CSP real-time dispatch model hikes revenues by over 15%
New software combining dispatch algorithms with probabilistic forecasting could reduce CSP bid prices by a quarter and spur fresh U.S. CSP deployment, Mike Wagner, Principal Investigator at the National Renewable Energy Laboratory (NREL), told New Energy Update.
On October 1, NREL launched a new three-year study into optimized CSP dispatch software that allows real-time operations of CSP plants. Partners in the project include U.S. developers SolarReserve, BrightSource and research institutions Northwestern University and the Colorado School of Mines.
One of seven new NREL CSP projects funded by the U.S. Department of Energy’s (DOE) Solar Energy Technologies Office (SETO), the research partners will develop an automated decision-making tool to better align CSP supply with power demand.
The researchers predict the optimized dispatch software could help reduce CSP bid prices by some 25% and remove the need for fossil-fuel power capacity during high demand periods.
The dispatch software will add value through "increased revenue, reduced downtime and better generation and reliability," Wagner told New Energy Update.
In the U.S. Southwest, the optimization software could allow operators to produce full capacity during the top 40 to 50 high-demand hours of the year and over 97.5% capacity during the top 1,000 hours, Wagner said.
“This gives CSP tremendous grid value and enables new renewable capacity without requiring additional low-capacity-factor non-renewable backup capacity,” he said.
U.S. researchers are also developing high-temperature CSP designs to lower storage costs. Together, these projects will help drive the U.S. towards DOE’s 2030 cost targets of $50/MWh for baseload CSP plants and $100/MWh for peaker CSP plants, Mark Mehos, NREL CSP Program manager, said in a statement.
"SETO wants to stimulate significant, market-driven deployment of CSP with thermal energy storage in the United States," he said.
US large-scale CSP plants
Source: CSP Today Global Tracker
The new optimized dispatch software will advance simulations from current hourly methods to real-time operations.
The research will build upon non-linear solvers developed by the Argonne National Laboratory, probabilistic forecasting models by Northwestern University and dispatch optimization algorithms built by the Colorado School of Mines.
The model factors in forecasted solar resource and electricity prices in the coming hours and days, allowing operators to leverage thermal storage energy (TES) to exploit favorable prices or operating conditions.
Compared with un-optimized "block dispatch" methods, real-time optimized dispatch can increase revenue from day-ahead electricity sales "by 15% or more, depending on the market,” Wagner told New Energy Update.
The optimization model includes uncertainties from external sources to increase accuracy.
Historical data on areas such as cloud cover and temperature are used to build forecasts and uncertainty boundaries.
This helps to minimize "high cost" events such as turbine trips due to low storage levels, missed production revenues during high value periods and dumped power output, Wagner said.
The uncertainty modelling has a major impact, increasing realized revenue by around 5%, he said.
CSP developers predict falling costs and growing wind and solar penetration could support a new wave of CSP development in U.S. and Europe.
The real-time optimized dispatch system will be most beneficial to CSP plants in markets with significant variations in electricity prices between daytime and night time periods, Wagner said.
“The California ISO market is a good example, because PV production has caused significantly reduced daytime rates,” he said.
California installed solar capacity was 22.8 GW at the end of June, of which around 10 GW was utility-scale capacity.
California hourly system energy prices in 2017
(Click image to enlarge)
Source: California ISO
SolarReserve is developing the 2 GW Sandstone CSP plant with 10 hours storage in Nevada, which will dispatch power to California during high demand periods in the afternoon and evening.
SolarReserve aims to achieve a price of "well" below $90/MWh for the project, Kevin Smith, CEO of SolarReserve, told New Energy Update in September.
If optimized dispatch software can reduce this price by 25%, that would put it at below $67.5/MWh. In comparison, Lazard consultancy estimates the generation cost of U.S. combined cycle gas plants at $41-$74/MWh and gas peaking plants at $152-$206/MWh, according to its latest figures.
PV plus battery costs are also falling and these projects will become highly competitive for evening peak demand in the coming years, analysts from Bloomberg New Energy Finance (BNEF) told New Energy Update last month.
Falling costs will push the price of PV with battery projects in the U.S. Southwest below gas plants within the next few years, without any state subsidies, Tara Narayanan, solar analyst at BNEF, said.
Optimized dispatch will allow operators to maximize component performance while reducing labor costs and downtimes.
By considering the trade-offs between component performance and market factors, the software reduces the number of turbine starts which can accelerate damage and lead to additional maintenance costs, Wagner said.
A recent study by NREL and SolarReserve showed optimized immediate synchronization between production and dispatch can reduce annual turbine starts by 85%.
NREL has found that fewer turbine starts can increase annual revenues by around 8.5%, Wagner said.
Based on a plant with net revenue of $70 million per year operating over 25 years, the increase in revenues from fewer turbine starts could be around $150 million, he said.
Optimized dispatch also allows better planning of maintenance staff resources, aids cleaning strategies and data can be used to improve project development, Wagner said.
“We are still working on quantifying these impacts,” he said.
NREL has also adapted the optimization models to incorporate hybrid CSP-PV systems.
Interest in hybrid CSP-PV is growing as operators seek lower cost energy for longer dispatch periods.
Combined CSP-PV plants can provide 24-hour baseload power at a lower cost per MWh than stand-alone CSP or separate CSP and PV plants, market experts told New Energy Update in August. Hybrid project savings include lower capex per MWh from smaller solar field requirements and logistics and labor efficiencies in construction and operations phases.
Early tests show similar benefits from optimized dispatch for hybrid plants, Wagner said.
Preliminary modelling shows a "significant reduction in overall cost of energy for hybrid CSP-PV plants without sacrificing reliability,” he said.
By Kerry Chamberlain