FUSE meeting notes: Aug 1 2023

[orso82]

Orso

• hackathon last week
  • Worked with David on issues for rebaselining FPP design
  • Worked with Brendan on time dependence
• Time dependence
  • Started coding and getting close to something
  • Brian: Heater time stepping in TRANSP, Astra algorithm in a IPP note

David

• Lot of good progress in hackathon
• Running on OMEGA
  • Issue with distributed optimizations in version 1.9x
  • Issue with Jupyter Lab in version 1.8.x
  • Will discuss with JuliaHub folks soon

Discussion: Savanah River

• Integration into FUSE
• Process to get non-citizens vetted takes months
• Necessarily a citizen
• Text file for inputs, executable, text outputs
• Tom volunteering, perhaps a good synergy with Jackson's BOP modeling?
• possible target of JuliaSim surrogate modeling?

Jackson

• Presented ThermalSystemsModel.jl package at hackathon
• Working to get steady state models into FUSE

Kevin

• PR to get max (vs total) leakage through the blanket
• Minor terminology changes
• Deterministic neutron solver
  • Can generate 2D geometry, but meshing hangs
  • 1D can make the geometry and mesh it, and run
  • 1D had some convergence issues
• Getting Amber up-to-date
  • Learning at least the paramak build

Brendan

• Optimization of TEQUILA
• Re-thinking fitting of flux surfaces in TEQUILA to make it more robust
• Discussed time dependence in FUSE

Discussion: Tokamak energy

• Visited by Tim
• Using TGLF in ASTRA and having difficulties converging
• Issue with community developed codes, no support, also TRANSP
• Cannot use ITER tools
• 4-5 people
• Michele Romanelli contact
• Interested in running OMFIT/STEP

Discussion: comparison with ASTRA modeling of AUG

• T. Luda paper
• how do we compare?
• FUSE SOL is 0D. Can we upgrade to 1D? Question for Jerome & INFUSE

Tom

• Setup Nan to validate TGLFNN in FUSE
• Found some bug
  • Number of thermal species
  • Timeout of TGLF simulations with large number of species
• Training FPP+DIII-D database for SAT1 (SAT0 is available)

Nan

• Thoughts about modeling of impurities in FUSE

Adriana

• Amani put Adriana in touch with Leo for financial modeling
• Leo wants to understand what we are doing
  • Prepared presentation
  • Bring him up to speed with what we are thinking
• Organizing notes for INFUSE collaboration about costing buildings
  • Who to partner with?
    • Mostly with fission background
  • What would be SOW?
  • What would be the deliverable?
  • A bite-size project, not multi-institutional, hopefully not sub-contractors
  • Next round call in December, with proposals in early February
  • Should ask Raffi to join, and Leo in his technical role

Tim

• SOFE
  • Talks are online
  • What the big players are doing
    • CHINA: huge progam, both BEST (2027) and CFTR being built (2030)
    • UKAEA
    • US is behind
    • CFS is small scale compared to others
  • Interesting of PROCESS using multi-objective optimization
    • Started with genetic algorithms
    • Looking at bayesian methods --> probability --> Risk
• JuliaCon
  • Computing oriented
  • Many interesting packages
    • Impressed by Gene framework
    • JuliaSim great for surrogate models

[jguterl]

Discussion: comparison with ASTRA modeling of AUG - T. Luda paper - how do we compare? - FUSE SOL is 0D. Can we upgrade to 1D? Question for Jerome & INFUSE Please add references (poster/paper) here.

…

On Wed, Aug 2, 2023 at 6:58 AM Orso Meneghini ***@***.***> wrote: Orso - hackathon last week - Worked with David on issues for rebaselining FPP design - Worked with Brendan on time dependence - Time dependence - Started coding and getting close to something - Brian: Heater time stepping in TRANSP, Astra algorithm in a IPP note David - Lot of good progress in hackathon - Running on OMEGA - Issue with distributed optimizations in version 1.9x - Issue with Jupyter Lab in version 1.8.x - Will discuss with JuliaHub folks soon Discussion: Savanah River - Integration into FUSE - Process to get non-citizens vetted takes months - Necessarily a citizen - Text file for inputs, executable, text outputs - Tom volunteering, perhaps a good synergy with Jackson's BOP modeling? - possible target of JuliaSim surrogate modeling? Jackson - Presented ThermalSystemsModel.jl package at hackathon - Working to get steady state models into FUSE Kevin - PR to get max (vs total) leakage through the blanket - Minor terminology changes - Deterministic neutron solver - Can generate 2D geometry, but meshing hangs - 1D can make the geometry and mesh it, and run - 1D had some convergence issues - Getting Amber up-to-date - Learning at least the paramak build Brendan - Optimization of TEQUILA - Re-thinking fitting of flux surfaces in TEQUILA to make it more robust - Discussed time dependence in FUSE Discussion: Tokamak energy - Visited by Tim - Using TGLF in ASTRA and having difficulties converging - Issue with community developed codes, no support, also TRANSP - Cannot use ITER tools - 4-5 people - Michele Romanelli contact - Interested in running OMFIT/STEP Discussion: comparison with ASTRA modeling of AUG - T. Luda paper - how do we compare? - FUSE SOL is 0D. Can we upgrade to 1D? Question for Jerome & INFUSE Tom - Setup Nan to validate TGLFNN in FUSE - Found some bug - Number of thermal species - Timeout of TGLF simulations with large number of species - Training FPP+DIII-D database for SAT1 (SAT0 is available) Nan - Thoughts about modeling of impurities in FUSE Adriana - Amani put Adriana in touch with Leo for financial modeling - Leo wants to understand what we are doing - Prepared presentation - Bring him up to speed with what we are thinking - Organizing notes for INFUSE collaboration about costing buildings - Who to partner with? - Mostly with fission background - What would be SOW? - What would be the deliverable? - A bite-size project, not multi-institutional, hopefully not sub-contractors - Next round call in December, with proposals in early February - Should ask Raffi to join, and Leo in his technical role Tim - SOFE - Talks are online - What the big players are doing - CHINA: huge progam, both BEST (2027) and CFTR being built (2030) - UKAEA - US is behind - CFS is small scale compared to others - Interesting of PROCESS using multi-objective optimization - Started with genetic algorithms - Looking at bayesian methods --> probability --> Risk - JuliaCon - Computing oriented - Many interesting packages - Impressed by Gene framework - JuliaSim great for surrogate models — Reply to this email directly, view it on GitHub <#384>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEESMZD5B2KFDRAEAVC3PKDXTHM5NANCNFSM6AAAAAA3AWIZ44> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>