Date of registration 2021-11-08
A Leap for Fusion Simulation Technology: Planning Study for V-DEMO Development in Korea
Fusion technology development in Korea faces three primary challenges: (i) Development of a high-performance operating mode for Tokamak, (ii) Building a blanket test facility to bridge KSTAR, ITER, and a fusion demonstration plant (DEMO), and (iii) Utilization of experimental data along with the integration of scientific and technological achievements to design a fusion DEMO. To achieve these three missions, KFE conducted a planning study for the development of V-DEMO.
V-DEMO will be a virtual fusion demonstration plant, built by applying the 4th Industrial Revolution technologies including supercomputers, artificial intelligence, big data, etc. The ultimate goal of V-DEMO is to quantitatively realize the core functions of a fusion power plant and comprehensively reproduce one by integrating simulations of tokamak, blanket, and BOP (Balance of Power) systems. With these features, V-DEMO can be used to identify various physical and engineering requirements that can be used in the detailed engineering designs of a fusion power plant. Clarifying these requirements can help in optimizing and verifying plant designs, as well as assisting security and safety clearance studies by investigating various accident scenarios.
Building V-DEMO will require examining available technologies and current technological levels, and then specifying strategies for each step. Therefore, the focus of this planning study was identifying the technological priorities required to resolve the challenges mentioned above.
Technology development trends and development direction for V-DEMO
Fusion simulation is the core technology for V-DEMO. It has been under development for the last two decades, mainly focusing on core plasma, heating, and PMI (Plasma Material Interaction). The current mainstream technologies are large-scale simulations capable of parallel expansion over tens of thousands of CPU cores. Korean researchers are presently working to elaborate and verify simulations using KSTAR’s advanced imaging diagnostics. The global goal in fusion simulation development is to develop the capability to quantitatively predict experiments using supercomputer-based simulations, and relevant studies are expected to play a large role in KSTAR studies.
The core plasma in tokamak is connected to power facilities via a breeding blanket system, which converts fusion energy into electricity. Simulating fusion electricity generation involves both fusion and nuclear power studies. The former can take advantage of the latter. As a matter of fact, the TBM (Test Blanket Module) currently being tested in ITER uses nuclear analysis tools and safety analysis codes from the nuclear power community, demonstrating that nuclear simulation technologies can be expanded to enable V-DEMO’s targeted functions.
AI (Artificial Intelligence) is also promising for fusion research innovation. AI can utilize machine learning based on numerous fusion experiments and simulation data to derive a data-driven fusion model. The big advantage of the data-driven fusion model is its combination of fast calculation and good precision. For example, according to research already published, a heating simulation can be generated in a much shorter time with machine learning, allowing real-time controls. Without it, the calculations would require thousands of CPU cores and considerable time. Various studies to accelerate simulations are ongoing, with the ultimate goal of developing a fusion simulator capable of the tremendous volume of repetitive calculations necessary for engineering design.
V-DEMO will need to integrate simulations from various fusion areas and therefore will need to develop an integrated platform. Fusion societies worldwide are working on simulation software integration frameworks. One of the most promising projects is the ITER-IMAS framework, which is currently under development for ITER. V-DEMO should expand its technological base to include plant features based on ITER-IMAS.
Digital twin is a technology that can be used to virtualize machines and facilities, using machine design data, and is considered the basis upon which simulation software and integrated framework can be deployed. For fusion, “Virtual KSTAR” is being developed to perform virtual experiments that combine KSTAR design, experimental data, and heating simulations. The technology developed through KSTAR will be applied to ITER in the near future, to pave the way for V-DEMO development.
Roadmap for V-DEMO development
Presently, various technological developments including simulations, virtualization, supercomputers, and machine learning are needed for V-DEMO development. In this planning study, a total of five technology groups were identified as the key technologies for V-DEMO: (i) tokamak simulation group, (ii) blanket-BOP simulation group, (iii) accelerated fusion simulation group, (iv) enabling technology group, and (v) fusion big data group. The development plans and strategies for each group were also established.
Roadmap for V-DEMO project
V-DEMO development will go through four stages by the year 2040. In the first stage, simulation and virtualization of a middle-sized tokamak such as KSTAR will be carried out. In the beginning, technology verification using KSTAR data will be the key objective. In the second stage, the technologies from the first stage will be further developed for ITER, and simulation acceleration using artificial intelligence will begin. In the third stage, blanket and BOP simulation technology development will be performed in collaboration with the nuclear energy community. In the final stage, V-DEMO will be completed, integrating the K-DEMO design data. Verification and optimization of a demonstration plant will be carried out utilizing V-DEMO.
Plan for experimental validations and key development targets in each phase
Fusion research in Korea has been ongoing for the last two decades, focusing on KSTAR and ITER. Now, another leap is needed to accomplish a demonstration of fusion technology. Alongside hardware technological achievements, another foundation for the fusion demonstration should be established, by proactively developing fusion simulation technologies, including V-DEMO.
The attached file(1)