TODAY’S STUDY: When There Is A LOT Of New Energy
Impacts of High Variable Renewable Energy Futures on Wholesale Electricity Prices, and on Electric-Sector Decision Making
Joachim Seel, Andrew Mills, Ryan Wiser, et al, May 2018 (Lawrence Berkeley National Laboratory)
Executive Summary
Increasing penetrations of variable renewable energy (VRE) can affect wholesale electricity price patterns and make them meaningfully different from past, traditional price patterns. Many long-lasting decisions for supply- and demand-side electricity infrastructure and programs are based on historical observations or assume a business-as-usual future with low shares of VRE. Our motivating question is whether certain electric-sector decisions that are made based on assumptions reflecting low VRE levels will still achieve their intended objective in a high VRE future. We qualitatively describe how various decisions may change with higher shares of VRE and outline an analytical framework for quantitatively evaluating the impacts of VRE on long-lasting decisions.
We then present results from detailed electricity market simulations with capacity expansion and unit commitment models for multiple regions of the U.S. for low and high VRE futures. We find a general decrease in average annual hourly wholesale energy prices with more VRE penetration, increased price volatility and frequency of very low-priced hours, and changing diurnal price patterns. Ancillary service prices rise substantially and peak net-load hours with high capacity value are shifted increasingly into the evening, particularly for high solar futures…
Key Findings
1-VRE Expansion Leads to Modest Retirement of Firm Capacity of 4-16%, Especially Coal, Oil and Steam Turbines
Total installed capacity increases with VRE growth as average capacity credit is 10-24% for new wind and 8-63% for new solar
SPP: firm capacity reduction by 9-12%
Retirement of Coal (4-8GW) and Other Gas (7GW, e.g. steam turbines)
Partially offset by Gas CT growth (4-7GW)
NYISO: firm capacity reduction by 13-16%
Dual Fuel (Oil) retirement (5+ GW)
Partially offset by Gas CT growth (1-2GW)
CAISO: firm capacity growth by 2-4%
Little overall changes in capacity
Minor growth in Gas CC (0.4-0.8GW) and Gas CT (0.4GW)
ERCOT: firm capacity reduction by 4-14%
Coal retirement largest in wind scenario (7GW) - none in solar
Largest Gas CT retirement in balanced (4GW vs. 1GW in solar)
Gas CC largely stable, growth by 1GW in wind scenario
2-Energy from VRE Primarily Displaces Coal and Natural Gas Generation
VRE generation offsets conventional generation 1-1, except when curtailed (in solar scenarios average VRE curtailment is 3-8% of all VRE generation)
SPP: fossil generation reduction by 27-32%
Reduction in Coal and Gas CC generation (30-35TWh each)
Minimal changes in Gas CT
11TWh of VRE curtailment, 14TWh of export in solar scenario
NYISO: fossil generation reduction by 44-50% Reduction in Gas CC (32-35TWh) and imports (17TWh)
Minimal drop in Gas CT CAISO: fossil generation reduction by 25-33%
Reduction in Gas CC (esp. in wind scenario: 17-28 TWh), imports (22- 26 TWh) and Gas CT (4-6 TWh)
Difficult to assess composition of imports as we lack fuel information
ERCOT: fossil generation reduction by 30-34%
Reduction in Coal (35-46TWh) and Gas CC (50-55TWh), esp. in solar, 60-80% Gas CT reduction (more in wind/balanced)
Up to 13TWh of solar curtailment, 5TWh of wind curtailment
3-VRE Changes the Marginal Carbon Emissions Rate
Total carbon emissions decrease with high VRE buildout by 21-47%
Marginal carbon emission rates decrease by 6-21% (ERCOT) to 28-38% (SPP)
VRE shifts timing of high marginal emissions, decreases by 750-1750lbs/MWh over the middle of the day in solar scenario
VRE leads to an increase in frequency of hours with very low marginal emission rates ranging from 5% of all hours in CAISO (wind scenario) to 31% in SPP (solar scenario)
4-Annual Average Energy Prices Decline with Increasing VRE Penetration
Load-weighted average electricity prices decrease with higher VRE penetration by $5 to $16 relative to low VRE baseline, depending on scenario and region
5-Average Energy Price Reduction From VRE Falls Within Range of Previous Studies
A common metric for comparisons across studies is the change in price ($/MWh) per % increase in VRE penetration
Accounting for the different starting levels of VRE penetration, the average reduction in electricity is $0.21-$0.87/MWh for each additional % of VRE penetration ($0.19-$.81/MWh for pre-curtailment VRE)
CAISO has greatest reduction due to carbon costs and relatively small incremental VRE generation growth
Decrease in average prices will reduce profitability of inflexible generators that are fully exposed to those prices (nuclear, solar, wind, to some extent coal and gas steam)
Our observation falls roughly in the range of established literature
6-Low Energy Prices Become More Frequent Under High VRE Scenarios
In some regions, the shape of the price distribution curve does not change dramatically but is merely shifted downwards (e.g. NYISO)
Other regions feature a more pronounced ‘cliff’, featuring a dramatic increase in hours with very low prices (e.g. ERCOT) Low prices driven by solar more than wind
7-High VRE Significantly Alters Diurnal Price Profiles, Particularly With High Solar
Substantial decrease in prices over the middle of the day in solar scenarios across all regions
Diurnal profiles vary by season
Morning: wind vs low VRE scenario in CAISO:
• -$25/MWh in Spring, but only -$10/MWh in Fall and Winter
Afternoon: solar vs low VRE scenario in NYISO:
• -$30/MWh in Spring and Summer, but only -$15/MWh in Winter
Evening: balanced / solar vs low VRE scenario in ERCOT:
• +$180/MWh in Summer (driven by few high-priced hours), but only +$5/MWh in Winter
Price peaks remain across most seasons in the early evening at levels similar to low VRE scenario
8-High VRE Increases Price Volatility; Prices Are Most Irregular with High Wind
Wider range in wind scenario during early morning hours
Change in average diurnal profile in balanced scenario & 5 th -95th range increases during the middle of the day
Coefficient of Variation is standard deviation of prices normalized by mean energy price to facilitate cross-regional comparison
High volatility in ERCOT in part due to few high priced hours ($1000- $9000/MWh) due to Operating Reserve Demand Curve
Total price volatility increases with VRE penetration, largest with solar Irregularity of prices (variability not captured by diurnal profiles, seasonal shifts and weekdays/weekends) is highest in wind scenarios
9-High VRE Leads to an Increase in Ancillary Service Prices
Increases for regulation reserve requirements with VRE are consistent with previous region-specific studies (an increase in the range of 1-1.5% of hourly VRE generation) VRE was not allowed to provide AS
Average prices for regulation (up and down) and spinning reserves increase by 2-8x across most regions in high VRE future to $15-$38/MWh due to high opportunity costs at low-net load levels
Non-spinning reserves tend to remain at lower prices
High solar penetrations often lead to the strongest increase, with peak prices above $190/MWh in CAISO across all AS-types
In SPP, downward regulation prices reach occasionally $200/MWh in all high VRE scenarios
Diurnal AS price profiles and their peaks can change significantly, as do price ranges
10-High VRE Has Modest Impacts on Capacity Prices; More Pronounced Shift In Timing of Peak Periods
Mixed trends in annual averages, solar often leads to higher prices Depending on region, top net-load hours are concentrated over fewer hours of the day and pushed later into the evening, especially in solar scenarios Top 100 net-load hours are spread however over more days (and months) in the high VRE scenarios in comparison to the low VRE scenario (from 22 to 45 days in ERCOT).
Conclusion and Discussion
VRE additions enable modest firm capacity and strong non-VRE generation reduction
Growth in VRE can decrease overall average wholesale market prices by $5-$16/MWh
Changing timing of cheap/expensive electricity and regularity/predictability of patterns:
Growth in frequency of very low priced periods (up to 20% of all hours in ERCOT)
Changing diurnal patterns especially with high solar
Increase in irregularity of wholesale prices especially with high wind
Lower average energy prices will increase relative importance of rising capacity and ancillary service prices Magnitude and importance of these shifts depends on response of other market participants (changing aggregate load shapes, DR participation, storage)
Results sensitive to our assumptions:
Not modeling intra-regional congestion, limited VRE leakage to neighboring regions
Fuel price and emission cost deviations impact optimal generator portfolio and marginal prices Focus on single exemplary year 2030 that doesn’t capture inter-annual variation or longer-term evolution of electric system
0 Comments:
Post a Comment
<< Home