Response ISO Day-Ahead $/MWH Hour-Ahead $/MWH 5-Minute $/MWH Average Within-Hour 5-Minute Range $/MWH NYISO $67.70 $64.93 $63.31 $91.18 ISO-NE $81.38 $80.76 $81.22 $24.40 CAISO $69.78 $68.32 $59.87 ERCOT 1 $71.69 $40.00 MISO $49.99 $48.62 $48.71 $67.75 1 ERCOT currently operate a 15 minute sub-hourly market rather than a 5 minute market. Average day-ahead, hour-ahead, and 5-minute prices are nearly equal 5-minute price is often slightly lower No premium for flexible generation Within hour 5-minute price range is very large Marginal generators receive a strong signal to move within the hour Source: Brendan Kirby, UWIG 2009 Spring Workshop SWREC AZ 09 -- 7
20% Wind Energy Can Be Managed Unit Commitment Costs $5.00 Integration Cost ($/MWH Wind Energy) $4.00 $3.00 $2.00 $1.00 $- 2003 2004 2005 15% Wind 20% Wind 25% Wind Penetration Level Wind integration costs for three penetration levels and pattern years. Cost of incremental operating reserves is embedded. Source: MN DOC SWREC AZ 09 -- 8
Time Scales of Interest System Load (MW) Time (hour of day) 0 4 8 12 16 20 24 Days Unit Commitment seconds to minutes Regulation tens of minutes to hours Load Following day Scheduling SWREC AZ 09 -- 9
Some Additional Reserves May Need to be Committed Reserve Category Base 15% Wind 20% Wind 25% Wind MW % MW % MW % MW % Regulating 137 0.65% 149 0.71% 153 0.73% 157 0.75% Spinning 330 1.57% 330 1.57% 330 1.57% 330 1.57% Non-Spin 330 1.57% 330 1.57% 330 1.57% 330 1.57% Load Following 100 0.48% 110 0.52% 114 0.54% 124 0.59% Operating Reserve Margin 152 0.73% 310 1.48% 408 1.94% 538 2.56% Total Operating Reserves 1049 5.00% 1229 5.86% 1335 6.36% 1479 7.05% Estimated Operating Reserve Requirement for MN BAs 2020 Load Source: MN DOC SWREC AZ 09 -- 10
Summary of Reserve Requirements Increase in reserve requirement Increase as % of wind capacity 10 % 9 % 8 % 7 % 6 % 5 % 4 % 3 % 2 % 1 % Four hours ahead Germany, Minnesota day-ahead others in-hour 0 % 0 % 5 % 10 % 15 % 20 % 25 % 30 % Wind penetration (% of gross demand) Nordic 2004 Finland 2004 Sweden Ireland 1 hour Ireland 4 hours UK Sweden (4 hours) dena Germany Minnesota 2006 Source: IEA Task 25 Report Different time scales for estimating the reserve requirement in-hour, 4 times ahead, day-ahead Less than 5 % of installed wind capacity - nearly 10 % if day-ahead forecast errors are left to short term reserves SWREC AZ 09 -- 11
Increased Balancing Cost Increase in balancing cost Euros/MWh wind 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0 % 5 % 10 % 15 % 20 % 25 % 30 % Wind penetration (% of gross demand) Nordic 2004 Finland 2004 UK Ireland Colorado Minnesota 2004 Minnesota 2006 California Greennet Germany Greennet Denmark Greennet Finland Greennet Norway Greennet Sweden SWREC AZ 09 -- 12
Forecasting and Balancing Markets Reduce Impacts SWREC AZ 09 -- 13
Different Forecasts for Different Time Periods Situational awareness forecast: used for severe weather events (real-time) Hour ahead forecast: uses rapid update cycle to produce 10 min forecasts 4-6 hrs ahead, updated every hour Day ahead forecast: Hourly forecasts 2-4 days ahead, updated every 12 hours, uses national weather service models Nodal forecast: hourly forecast of transmission system nodal injections for managing transmission congestion Different performance metrics for different forecasts Need solar data! SWREC AZ 09 -- 14
How Good is the Forecast? Wind plant output can be forecast within some margin of error, and forecasts are getting better Hour Ahead Energy (% actual) Capacity (% rated) Forecast Error Single Plant Large Region 10-15% 6-11% 4-6% 3-6% Day Ahead Hourly Energy (% Actual) Hourly Capacity (% Rated) 25-30% 10-12% 15-18% 6-8% SWREC AZ 09 -- 15
What If the Wind Stops Blowing Everywhere at the Same Time? Meso-scale wind forecasting techniques provide the answer Significant benefit to geographical dispersion Dispersion provides smoothing in the long term Aggregation provides smoothing in the short term Extensive modeling studies have shown no credible single contingency leading to simultaneous loss of capacity in a broad geographical region SWREC AZ 09 -- 16
The Power of Aggregation Source: Thomas Ackermann, Energynautics SWREC AZ 09 -- 17
What To Do When the Wind Doesn t Blow Good question! Must deal with energy resource in a capacity world Dealt with through probabilistic reliability methods used to calculate Effective Load Carrying Capability (ELCC) Contribution may be large (40%) or small (<5%) Once the ELCC is determined, get on with the job of designing a reliable system And that means adding more flexible capacity in the future! SWREC AZ 09 -- 18
An Energy Resource in a Capacity World 45% 40% 35% ELCC as % Rated Capacity 30% 25% 20% 15% 10% 5% 0% NYSERDA Onshore NYSERDA Offshore MN/Xcel(1) CO Green MN/Xcel(2) MN 2006 PacifiCorp CA/CEC SWREC AZ 09 -- 19
What About Energy Storage? Valuable component of a power system, can provide many benefits Greatest value when operated for benefit of entire system, not dedicated to a single resource One of many sources of flexibility available to the system Expensive, and benefits accrue to different parties, i.e. generation owner, trans. system operator, power marketer Seldom sufficient value in revenue stream for any single party to justify the investment Integration studies do not show need for storage at 20% wind except possibly on small, isolated systems SWREC AZ 09 -- 20
Value of Electrical Storage - Wind Medium run ~ 80% Incumbent Generation High Diversity Wind Data Source: LBNL SWREC AZ 09 -- 21
Won t Too Much Wind Power Cause the System to Collapse? Often comes up as a question after a system disturbance resulting in a blackout Related questions about system stability are driving world-wide wind turbine and wind plant model development and verification efforts (IEEE, UWIG, WECC, manufacturers, TSOs, utilities) Detailed simulations of DFIGs shows that wind plants can actually aid system stability by providing LVRT and dynamic var support to reduce voltage excursions and dampen swings SWREC AZ 09 -- 22
System Stability Case Study Wind integration and interconnection study conducted by GE for NYISO, supported by NYSERDA Looked at impacts of 3,300 MW of wind generation on 33,000 MW peak load system (10%) Stability case study investigated differences in behavior with 3,300 MW of wind plant with generic doubly fed induction machines, distributed throughout the state, replacing 3,300 MW of conventional plant SWREC AZ 09 -- 23
Marcy 345kV Bus Voltage (pu) With Wind u u u u Impact of Wind Generation on System Dynamic Performance Fault at Marcy 345 kv bus Severe contingency for overall system stability Simulation assumes vectorcontrolled wind turbines Wind generation improves postfault response of interconnected power grid Without Wind Total East Interface Flow (MW) Without Wind With Wind SWREC source:ge/nyserda AZ 09 -- 24
Turbine Technology Advances Reduce Impacts SWREC AZ 09 -- 25
Wind Plant Control Capability Source: GE Energy SWREC AZ 09 -- 26
What s Different about Planning Transmission for Wind Energy Planning for energy resource different from planning for capacity resource look at 8760 hours Expect shift in time of peak period of transmission loading to shoulder periods - 60% of peak load New contingencies likely around times of minimum load and minimum conventional generation Perform reliability transmission design-nerc criteria Perform wind integration study to determine A/S requirements (ramping, operating reserves, extreme events) LOLE and ELCC calculations are likely to modify the reserve margins SWREC AZ 09 -- 27
System Planning and Operation Recommendations Perform detailed wind integration studies Deploy more flexible generation and load technologies Improve wind plant output forecasting tools Improve grid codes and wind plant models Aggregate wind plant output over large regions Improve balancing area consolidation and ACE sharing Ongoing forums to share operating experience SWREC AZ 09 -- 28
Transmission Planning Recommendations Develop adequate transmission capacity Comprehensive regional planning processes Federal leadership in developing transmission in support of national energy policy Reassessment of transmission financing approach Customers in remote regions can t afford it load pays in the end More certainty of transmission cost recovery More robust and flexible smart grid SWREC AZ 09 -- 29
Market Operation and Transmission Policy Recommendations Develop well-functioning real-time, hour-ahead and day-ahead energy and price responsive load markets and expand access to those markets Adopt market rules and tariff provisions that are more appropriate to weather-driven resources Make better use of physically (in contrast with contractually) available transmission capacity Eliminate pancaked rates SWREC AZ 09 -- 30
and the conclusion is There are no fundamental technical barriers to the integration of 20% wind energy into the electrical system, but There needs to be a continuing evolution of transmission planning and system operation policy and market development for this to be achieved. SWREC AZ 09 -- 31
For More Information Visit www.uwig.org Email info@uwig.org Phone Charlie Smith 703-860-5160 Sandy Smith 865-218-4600 Mail Utility Wind Integration Group PO Box 2787 Reston, VA 20195 USA SWREC AZ 09 -- 32