Catalyzing Reactor Research
CASL is the Department of Energy's first energy innovation hub
ORNL's recently dedicated Consortium for Advanced Simulations of Light Water Reactors (CASL) is the Department of Energy's first energy innovation hub. The facility is designed to catalyze nuclear research and development with its unprecedented simulation and modeling capabilities.
ORNL's computing resources enable scientists to visualize reactors in unprecedented detail.
CASL is supported by a consortium of 10 core partners—three from industry, four DOE laboratories and three universities. In the early days of the partnership, CASL's industrial partners are the key players, setting the organization on a heading that will ensure developments can be sucessfully implemented in the nuclear industry. The partners, Westinghouse, the Tennessee Valley Authority and the Electric Power Research Institute, represent a cross section of the nuclear energy industry. Westinghouse designs and sells reactors and reactor fuel, TVA operates nuclear power plants, and EPRI acts as the research and development arm of the nuclear industry. These organizations provide CASL with a broad perspective on issues and requirements that affect the entire nuclear sector.
"Overall, nuclear energy performance goals drive our activities," says CASL Director Doug Kothe. The CASL research team focuses on improving the performance of pressurized water reactors, which account for 60 percent of the hundred or so reactors currently on line in the U.S. This includes devising safe, innovative ways to run these reactors more efficiently, driving operating costs down, and burning reactor fuel longer to minimize waste products.
To accomplish these goals, the CASL team is currently concentrating on developing a new suite of simulation tools called the Virtual Environment for Reactor Analysis. Researchers will use VERA to model what happens inside a working reactor vessel—specifically the kind of pressurized water reactors that TVA operates. (Three of the six reactors TVA currently operates are PWRs.) In the longer term, CASL will broaden VERA's scope to encompass modeling operational and safety scenarios for other types of reactors, including proposed next-generation designs.
Kothe notes that in the wake of the recent events at Japan's Fukushima nuclear site, a lot of people want to know if CASL's simulations might help scientists understand what went wrong with the Fukushima reactors and how to prevent similar problems in the future. He explains that while CASL's goal isn't to model severe accident scenarios, one aspect of CASL's scope of work is Fukushima related: understanding how reactor fuel behaves when coolant is lost or when it is suddenly exposed to steam or air. "What we learn will be relevant to a range of safety analyses," Kothe says, "including studies of the problems that occurred at the Fukushima site—but we're not modeling any sitespecific accident scenarios."
Virtual reactor
The first issue CASL will address is the efficiency of the current generation of nuclear reactors. This has been a concern of the nuclear industry for several years. In fact, utilities have put an additional 6 GW of power on the grid without building a new reactor through improvements in operational efficiency. "This means running the plants more efficiently and at higher power," Kothe says, "which usually involves using more fuel, fresher fuel and increasing the flow of coolant through the reactor core." One of the main challenges CASL researchers will face is understanding how fuels perform under these conditions and determining how to generate more power while producing less waste.
CASL will investigate fuel performance to help nuclear plants produce more power with less waste.
In the longer term, CASL will also investigate what Kothe calls "localized issues," those involving a specific fuel rod or fuel pellet in a scenario where the reactor has suffered a loss of coolant. "We're trying to determine with a high level of confidence what the behavior of the rod or pellet would be under those conditions," he says. "How hot does the fuel get? Does it melt? Does it release gases?" To accomplish this, researchers will build computer models of fuel behavior and then benchmark them against known behaviors. For example, if researchers know from experience how hot the cladding on a fuel rod gets when a reactor loses coolant for a given period of time, they want to ensure that models accurately reproduce this.
The computational power available on ORNL's Jaguar supercomputer will enable researchers to build models in unprecedented detail. CASL scientists will be looking at fuel behavior at the pellet level in reactors that have 51,000 fuel rods, each filled with 300 to 400 pellets. "It might look like we're losing the big picture by zeroing in on one pellet," Kothe explains, "but it's important to understand what's happening on that level before we can understand the overall performance of the reactor." CASL researchers expect VERA to start with a full-core simulation and then find vulnerable or weak points of the design—down to individual fuel pellets.
Of course, CASL's main goal is to give researchers a better understanding of the overall performance of a working reactor core. For example, VERA's simulation abilities allow nuclear fuel designers to experimentally enhance the performance of a realworld reactor by changing the arrangement or composition of the fuel in its virtual twin. Kothe notes that VERA's simulations will be limited to the reactor core. "CASL is not going to model entire nuclear plants," he says. "Other groups have already developed simulations that model the systems outside the reactor. The CASL virtual reactor is initially going to be a core simulator, not a plant simulator."
However, in anticipation of pairing VERA's modeling capabilities with those of other simulation tools, CASL researchers are coordinating their work with an Idaho National Laboratory research team that has developed software to simulate safety-related situations at a nuclear plant. This partnership ensures that the INL system simulator can be plugged directly into VERA. "We're making sure that VERA can be integrated into efforts to model core and system failures," Kothe says. "We can envision a real-time nuclear power plant simulation in which VERA is simulating the reactor and additional software is simulating the behavior of other systems."
Measures of success
Kothe says that CASL researchers want to deliver three things as measures of success. First is an accurate, detailed simulation technology that helps nuclear professionals anticipate what happens in an operating reactor core in a more predictable way. This will be the VERA virtual reactor. They also want to provide simulation-guided solutions to industry problems related to reactor power upgrades, reducing waste and enhancing safety. Finally, Kothe's team would like to help bridge the gap between fundamental research and commercialization by accelerating the deployment of basic science advances into industry.
One step toward bridging this gap will be making VERA accessible to a number of different users. Even though VERA will be optimized to run on Jaguar, the software must be portable enough to run in less exotic environments, including desktop and laptop computers. "Although the models generated on laptops will be less complex than those generated by Jaguar," Kothe explains, "developing software that will produce high-quality simulations across a range of computing platforms is a priority."
A new state of the art
Over the next three years, CASL expects to release several versions of VERA. The big advantage VERA will have over the current state of the art in reactor modeling is its ability to model all the significant aspects of reactor core operation and their simultaneous interactions. "In order to do a really good job of predictability and simulation you have to do these things simultaneously," Kothe says.
VERA software will eventually be able to create reactor simulations on desktop and laptop computers.
By the end of the consortium's first five years of operation, the expectation is that nuclear vendors and operators will use VERA in engineering their designs and increasing the output and efficiency of their nuclear plants. Kothe also expects that, through research, discussions with industry and published papers, CASL will help utilities with their goals of burning fuel more thoroughly, decreasing the time it takes to power-up a reactor and extending the lives of their facilities.
Kothe emphasizes that CASL's success will depend, to a great extent, on meeting the needs of nuclear utilities and reactor vendors who have already invested a lot of time and money in their own modeling capabilities. "We're making sure we incorporate these capabilities into VERA as well," he says. "It's important for us to understand our partners' requirements—otherwise we can't help them improve their operations. Working with the nuclear power industry is not an 'if you build it they will come' thing. It's a cultural thing. We're starting from their baseline, rather than our own."—Jim Pearce