ORNL mapping tool identifies suitable locations for new nuclear plants
Constructing a layer cake with 22 tiers would be a baking experiment of epic proportions—an architectural balancing act worthy of a pastry professional. Although Olufemi Omitaomu would not call himself a chef, he and his ORNL colleagues are concocting a multilayered cake of their own, except instead of being on a cake stand, it's on a computer screen.
OR-SAGE maps reflect reactor siting needs, such as level ground. This map shows areas with excessive slope in orange. The system also helps researchers determine if siting issues can be addressed, for example, by leveling a hill to reduce the slope of the site.
Instead of flour and sugar, Omitaomu deals with data—and lots of it. Twenty-two datasets known as layers make up an ORNL tool called OR-SAGE, short for Oak Ridge Siting Analysis for power Generation Expansion. The mapping tool, developed from a wide array of geographic information systems (GIS) data sources, is designed to support siting evaluations for power plants of all shapes and sizes.
The development of the OR-SAGE tool, which is being used to support a project funded by the nonprofit Electric Power Research Institute, was prompted by a 2008 ORNL internal study that examined the key issues associated with the country's future energy needs. A principal finding was that 300 gigawatts of new nuclear electric-generating capacity would be needed by 2050. What was not clear, however, was where those power plants could be located.
"We don't know if we have enough land area to support that analysis," says Omitaomu, a research scientist in ORNL's Computational Sciences and Engineering Division. "That's not a single power plant. It's hundreds."
Identifying candidate sites
An EPRI siting guide, based on guidelines from the Nuclear Regulatory Commission, formed the foundation for the OR-SAGE tool. The written criteria, representing factors such as population density, seismic activity and proximity to cooling water sources, were matched with appropriate datasets to provide values for each criterion. The ORNL team assembled the datasets from diverse sources, including the U.S. Geological Survey, the U.S. Census Bureau, the Department of Transportation, the Federal Aviation Administration and the Federal Emergency Management Agency.
One source unique to ORNL was LandScan, a population distribution model. LandScan models population distribution at resolutions down to the level of a city block. Unlike census data, which only provides a broad sense of where people are located within a large area, "LandScan is able to tell you exactly where those people are," says Omitaomu.
Finding the appropriate dataset for each written criterion was not a straightforward process. For example, as a rule of thumb, nuclear power plants need to be placed within 20 miles of a large body of water to meet cooling water needs. This required the ORNL team to limit acceptable sites to a 20-mile zone around water sources where plants could be reasonably sited. Locating a plant outside the zone might be prohibitively expensive in terms of water transportation costs.
After tweaking the GIS datasets to reflect the realistic siting needs, ORNL researchers divided up the continental United States into millions of cells, each measuring 100 meters by 100 meters— about 2.5 acres. Each dataset was then computed for every cell to see if the cell met the criteria. "For each cell, I want know how many people are in that cell. I want to know the slope of that cell; I want to know if it is protected land, if there is a stream in that cell, and so on," Omitaomu says.
The high-resolution nature of the 100-meter by 100-meter cell is a critical component of the analysis. If the cell size is too large, then the GIS tool might exclude large swathes of land, missing smaller viable areas within the cell. A cell can be excluded for siting eligibility for a number of reasons: if its slope is too steep, if too many people live within its boundaries, or if it happens to be part of a national forest, for instance. Omitaomu describes the process as looking for holes in the dataset: "You want to see which cell has a hole throughout all the layers. Then you pull all that together to get a base map of those areas that pass all those criteria."
Even if a given cell does not pass all the criteria, the OR-SAGE method can identify the factors behind its failure. This information can help researchers determine whether the reason for its exclusion can be addressed by, for example, leveling a hill to alleviate steep slope issues. Other factors, such as a high risk of seismic activity, may make the area completely unsuitable for siting a power plant.
In addition to identifying siting eligibility for individual cells, the OR-SAGE tool can be used to specify contiguous areas that are large enough to host a power plant. Traditionally, a large nuclear plant has required about 500 acres of land; however, recent small modular reactor designs require only about 50 acres.
Missing the forest for the trees
In the past, utilities looking for a suitable site for a nuclear power plants may have limited their options because they lacked the broad vision and detailed knowledge that OR-SAGE can provide. "Industry has never had a tool that can give them a view of the entire national landscape," Omitaomu says.
Typically, power plant locations are chosen from a pool of predetermined sites, in large part because scouting out new locations with traditional surveying methods can be time-consuming and costly. The OR-SAGE tool, on the other hand, can be run quickly and easily from a personal computer.
Omitaomu emphasizes that the OR-SAGE method is not intended to replace on-the-ground field studies and data collection—a necessary part of nuclear power plant licensing. "This is a screening tool; it does not tell you, go put something here," Omitaomu says. What OR-SAGE can do, however, is broaden the horizons of groups looking to site power plants.
Even if a company or a utility has preselected a handful of candidate sites, the ORNL tool can save them time and money by narrowing down the options. "If you have several areas in mind," Omitaomu explains, "instead of sending people to these areas and doing field analyses, this tool can easily tell you that, perhaps, two out of the five areas are not suitable. So instead of wasting money to do detailed analyses on all five, you can focus on the remaining three."
The icing on the cake
The short answer to the team's initial question—does the U.S. have enough land to accommodate a large increase in the number of nuclear power plants—is a solid yes. Taking into account the need for contiguous land areas, the ORNL team in one baseline scenario assessed that 13 percent of land in the continental United States may be suitable for siting a large nuclear reactor, while 24 percent may be appropriate for hosting a small modular reactor.
EPRI is using results from the OR-SAGE project as input for economic analyses to explore options for deploying various types of electrical generation plants. As ORNL researchers crafted the modeling system, EPRI realized the value of OR-SAGE for evaluating sites beyond those for nuclear plants. "We started with nuclear, and once they saw it was a good tool, they wanted us to extend it to renewables, like solar, advanced coal and compressed air energy storage," Omitaomu says.
The national scope of the OR-SAGE tool, combined with its ability to analyze areas almost as small as a football field, means that it can be used to help make siting decisions on many different scales. By producing maps that show where different types of power plants can realistically be placed, the GIS model can also help policymakers who want to develop energy profiles or portfolios for a state, a region or the nation as a whole.—Morgan McCorkle