THE POWER OF WIND AND SOLAR TOGETHER

How do Wind and Solar Power Affect Grid Operations: The Western Wind and Solar Integration Study
D. Lew, M. Milligan, G. Jordan, L. Freeman, N. Miller, K. Clark, and R. Piwko, September-October 2009 (National Renewable Energy Laboratory)

SUMMARY
The dream: Powering the nation on New Energy.

The very real parts of the dream: Wind off the Atlantic coast, on the Great Lakes and the Midwestern plains and in the Pacific Northwest; Sun on the Southwestern deserts; Ocean energies in the Florida Gulf Stream, along the Mississipi and other great rivers and off the Pacific coast; geothermal resources in the Mountain and Far West; and biomass plants in woody and agricultural areas everywhere.

The challenge: Getting the electricity from where it is generated to where it is needed.

How Do Wind and Solar Power Affect Grid Operations: The Western Wind and Solar Integration Study (WWSIS) is the most authoritative study yet on how to meet the challenge with real resources and realize the dream.

Also one of the biggest studies so far done on how to integrate New Energy resources into a steady power supply, The Western Wind and Solar Integration Study (WWSIS) began in 2007 and has been looking at how it would work if as much as 35% of the grid’s power was coming from wind, photovoltaics (PV), and concentrating solar power (CSP. The study has focused on the WestConnect grid in Arizona, Colorado, Nevada, New Mexico, and Wyoming, using 3 scenarios.

The conclusion: There are very real and achievable ways to make the dream come true.

click to enlarge

COMMENTARY
The study, sponsored by the U.S. Department of Energy (DOE) and performed by the National Renewable Energy Laboratory (NREL) in partnership with WestConnect, was designed to answer questions crucial to utilities, Public Utility Commissions (PUCs), developers, and regional planning organizations:
(1) Does diversity of supply across regions ease the challenge of the New Energies’ variability ?
(2) How does using local resources compare to using out-of-state resources?
(3) Does cooperation between transmission systems to balance supply ease the variability question?
(4) How does New Energy storage fit in?
(5) Does adding New Energy mean grid operators have to rethink reserve requirements?
(6) How important is forecasting sun and wind?
(7) Can the steadiness of hydroelectric power ease the integration of the New Energies into the grid supply?

click to enlarge

The study modeled the year 2017, using historical load and weather patterns from 2004, 2005, and 2006. It found significant variations in the 3 years' sun and wind patterns.

It applied a 35% New Energy load in the studied area and as well as up to a 23% New Energy load in the surrounding regions. This tested both the grid in the studied area and the ability of that grid to work with significant New Energy supplies around it.

click to enlarge

Over 75 gigawatts of wind power sites needed to be modeled. The award-winning 3TIER Group developed the wind dataset using rigorously detailed Weather Research and Forecasting (WRF) mesoscale Numerical Weather Prediction Model (NWP) information. 3TIER Group also addressed forecasting concerns with day-ahead wind forecasts for each hour across 960 gigawatts of wind sites (32,000 sites of 30 megawatts each). Even array and electrical losses were considered. The wind dataset is publicly available.

Modeling for the solar resource, done by the State University of New York (SUNY)/Albany, was based on a more limited dataset. Experience from the 4.6 megawatt Springerville Generating Station Solar System in Arizona indicated that central station photovoltaic (PV) supplies could have significant impacts on the grid but there was little corroborating data. SUNY/Albany used the data to build 100-megawatt blocks for assessment purposes and NREL broke those blocks into 10-minute periods. The hourly PV profiles are available on the web.

click to enlarge

Concentrating solar power (CSP) was modeled as 100-megawatt blocks of parabolic trough plant outputs and included a hypothetical 6 hours of storage. 200+ gigawatts of concentrating solar power plants were modeled and the models are available on the web. The stored power construct was modeled so that electricity was dispatched in a typical utility load pattern (using Southern California Edison as the model). Minimal losses and a stable output were assumed.

NREL created 3 scenarios, 2 “bookend” scenarios and a “middle ground” scenario:
(1) An In-Area (IA) scenario used local resources. It included the best possible sites, a mix of supply and geography, and adequate local transmission.
(2) A Mega-Projects (MP) scenario used higher quality but more remote resources. It assumed $1600/MW-mile transmission capital cost; $2000/kW wind capital cost; $4000/kW solar capital cost; and 1% losses per 100 miles cost penalty. It then assumed high capacity and transmission development.
(3) The Local Priority (LP) scenario looked at a more “realistic” development of projects and new transmission. Its assumptions were similar to MP but with a 10% cost-benefit for local wind and solar projects and the assumption of some, but not as much, new transmission.

click to enlarge

WWSIS is an ongoing study, expected to be completed in 2010, but much vital data has been developed, offering 6 major insights:
(1) Balancing area coordination is imperative to integrate 35% New Energy into the grid.
(2) Aggregating wind and solar outputs does mitigate variability and ramping supplies.
(3) Impacts do not significantly change whether local or remote sources comprise the aggregated supplies. Overall cost savings, displaced generation, spot prices, and other factors are very similar.
(4) Activity in the transmission system is very dependent on activity in the surrounding systems. Export of excess production decreased significantly as New Energy capacity of the surrounding systems went up from 11% to 23%.
(5) Pumped hydro storage is useful in such a New Energy system but it is not necessary.
(6) Absolutely perfect forecasts are only modestly better for the system than current state-of-the-art forecasts.

click to enlarge

A very detailed statistical analysis was done by NREL’s GE partners. Further insights emerged:
(1) There are significant year-to-year and month-to-month variations for both solar and wind energies but less scenario-to-scenario variation.
(2) Size matters. Some small areas have adequate wind or sun to meet 100% of local needs but over large regions, cooperation and integration are needed to put large percentages of New Energy on the grid.
(3) More geographic diversity mitigates the variability of solar and wind energies more effectively.
(4) The biggest ramping up of demand is in the late afternoons in the late fall and winter when both wind and solar tend to drop off at the same time and there is increasing demand for electricity in the chilly early darkness. The biggest ramping down of demand is in the evening in the late summer and early fall, when consumer demand falls off in the dimmed cooling sunlight of evening.

click to enlarge

An extremely interesting operational analysis was done of the 3 scenarios at 3 New Energy percentages by the GE researchers. It assumed a $2/MBTU coal price, a $9.5/MBTU natural gas price, and a $30/ton of carbon dioxide tax. It assumed a reasonable participation by grid operators throughout the region. It added 24 gigawatts of capacity to the existing system as a reserve margin. The avoided cost of fuel, emissions and efficiency losses from subnormal plant output were included in operating savings. ‘Wear and tear’ costs from harder cycling of plants could not be accounted for because they have not yet been definitively calculated.

Operational findings:
(1) There was no significant operational variation between the 3 scenarios using local and remote resources. The significant differences were dependent on how much New Energy was in the system.
(2) Forecasting is critical to operations.
(3) Operation was unaffected for New Energy of up to 23% of grid supply in the studied system and 11% outside it. At 35% in the studied system and 23% outside it, operations became more complicated.
(4) As New Energy grid supplies go from 23% to 35% in the system and 11% to 23% outside it, about 10% less New Energy gets exported to surrounding regions because there is less to export and less required.
(5) Given a perfect forecast, a 23% New Energy grid supply drives spot power prices lower but as New Energy grows toward 35% of grid supply, forecast errors also grow and spot prices follow.
(6) The more New Energy there is on the grid, the more combined cycle and gas turbine generated power is displaced. Coal begins to be displaced when New Energy gets to be 35% of grid supply.

click to enlarge

(7) In the local sources scenario, a 35% New Energy grid supply saves $20 billion ($82/MWh) and should bring a $2.20/MWh value to the western transmission system.
(8) Spot price revenues' value on the studied system in the local sources scenario is ~$60/MWh for solar (both CSP and PV) and ~$38/MWh for wind. For the entire western transmission system, spot price revenues' value is ~$50/MWh solar and ~$30/MWh for wind.
(9) The more New Energy there is in the system, the more energy goes unserved but discounting the forecast reduces unserved energy to almost nothing and adds little in spilled energy.
(10) Going from 23% to 35% New Energy requires the grid operator to interrupt the demand on the system in cases of sudden changes in supply. This does not have to add a significant cost burden.
(11) The major, statistical and operational findings did not change in the 3 projected years, despite variations in other conditions.

click to enlarge

Finally, a reliability analysis showed that capacity values of a combined wind, PV and CSP grid supply were also consistent across the scenarios and years modeled.

Future studies will look at more detail about the increase from 23% to 35% New Energy, at the role of storage and at the roles of plug-in vehicles and demand response in grid management.

What is already clear is that New Energy on the grid is a manageable concept and putting more on the grid is rapidly moving from dream to reality.

click to enlarge

QUOTES
- From the “Introduction” to the study: “…The [WWSIS] study was originally established to follow onto DOE’s 20% Wind Energy by 2030 report…which did not find any technical barriers to reaching 20% wind energy in the continental United States by 2030. This study and its partner study, the Eastern Wind Integration and Transmission Study, performed a more in-depth operating impact analysis to see if 20% wind energy was feasible from an operational level. In DOE/NREL’s analysis, the 20% wind energy target required 25% wind energy in the western interconnection; therefore, this study considered 20% and 30% wind energy to bracket the DOE analysis. Additionally, since solar is rapidly growing in the west, 5% solar was also considered in this study…The goal of the WWSIS is to understand the costs and operating impacts due to the variability and uncertainty of wind, PV, and CSP on the grid. This is mainly an operations, not transmission study, although different scenarios model different transmission build-outs to deliver power. The study does not focus on the cost of generating wind or solar power, but rather the operational costs and savings due to fuel and emissions…”

click to enlarge

- From the study: “[One graph] shows the operation of the system going from no new renewables to 11% to 23% to 35% renewables using the IA scenario during a week in April 2006. At 35% renewables, the combined cycle units are almost completely off, gas turbine output has increased, and the coal plants are cycling significantly. Even the nuclear units are trying to cycle some, which more likely would indicate a need to spill some of the wind generation. Total generation inside the study area drops because the rest of the WECC, with 23% renewables penetration, is now reducing their need for imports from the study area. This particular week in April was a severe wind week. Other months did not have as much impact. The 5th plot shows operation for a week in mid-July for the 35% penetration scenario, which did not seem to present any operational strain on the system…”

click to enlarge

- From the study’s concluding “Next Steps” section: “The study originally anticipated undertaking additional scenarios, but because the operational impacts of the three scenarios did not show great differences, the study will instead focus on other analysis before it is completed in early 2010. The LP scenario will be re-run with a 23% renewable energy penetration in both the study footprint and the rest of WECC so that the results from stepping up from the 23% renewables (11% in rest of WECC) to the 35% renewables (23% in rest of WECC) can be understood. A second analysis will examine the role of storage on different timescales and look at the value of storage for various penetration levels. Plug-in hybrid electric vehicles and demand response will be considered in this analysis”