580 likes | 592 Views
This study explores the costs and operating impacts of wind, PV, and concentrating solar power (CSP) on the grid, focusing on the variability and uncertainty of these renewable energy sources. It includes scenarios for different wind and solar penetrations and provides insights into the challenges of modeling such a large footprint. The study also analyzes load, wind, and net load data, as well as the validation results of wind and solar generation.
E N D
Western Wind and Solar Integration Study Update Michael Milligan, NREL (speaking for a large team at GE, 3TIER and NREL) Westconnect, Apr 21, 2009
Overview • Goal • To understand the costs and operating impacts due to the variability and uncertainty of wind, PV and concentrating solar power (CSP) on the grid • Not the cost of wind or solar generation • Scope of study • Operations, not transmission study • Study year – 2017 to line up with WECC studies • Simulate load and climate of 2004, 2005, 2006 forecast to 2017
Scenario Overview • Baseline – no new renewables • In-Area – each transmission area meets its target from sources within that area • 30% wind, 5% solar in footprint (20% wind, 3% solar in the rest of WECC) • 20% wind, 3% solar (10% wind, 1% solar rest of WECC) • 10% wind, 1% solar (10% wind, 1% solar rest of WECC) • Mega Project – least cost of delivered energy • Local Priority – similar to Mega Projects but with small bonus for in-area sites • Plus other scenarios yet to be determined (high solar, high capacity value, high geographic diversity) Solar is 70% CSP and 30% PV. CSP has 6 hours of thermal storage. Penetrations are by energy
Modeling such a large footprint is challenging • Many terabytes of data • “Seams” issues occurred in initial runs • Large wind ramps that were not real • Validated against meteorological towers • Corrections improved the realism of the wind speed databetter estimates of wind power generation
Wind Selected for 30% In-Area Scenario Red dots = Pre-selectedBlue dots = New sites 30% Wind In Footprint20% Wind Out of Footprint 2,340 MW 8.4 TWh 7,050 MW 17.3 TWh Legend: Wind MW Annual Energy 5,640 MW 18.5 TWh 900 MW 2.6 TWh 11,220 MW 29.9 TWh Total Installed MW: 29,940 MW (998 sites) Total Wind Capital Cost: $59.9 B 2,790 MW 9.4 TWh
In-Area Scenario - Solar 5% Solar In Footprint3% Solar Out of Footprint 0 MW 300 MW 0.4 TWh Legend: CSPw/S MW PV MW SolarAnnual Energy 600 MW 500 MW 2.8 TWh 700 MW 700 MW 3.3 TWh 200 MW 100 MW 0.5 TWh Total In-footprint Installed MW: 2900 CSP w/S 2900 PV Total In-footprint Energy Solar: 15 TWh Total Solar Capital Cost: $23 B 1000 MW 1000 MW 5.3 TWh 400 MW 300 MW 1.6TWh
30% In-Area Scenario Energy Summary In Footprint Out of Footprint 30% Wind, 5% Solar In Footprint20% Wind, 3% Solar Out of Footprint
30% In-Area Scenario Power Summary In Footprint Out of Footprint Penetration = Wind Plant MW Rating Load MW
Load, Wind, and Net Load • Data for load, wind, and solar are time-synchronized • For each MW of wind or solar that is generated, 1 MW of load need not be supplied by conventional generation • Net load: Consumer load less • Wind generation • Solar generation • Net load is what the system must operate to
Study Area Total, Load Wind, Solar and Net Load for Oct 8th 2006(30% In-Area Scenario) With new data With old data
Texas wind data validation 536 MW in 5 wind plants
Another site wind validation Mesoscale wind data capacity factor is off by 2-9%
Assessment of validation results • NWP data has good overall match to expectation • …but does not fully capture wind behavior and is not adequate for bankers or project assessment • State of the art needs improving • …but probably does a good job representing overall variability of wind over this large study footprint
Solar PV data – hourly and 10 min Satellite cloud cover model from R. Perez at SUNY/Albany produced 10 km hourly solar radiation data. R. George at NREL generated 10 minute data using measured PV output data.
New Mexico (2006) Benefits of aggregated wind: Actual wind output vs. One-Hour delta as a Percentage of Installed Wind Capacity (30% Scenario) Study area aggregation tends to mitigate relative impact of large ramps
Distribution of Extreme Hourly Wind Deltas 2004 – 2006(30% In Area)
Study Area Monthly Energy from Wind and Solar for 2004 – 2006 (30% In Area Scenario)
Study Area Average Daily Energy from Wind and Solar for 04 - 06 (30% In Area Scenario) 2004 2006 2005 Hour of Day Hour of Day Study Area Average Energy by hour, from Wind and Solar 2004 2006 2005 Hour of Day Hour of Day Study Area Average Percent Energy by hour, from Wind and Solar
State Monthly Energy from Wind and Solar for 2004 – 2006 (30% In Area Scenario)
Net Load Duration, Wind and Solar Penetration Over Year 2006 113% Instantaneous penetration, based on load MW Min load 22200 MW Below existing min load ~57% of year for 30% scenario 24 hours with over 100% penetration 11.7% No PV production for half the year 7.3% CSP output for 60% of the year
Study Area Net Load Duration for 2004, 2005, 2006 (30% In Area Scenario)
Study Area Load and Wind Average Daily Profiles By Seasonal Month for 2006 (30% In Area Scenario) Apr Jan Jul Oct
Study Area CSPws and PV Average Daily Profiles By Seasonal Month for 2006 (30% In Area Scenario) Jan Apr Jul Oct
Study Area Total Load, Wind and Solar Variation Over Month of April(30% in Area Scenario) Minimum net load: –2887 MW Instantaneous penetration: 112% Substantial increase in net load variability driven largely by wind variation
WECC Total Load, Wind and Solar Variation Over Month of April(30% in Area Scenario) Minimum net load 20,525 MW Instantaneous penetration: 73% Impact of wind on WECC net load less dramatic over a larger control area
Variability Analysis - Deltas • Statistics used to characterize variability: • Delta (∆) – The difference between successive data points in a series, or period-to-period ramp rate. • Positive delta is a rise or up-ramp • Negative delta is a drop or down-ramp • Mean () – The average of the deltas (typically zero within a diurnal cycle) • Sigma (σ) – The standard deviation of the deltas; measures spread about the mean For a normal distribution of deltas, σ is related to the percentage of deltas within a certain distance of the mean
Wind Deltas vs Load Deltas by season for 2004-2006 (30% in Area Scenario) Increased L-W down-ramps (186, 7528) Load and wind deltas offset Q2 29 hours (342 GWh) where wind pushes L-W delta beyond largest load delta -7000 MW -6000 MW (-4125, 2950) -5000 MW Q1 (2985, -4372) -5000 MW -6000 MW -7000 MW Q4 Q3 Increased L-W up-ramps Load and Wind deltas offset (1199, -5926)
Wind Deltas vs Load Deltas by season for 2004-2006 (30% in Area Scenario) Load and wind deltas offset Increased L-W down-ramps Increased L-W up-ramps Load and wind deltas offset
Study Area Average Daily Profile of Deltas Over Year 2006(30% In Area Scenario)(Avg. +/- sigma, Minimum, Maximum) 7874 MW Total Load and Net Load(MW) Load and Net Load Delta (MW) -6851 MW Hour of Day
Study Area Total Load, Wind and Solar Variation for Selected Days(30% In Area Scenario) L-W-S Maximum Negative Delta Day - Mar 26, 2006 L-W-S Maximum Positive Delta Day - Jan 1 2006 Minimum Net Load Day - Apr 15 2006 Solar Maximum Negative Delta Day - Jan 20 2006
Study Assumptions • 2017 Fuel Prices: • Coal ~ $2.00/MBtu • Natural Gas ~$8.00/MBtu • Carbon Tax : $30/ton • Energy Velocity Database • ~24 GW capacity added 2009-2017 timeframe to maintain reserve margins (~11GW not in plans) • NERC ES&D Peak Load Projections • Economically Rational, WECC-wide Commitment and Dispatch
Forecast Error • Forecast error varied annually and regionally • Reduced forecast 10% in study footprint and 20%outside • “R” scenarios • Also considered perfect forecasts • “P” scenarios
Combined cycle units most displaced Pumped storage has moderate increase
Weekly Operational Analysis • Examined hourly operation for two specific weeks in mid-April and mid-July • Results show hourly variation in generation by type as renewable penetration increases