The role of fusion energy in a decarbonized electricity system October 17, 2024 energy.mit.edu | @mitenergy MITEI | MIT The Role of Fusion in a Decarbonized Electricity System Summary of Findings A study from the MIT Energy Initiative in collaboration with the MIT Plasma Science and Fusion Center September 2024 Overview of Fusion - Opportunity and Challenge • When two nuclei combine they release enormous amounts of energy: – More than 1,000,000 times the energy of combustion on a mass basis – More than nuclear fission on a mass basis • Required conditions for fusion: – Very high temperatures (>> 100 million degrees C) – Confinement of fusion fuels to bring them together long enough to fuse • Primary confinement methods: – Magnetic confinement – Inertial confinement – Magneto-inertial confinement • The Challenge: – Building an economically viable system that can handle the temperatures and high-energy particles released, plus the ability to convert that energy into electricity Timeline and Structure of this Study • Project period: fall of 2022 through spring 2024 • Structured with core workstreams: – Global deployment – Subregions of United States – Critical materials and supply chains – Key cost drivers • We do not predict: – When fusion will first be deployed commercially – Which fusion technology will deploy first – What it will cost • Instead, we focus on conditions required for fusion commercial viability Fusion has a potential societal value in the trillions of dollars • Fusion technology can reduce the total cost of decarbonization by a cumulative discounted $3.6 trillion if fusion power plants cost $8,000/kW in 2050 and fall to $4,300/kW in 2100 • Fusion technology can reduce the total cost of decarbonization by a cumulative discounted $8.7 trillion if fusion power plants cost $5,600/kW in 2050 and fall to $3,000/kW in 2100 Note: Dollars in this presentation are reported in 2021 dollars, except were explicitly indicated Chart: Annual GDP benefits (trillion 2021$) plotted from 2035 to 2100 for $5.6K and $8K scenarios. The scale of fusion deployment will depend on costs • For a 1.5°C stabilization decarbonization scenario, the total global share of electricity generation from fusion in 2100 ranges from less than 10% to about half depending on the assumed cost for fusion. • Fusion costs shown are for the overnight cost of constructing a fusion power plant in the U.S. in the year 2050. At the end of the century, costs are about half the assumed 2050 costs. Scenarios shown: $11,300/kW, $8,000/kW, $5,600/kW, and $2,800/kW in 2050. The scale and timing of fusion deployment is highly variable across regions • Initial deployment is strongest in the United States and Europe • Largest increase in fusion takes place in India during the last three decades of the century • Africa is a late adopter of fusion but sees strong growth late in the century • These trends are driven by: o economic growth o population density o electrification needs o regional costs o decarbonization targets o relative prices of electricity o limits on fission-based nuclear generation o renewable resource availability Fusion deployment will highly depend on the availability and cost of other low-carbon technologies - High Penetration, Low Sensitivity: • U.S. Subregions: Atlantic and Southeast • Renewable attributes: Poor onshore wind, hydro, and geothermal resources • Fusion penetration at $6,000/kW: Required at all emission caps from 1 to 50 gCO2/kWh - Medium Penetration, Medium Sensitivity: • U.S. Subregions: California, Northeast, Southwest • Renewable attributes: Northeast has best offshore wind; California has best geothermal; Southwest has best solar; all three have modest onshore wind capacity or quality • Fusion penetration at $6,000/kW: No penetration at 50 gCO2/kWh, but capacity reaches 33%–55% of demand at 1 gCO2/kWh - Low Penetration, Low Sensitivity: • U.S. Subregions: Northwest • Renewable attributes: Below average solar and wind resources, but excellent diversity of renewable resources including good hydro and moderate geothermal • Fusion penetration at $6,000/kW: Required at all emission caps 1–20 gCO2/kWh but capacity is never more than 26% of demand - Low Penetration, High Sensitivity: • U.S. Subregions: Central, North Central, Texas • Renewable attributes: Abundant, high-quality, and low-cost onshore wind; limited renewables beyond onshore wind and solar • Fusion penetration at $6,000/kW: Required only at 4 gCO2/kWh and below, but capacity reaches 25%–45% of demand at 1 gCO2/kWh * Sensitivity refers to the sensitivity of fusion penetration with respect to changes in the emissions cap. The role of fusion power plants is also highly sensitive to costs • Fusion power plants serve as: o Low-capacity-factor, dispatchable electric generation when fusion costs are high o Baseload resource when FPP costs are moderate o Dispatchable generation with a moderate capacity factor when FPP costs are low. • This trend was observed in our analysis of the New England subregion of the U.S. Graphs show: Fusion installed capacity as % of peak load vs FPP capital costs; Fusion fleet generation capacity as % of annual demand vs FPP capital costs; Fusion annual capacity factor vs FPP capital costs. The availability of firm, low-carbon natural gas power plants can have a large impact on the deployment of fusion power plants • NGCC power plants with high carbon capture and low upstream methane emissions can have a large impact on fusion deployment • Threshold cost point at which fusion becomes competitive is $4,000/kW lower when NGCC with 95% carbon capture is available than when NGCC with 95% carbon capture is not available. Comparison charts displayed for (A) With NGCC with 95% capture and (B) Without NGCC with 95% capture. Supply chains for the processed materials and manufactured parts needed to build fusion power plants vary widely in maturity • Different technologies are at varying stages of maturity with identifiable issues and bottlenecks • R&D is needed to develop materials and manufacturing capabilities essential for fusion at the scale outlined in this report • Fusion components can be broken up into two categories: o Niche (e.g. tungsten heavy) with limited non-fusion market opportunities o Components (e.g. high-temperature superconductor, radio-frequency devices) with strong potential for commercial non-fusion use • For raw materials, there are no anticipated showstoppers, however beryllium resources and markets remain an uncertainty Key cost drivers for fusion power plants include reactor equipment cost, regulatory considerations, and operations and maintenance costs • Fusion reactor equipment is the leading cost contributor at 30% to 65% of the total capital cost • Regulation can be a potentially large cost driver and motivates: o fusion companies to minimize their footprint with respect to fuels and activated materials o governments to adopt appropriate and effective regulatory policies to maximize their ability to use fusion energy in achieving decarbonization goals • Operating and maintenance (O&M) costs can be significant for a fusion power plant Key Takeaways • Fusion has potential societal value in the trillions of dollars in a decarbonized world. • Deployment and operation of fusion power plants is highly dependent on: o Fusion costs o Cost and availability of alternative low-carbon technologies in each region o Carbon emission constraints o Economic and electricity demand growth o Market design • The ability of fusion to scale requires development of materials and manufacturing capabilities for niche components • For raw materials, there are no anticipated showstoppers Acknowledgements Study Participants: Robert Armstrong, Randall Field, John Parsons, Koroush Shirvan, Nirmal Bhatt, Sara Ferry, Theodore Mouratidis, Jeffrey Freidberg, Dharik Mallapragada, Jennifer Morris, Elsa Olivetti, Tim Schittekatte, Robert Stoner, Dennis Whyte, Ruaridh Macdonald, Sergey Paltsev, Layla Araiinejad, Amanda Farnsworth, Angelo Gurgel, Y.-H. Henry Chen, Emre Gençer, Samuele Meschini, Matthew Munderville, Richard Roth, Michael Short. Subject Matter Experts: Zachary Hartwig, Alexander Molodyk, Keith Smith, Steve Wukitch, Stuart Malloy, Daniel Ng, Lance Sneed, Yifei Zhang. Project/Report Team: Gail Monahan, Tom Witkowski, Elizabeth Hamblin, Kelley Travers, Kelly McGinity, Marika Tatsutani. Funding: MITEI and PSFC gratefully acknowledge Eni for supporting this research and for engaging in discussions with MIT during this research project. Overview of PSFC’s mission Nuno Loureiro Director, Plasma Science and Fusion Center MIT This is the PSFC. Great things happen here. Founded in 1976 | 160,000 sq ft of lab space | 150 employees | 90 graduate students | 5 research divisions | 86 ongoing research projects The fusion world is unrecognizable compared to what it looked like five years ago • We launched the MIT-CFS partnership about 5 years ago • SPARC’s success has spurred on the financing and commercial world to take fusion seriously as a disruptive new energy source • Goal is to bring fusion energy to the world NOW. US government aggressively pursuing fusion - White House Summit: Developing a Bold Decadal Vision for Commercial Fusion Energy - DOE Announces $46 Million for Commercial Fusion Energy Development - Companies: Commonwealth Fusion Systems, Focused Energy, Princeton Stellarators, Realta Fusion, Tokamak Energy, Type One Energy, Xcimer Energy Corporation, Zap Energy Private Fusion Industry has experienced staggering growth • >6 B$ raised • Over 40 companies • Industry association • Wider geographic spread • Wide array of technical approaches • Competition is good • Private funding encourages breakthroughs with higher risk on faster timelines To create a device that can produce more energy than it consumes, we need: - Strong and efficient magnets to confine fusing plasmas (TFMC and CSMC) - Peak plasma fusion performance (Machine learning, quantum computing) - A way to convert energy into power (LIBRA lithium blanket) - Extra-durable materials that can withstand plasma and neutron exposure (Cyclotron materials testing) ARC fusion device - Rendering by CFS NIF laser facility demonstrated scientific fusion energy gain in Dec 2022 • 2 MJ laser, 3 MJ fusion energy • The important science is that the improved fusion energy was driven by the alpha heating resulting from the fusion reactions themselves • This feature is common to all fusion and so should be celebrated as another important science milestone • MIT (PSFC) is the leading university in measurements of fusion products • Now being adapted for SPARC! We have made amazing progress but.. • Need to expand and innovate on fusion education. At MIT and beyond! o The US academy is not prepared o We are strapped for fusion instructors. (Features: Launched Proxima Fusion, Leaders @ NIF) We have made amazing progress but.. • Need to expand and innovate on fusion education. At MIT and beyond! o The US academy is not prepared o We are strapped for fusion instructors. • Capable, but fast, experimental facilities for addressing the next set of technical innovations that make fusion economical. (LIBRA experiment progression: BABY-0.1L, BABY-1L, LIBRA-Pi: 125L, LIBRA: 500L) We have made amazing progress but.. • Need to expand and innovate on fusion education. At MIT and beyond! o The US academy is not prepared o We are strapped for fusion instructors. • Capable, but fast, experimental facilities for addressing the next set of technical innovations that make fusion economical. • Need to launch the next generation of companies that will provide critical new technologies for industry. Report reference: Enabling Commercial Fusion: Venture & Technology Opportunities for a Growing Fusion Industry (Proto Ventures / PSFC) Materials testing under fusion relevant conditions Zach Hartwig Nuclear Science and Engineering, Plasma Science and Fusion Center 10/24/2024 © MIT Plasma Science and Fusion Center MIT Visit to DOE FES (Mon Aug 5 2024) “Fusion energy is becoming a materials problem” • Dozens of private fusion companies (>$7B investment) are developing fusion for power production in the 2030s. • National programs (US, UK, China, etc.) are supporting these efforts and/or launching their own programs with objectives in the 2030s. • Every approach to fusion energy requires solving materials challenges. These challenges are ultimately linked to economic viability. Some materials challenges in a fusion power plant Common challenges: • D-T fusion makes 14.1 MeV neutrons, causing bulk H and He generation in addition to cascade damage • Neutron energy spectrum varies widely depending on subsystem of interest, and material response will vary accordingly References: [0] B. N. Sorbom et al, FED 100 (2015) 378-405; [1] A Q. Kuang et al, FED 137 (2018) 221-242 ARC Fusion Power Plant rendering Some materials challenges in a fusion power plant Structural material examples: • Radiation induced damage: embrittlement, swelling, enhanced creep, and more! • Chemical compatibility with molten salt under simultaneous irradiation Common challenges: • D-T fusion makes 14.1 MeV neutrons, causing bulk H and He generation in addition to cascade damage • Neutron energy spectrum varies widely depending on subsystem of interest, and material response will vary accordingly Some materials challenges in a fusion power plant Structural material examples: • Radiation induced damage: embrittlement, swelling, enhanced creep, and more! • Chemical compatibility with molten salt under simultaneous irradiation Plasma-facing material examples: • Tritium retention and migration in bulk damaged materials • Radiation damage changes to material properties required for high heat flux handling (10 MW/m2) Common challenges: • D-T fusion makes 14.1 MeV neutrons, causing bulk H and He generation in addition to cascade damage • Neutron energy spectrum varies widely depending on subsystem of interest, and material response will vary accordingly Some materials challenges in a fusion power plant Functional material materials: • Degradation of the superconducting magnets while irradiated at cryogenic temperatures (~20 K) • Browning of optical glass and mirrors required for plasma diagnostics Plasma-facing material examples: • Tritium retention and migration in bulk damaged materials • Radiation damage changes to material properties required for high heat flux handling (10 MW/m2) Structural material examples: • Radiation induced damage: embrittlement, swelling, enhanced creep, and more! • Chemical compatibility with molten salt under simultaneous irradiation Neutron irradiation of materials is insufficient • Today: Irradiation in reactors Disadvantages: Low damage rates, High cost and rare facilities, Long + few learning cycles, Low fidelity for fusion • 2040s: Irradiation in IFMIF (~80 meters) Disadvantages: Low damage rates, Limited exposure conditions, Operational in the 2030s? 2040s?, Impossible to scale to meet demand • We have to think differently and boldly to tackle fusion materials on relevant timescales. Facility example #1: Cryogenic proton irradiation in high fidelity conditions • Experimental schematic showing Proton beam, Thermal radiation shield + secondary electrons, Electric insulator (650 µm thick sapphire), Mounting screws, Superconductor @ 20 K with current, Gold coated glass collimator, Cold-head and heaters (Base temperature: 15 K, Cooling power: 25W at 20 K, Heating power: 400 W, CX-CH: PID temperature sensor). • Reference: A. Devitre et al. Rev. Sci. Instr. 95 (2024) 063907. • Facility generously supported by Eni. S.p.A Facility example #1: Cryogenic proton irradiation in high fidelity conditions Graph comparing Superconducting Performance vs Radiation damage (damage level in mDPA): - Room temperature irradiation @ 300 K - Cryogenic irradiation (irradiated at 20-200 K) Reference: D. X. Fischer, et al. Submitted to Supercond. Sci. and Tech., 2024 Facility example #1: Cryogenic proton irradiation in high fidelity conditions • 60% faster degradation under magnet conditions! • Innovative facilities can get answers in the right conditions on fast timescales Reference: D. X. Fischer, et al. Submitted to Supercond. Sci. and Tech., 2024 Facility example #2: A high throughput bulk materials irradiation facility • Process breakdown: Bulk materials (0.1-1 mm irradiated material, 10-30 MeV proton beam) → Uniform damage → High-energy recoils → Controlled transmutation (H, He) → Less radioactivity → Macroscale testing (stress-strain curves). • A technique that is faster, lower cost, lower activity, higher fidelity, and scalable. Reference: S. J. Jepeal, L. L. Snead, Z.S. Hartwig. Materials and Design, 200 (2021) 109445. Facility example #2: A first-of-kind irradiation facility for fusion materials PSFC is readying to construct a new facility capable of high fidelity, high throughput bulk materials irradiation: • Replicate fusion irradiation responses in materials • Damage materials up to 100x faster than application • Reuse existing experimental space at PSFC worth $50M • Utilizes commercial off-the-shelf accelerator technology • Ready to begin science operation 18 months from funding 4 beamlines with user-driven modular end station to maximize science output: • Thin-film and bulk materials irradiation (in-situ and post-irradiation examination) • Complex environmental conditions like linear plasma device for plasma-material and molten salt radiation-corrosion issues Provide decision-quality materials data for fusion companies in years not decades. Artificial Intelligence for preventing plasma disruptions Cristina Rea Principal Research Scientist, Group Leader MIT Plasma Science & Fusion Center MITEI - 10/17/2024 Fusion plasmas → highly nonlinear interactions + loss of control = disruptions! Normal plasma vs disruptive plasma: Final loss of control evolving on timescales of milliseconds: • Fast current drop leads to loss of confining poloidal field • Fast current transient causes large induced currents & forces • Rapid thermal losses cause surface damage (Data courtesy: R.S. Granetz showing current, temperature, radiated power, current centroid location, induced currents) Different instabilities grow when plasma pushed to controllability limits Visible camera view of runaway electron (RE) beam hitting Alcator C-Mod first wall. Courtesy R.A. Tinguely Disruption prevention strategies needed for next generation devices Hierarchy of strategies: 1. Accept the damage and live with it 2. Mitigate the damage by [...] to radiate away as much/fast as possible 3. Avoid altogether by detecting precursors & steer plasma away from disruptive boundary (Preferred) Visual comparisons: Alcator C-Mod RE beam hitting first wall vs JET runaway electrons damage resulting in melting (iter.org/newsline/-/2234) While first principle models lack, data-driven / ML ones can provide: Explainable proximity to unstable operational space • Probability of impending disruption + contributions • Available on DIII-D and EAST in real-time systems for proximity control [1,2] References: [1] Rea et al, Nucl. Fusion 59 (2019) 096016; [2] Rea et al, 2021 IAEA EX/P1–25 While first principle models lack, data-driven / ML ones can provide: Explainable proximity to unstable operational space & interpretable tracking of instability onset • Collisionality at plasma edge identified as indicator for precursor to L-mode density limit [3] • Yields 6x fewer false positives than literature results • Algorithm available in DIII-D real-time control system References: [3] Maris, Rea et al, under review Nuclear Fusion, 2024; Rea et al, 2019 & 2021 Additional examples of ML/AI research aiding disruption prevention • Disruption-py: interoperable, open-source Python library for data access across different fusion devices** • Enables assembly of big databases for ML applications Architecture workflow: Tokamak (Alcator C-Mod, DIII-D, HBT-EP) → EFIT/PCS/Diagnostics MDSplus Trees → Physics Routines → Interpolate to Requested Time Base → Assemble Data Frame → Storage (csv, h5, ...) → Analysis References: Trevisan et al, APS-DPP 2024 PP12.9; Wei et al, APS-DPP 2024 PP12.10; github.com/MIT-PSFC/disruption-py; zenodo.org/records/13935224 The Disruptions team includes >20 scientists / postdocs / students / collaborators Website: https://disruptions.mit.edu/ ML/AI tools used ubiquitously across different projects: - Runaway Electron Mitigation Coil - Off-Normal Warning System - ML-Driven Disruption Prediction - Tearing-Free Pilot Plant Scenarios - Runaway Electron Wall Impacts - Risk-Aware Disruption Prediction - Density-Limit Disruption Analysis - Radiation Modeling and Bolometry - ML-Driven Trajectory Design Notable collaborations: CFS, Google DeepMind, Eni, CAFFE MIT-PSFC, international universities and entities. IAEA Collaborating Centre in AI in Fusion and Plasma Science: Enabling Research and Education 5yr cost-free agreement (2023-2027). Builds on IAEA Coordinated Research Project “AI for Fusion” [1] and DOE-funded activities sharing focus on: • Open Science and FAIR access to fusion data/metadata • Accessible research products (data, software, …) enabling cross-pollination • Education, training, capacity building in ML-applied fusion science Selected examples: o Computational Physics School for Fusion Research (CPS-FR) [2] - HPC, Parallel Programming, Computational Statistics, and ML (grad students + postdocs) o $5M multi-institutional collaboration (W&M, U. Auburn, U. Wisc-Madison, HDFG) to develop open research products [3] - AI for Fusion Energy Summer School (undergrads) o ITU data challenge [4] leveraging raw Alcator C-Mod dataset References: [1] nucleus.iaea.org; [2] sites.google.com/psfc.mit.edu/cps-fr-2024; [3] crea-psfc.github.io/open-fair-fusion; [4] aiforgood.itu.int Tritium management Sara Ferry MIT Plasma Science & Fusion Center MITEI - 10/17/2024 Most proposed fusion power plant concepts use deuterium and tritium as fuel. • The D-T fusion reaction in the plasma produces high-energy neutrons and helium particles: Deuteron (2H) + Triton (3H) → Neutron (1n, 14.1 MeV) + Helium ash (4He, 3.5 MeV) • Diagram illustrates neutron travelling to Blanket and helium/particles going to Exhaust. Rendering of an ARC-class fusion power plant courtesy of Commonwealth Fusion Systems Tritium is scarce, so D-T fusion power plants must breed their own. Reactions: - Plasma: D + T → n + ⍺ - Breeder Blanket: 6Li + n → T + ⍺ + heat 7Li + n → T + ⍺ + n’ - heat Functions of the breeding blanket: 1. Shielding: Moderate + absorb neutrons to shield the magnets 2. Heat capture: Generate power 3. Tritium breeding: Make more fuel in the breeder, which contains lithium Tritium is scarce, so D-T fusion power plants must breed their own. • Breeding material contains lithium: – Lithium ceramic solids (e.g. Li4SiO4) – Pure liquid lithium – Molten lead-lithium (PbLi) – Molten salts like FLiBe • Beryllium and lead are neutron multipliers and boost breeding • Blanket also needs infrastructure for chemistry control and heat removal • Blanket performance characterized by the tritium breeding ratio (TBR): TBR = (# of tritons produced in the blanket) / (# of tritons consumed in the plasma) MIT PSFC is investigating molten salt “liquid immersion blankets” for ARC-class fusion power plants. Key advantages of the LIB approach: • Maximize solid angle coverage of the neutron source (the plasma) to increase TBR • Can pump out breeder for maintenance • Online chemistry control • Minimize complexity • Efficient heat removal Drawbacks of conventional blankets: • Complex solid structures are challenging for fabrication, maintenance, and assembly • Activation concerns • More structural material = less breeder = lower TBR Example: helium-cooled lead-lithium (HCLL) test blanket module for ITER devotes lots of volume to non-breeding structures (Aiello G et al., Fusion Engineering and Design 2011). The high breeding efficiency of the LIB will let us put fusion plants on the grid much faster. • Assume we were able to begin building fusion power plants in 2030, and tritium supply was our only rate-limiting step. • Tritium inventory doubling times: – Liquid Immersion Blanket: 2 years – ARIES blanket: 5 years – EU-DEMO blanket: 10 years • Using conventional blankets, it would take ~70 years to reach 0.1 TW of installed fusion power capacity (<10% of U.S. installed capacity as of 2023). This isn’t fast enough! The LIBRA experiment at MIT is advancing molten salt breeder blanket technology. • LIBRA (Liquid Immersion Blanket: Robust Accountancy) enables us to study tritium breeding and behavior in FLiBe in a prototypical neutronic and chemical environment. • LIBRA requires a staged experimental approach of increasing complexity and volume: BABY-0.1L → BABY-1L → LIBRA-Pi: 125L → LIBRA: 500L (supported by ARPA-E). • Experimental setup for BABY-0.1L in the MIT PSFC Vault Laboratory on Albany Street shown. The LIBRA experiment at MIT is advancing molten salt breeder blanket technology. • There are enormous uncertainties regarding how tritium behaves in FLiBe. Literature data shows a 3-4 order-of-magnitude spread for solubility of tritium in FLiBe at the same temperature! • The LIBRA initiative will provide reliable, repeatable data on tritium behavior in prototypical breeder blanket environments that can be used to reliably design full-scale blanket systems for power plants. See also: U.S. fusion blanket and fuel cycle technology community roadmaps published by EPRI. The breeder blanket (marked with a ⭐️) is just one piece of a complex fuel cycle system. • We need to understand tritium’s dynamics through the entire fuel cycle: achieving high breeding rates in the blanket is just part of the engineering challenge! • Tritium is highly mobile, and losses due to trapping and permeation (as well as radioactive decay) must be accounted for. • Most of the necessary technologies have not been demonstrated at scale, but are critical to the success of the fusion industry. Reference: S. Meschini, S. E. Ferry, R. Delaporte-Mathurin, D. Whyte, “Modeling and analysis of the tritium fuel cycle for ARC- and STEP-class DT fusion power plants,” Nuclear Fusion 63 (2023). MIT PSFC is working to fuel the future fusion industry. • D-T fusion plants must breed their own tritium: D + T → n + He; Neutron + Lithium → Tritium + He • Advanced tritium breeding technology is the key to unlocking fusion power’s market potential. • MIT PSFC is undertaking an ambitious, staged approach to derisk the FLiBe liquid immersion blanket (BABY-0.1L, BABY-1L, LIBRA-Pi: 125L, LIBRA: 500L). LIBRA team: Kevin Woller (PI), John Ball, Remi Delaporte-Mathurin, Collin Dunn, Kaelyn Dunnell, Emily Edwards, Sara Ferry, Nikola Goleš, Ed Lamere, Andrew Lanzrath, Rick Leccacorvi, Samuele Meschini, Ethan Peterson, Stefano Segantin, Rui Vieira, Colin Weaver, Dennis Whyte, Weiyue Zhou. Impacts of neutron transport on fusion power plant design Prof. Ethan Peterson MIT Plasma Science & Fusion Center Neutrons greatly impact fusion power plant design • Neutrons carry 80% of D-T fusion energy • Monte Carlo simulations compute the distribution of neutrons throughout the power plant and their effects • Born at 14 MeV where they travel 5 cm on average between collisions in steel; requires ~500 collisions to lose energy • Fusion neutrons travel far, at high energy and cause damage, heating, dose, and transmutation Neutrons greatly impact fusion power plant design 1. Capture neutron energy and generate tritium 2. Protect magnets from nuclear heating and radiation damage 3. Engineer plasma facing materials that can withstand high fluence 4. Protect workers and the public from direct and shutdown dose • Neutron responses determine system requirements, yet take the longest to execute. • We need fast, accurate methods to compute them as well as their uncertainty. Fragmentation of fusion neutronics workflows impedes analysis speed and uncertainty quantification • Consider the analysis of “shutdown dose rate” (SDR) for the maintenance of fusion power plants: Model definition & inputs + Cross section data → Neutron transport solve → Reaction rates (Energy deposition, damage, tritium production) → Activation solve → Activated material compositions (Waste disposal ratings, transmutations, decay heat) → Photon source generation → Photon spectrum → Photon transport solve (Radiation sources for shielding analysis) → SDR Fragmentation of fusion neutronics workflows impedes analysis speed and uncertainty quantification • Disconnected workflow requiring separate codes: – Code A: Neutron transport solve – Code B: Activation solve – Code C: Photon source generation – Code A: Photon transport solve • Barriers prevent uncertainty propagation, solution acceleration, and robust verification and validation (V&V). Streamlining these workflows enables scalable and sustainable nuclear analysis for fusion power plants • Integrated pipeline inside OpenMC: Model definition & inputs + Cross section data → Neutron transport solve → Reaction rates → Activation solve → Activated material compositions → Photon source generation → Photon spectrum → Photon transport solve → SDR • Open-source, community-driven ecosystem • Designed for exascale architectures • Extensible for coupling with other applications • Single source of truth for transparent V&V Reference: E. E. Peterson, P. K. Romano, P. C. Shriwise, P. A. Myers, “Development and validation of fully open-source R2S shutdown dose rate capabilities in OpenMC,” Nucl. Fusion 64 (2024). Goal: An analysis platform for the first rigorous assessment of nuclear-driven engineering margins in fusion power plants • Streamlining analyses was the first step towards a comprehensive assessment and reduction of uncertainties for complex workflows • Uncertainties come from many places: nuclear data, material and geometry specifications, method approximations • Discrepancy grows dramatically with size and complexity (from 0.01x up to 10x) • To have confidence in our predictive modeling tools we have to find better agreement between experiments and simulations for large complex systems The role of fusion energy in a decarbonized electricity system October 17, 2024 energy.mit.edu | @mitenergy MITEI | MIT