So you want to build a solar or wind farm? Here’s how to decide where

The findings show the benefits of coordinating the siting of solar farms, wind farms and storage systems, taking into account local and temporal variations in wind, sunlight and energy demand, to maximize the use of renewable resources. The researchers found that this approach could reduce the overall cost of the system by eliminating significant investments in storage, while maximizing the availability of clean energy when needed.

The research will be published in the journal Cell Sustainability Reportco-authored by Liying Qiu and Rahman Khorramfar, postdocs in the MIT Department of Civil and Environmental Engineering, and professors Saurabh Amin and Michael Howland.

Lead author Qiu said that with the team’s new approach, “we can take advantage of resource complementarity, which means that different types of renewable resources, such as wind and solar, or renewable resources in different locations can compensate for each other in time and space .

Such complementarities will become increasingly important as variable renewables make up a larger share of electricity entering the grid, she said. By more smoothly coordinating the peaks and troughs of production and demand, she said, “we’re actually trying to use natural variability itself to solve the variability problem.”

Typically, when planning large-scale renewable energy installations, “some work is done at the national level, for example, 30% of the energy should be wind and 20% solar. That’s very general,” Qiu said. In this case In the study, the team looked at weather data and energy system planning models with a resolution of less than 10 kilometers (about 6 miles). “It’s a way of identifying where we should build every renewable energy plant, rather than just saying the city should have this many wind or solar farms,” ​​she explains.

To compile data and enable high-resolution planning, researchers rely on a variety of sources that have not been previously integrated. They used high-resolution meteorological data from the National Renewable Energy Laboratory, which is publicly available at 2-km resolution but is rarely used in planning models at such fine scales. This data is combined with an energy system model they developed to optimize site selection at a resolution of less than 10 kilometers. To understand how granular data and models make a difference in different regions, they focused on three regions of the United States—New England, Texas, and California—while analyzing up to 138,271 possible site locations in a single region. .

By comparing site selection results based on typical and high-resolution methods, the team showed that “resource complementarity really helps us reduce system costs by aligning renewable generation with demand,” which should translate directly to The real world, Qiu said, is decision-making. “If an individual developer wants to build a wind or solar farm and just goes where the average wind or solar resource is, it may not necessarily guarantee the best fit for a decarbonized energy system.”

This is because of the complex interplay between electricity production and demand, both of which change hourly and month-to-month with the seasons. “What we are trying to do is minimize the difference between energy supply and demand, rather than simply provide as much renewable energy as possible,” Qiu said. “Sometimes you have a generation that can’t be utilized by the system, and sometimes you don’t have enough generation to meet the demand.”

In New England, for example, new analysis suggests that more wind farms should be built in areas where winds are stronger at night and solar power cannot be harnessed. Some places are windier at night and others are windier during the day.

The insights were revealed through the integration of high-resolution weather data and energy system optimization used by the researchers. Complementarity between renewable energy power plants is much smaller when planning using lower resolution weather data (generated at global resolution of 30 km and more commonly used for energy system planning). Therefore, the total system cost is much higher. The high-resolution model enhances the complementarity between wind and solar farms due to improvements in renewable resource variability.

The researchers say their framework is flexible and can be easily adapted to any region to take into account local geophysical and other conditions. For example, in Texas, peak winds in the west occur in the morning, while peak winds along the south coast occur in the afternoon, so the two naturally complement each other.

Khorramfar said the work “highlights the importance of data-driven decision-making in energy planning.” The work shows that using such high-resolution data coupled with carefully crafted energy planning models “can reduce system costs and ultimately provide a more cost-effective path to the energy transition.”

Amin, a principal researcher at the Information and Data Systems Laboratory, said one of the surprising things about the findings was how significant the gains were from analyzing relatively short-term changes in inputs and outputs that occurred over a 24-hour period. “Prior to this study, people would not have thought of the potential for cost savings by leveraging complementarities within a single day,” he said.

Furthermore, Amin said, surprisingly, such modeling could reduce the need for storage as part of these energy systems to a significant extent. “This study shows that there is actually hidden cost-saving potential in taking advantage of local weather patterns, which could reduce storage costs.”

Howland said the system-level analysis and planning presented by this study “changes the way we think about where renewable power plants are sited and how to design them so that they best serve the energy grid. It has to go beyond just We continue to collaborate across traditional research areas, integrating expertise in fluid dynamics, atmospheric science and energy engineering to lower the cost of energy from a single wind or solar farm.

This research was supported by the MIT Alliance for Climate and Sustainability and the MIT Climate Grand Challenge.