FUSE meeting notes: Oct 29 2024

[orso82]

Orso

• Costing now supports propagation of uncertainties
  • Noticed that Sheffield and ARIES costs have diverged once materials costs got changed
  • We should review carefully materials costs @adrianaghiozzi
  • Also an opportunity to add uncertainties to materials costs
• Now all quantities in dd can be uncertain
  • includes support to do I/O
  • Important as we move to support experimental data analyses
  • _error_upper in IMAS becomes _σ in IMAS.jl
  • Methodology for handling uncertainties can be used for supporting automatic differentiation operating on the data structure
• Now ActorEquilibrium.model and ActorPedestal.model can both be set to :none effectively turning them off in compound actors
• Fixed issue with WPED pedestal model not working
• Fixed issue with controlling line average density
• Fixed issue with setting quasi-neutrality by varying density of impurity species

Brendan

• Working on Grad-Hogan solver
  • Conducing surfaces in the equilibrium solve, we keep the flux constant
  • Can model vertical stability
  • Now working on coupling to current diffusion
  • Still some issue with VDE timescales that we're ironing out
  • Working with control group to setup equilibrium control in FUSE
• Need to implement the q-solver in FRESCO
  • Requires flux-surface averaging in FRESCO

Min-Gu

• TrojonBetaNN
  • two models Yuequiang (n=1,2,3) and Yifei Zhao (n=1)
  • Yifei Zhao's is only for HL2M
• Opened PR in EFIT.jl to add support for gEQDSK
• First trial to develop a NN of TEQUILA
  • Need to generate more data

Tim

• Database generator notebook
  • Can work with Min-Gu
• Codecamp signup: 8 people confirmed outside of GA
  • Discussions with Jeff about how to provide internet access without needing to go through cyber-access requests. Hot-spotting with Google Fi seems promissing.

Adriana

• Participated to costing model task force
  • What are the weak points that they are trying to improve?
  • They try to follow convetions of GEN4 fission reactors
  • They'll send out slides
• Getting neutron rates out of RABBIT, storing it in core_sources.neutrals
• FACIT model coefficients as matlab file

Galina

• PAM
  • Some differences between Julia and OMFIT because of different ways of defining grid
  • OMFIT looks a bit suspicious
  • Will work with Joey to figure it out
  • Then will work on ActorPAM for FUSE

Tom

• TGLF-EP finished, now trying to set it up from IMAS
  • Cleaning up TJLF repo