FUSE meeting notes: Sep 17 2024

[orso82]

Orso

• FRESCO
  • Will need to run from P and J as function of rho_tor_norm (open an issue)
• More work getting tutorial ready for public consumption
• Updated FUSE checkpoint with @checkin and @checkout macros
• APS preparations
  • Pamphlet
  • Rollup
  • Press release

Tim

• SAGA optimization went through for SNARC
• Optimization with 250k cases with 3 objectives
  • Some challenge to analyze those cases
• Hypercube of BOP actor
  • Generate a hypercube of the optimal balance of plant #689

Brendan

• FRESCO works in it's most basic form
  • Will need to add (fast) flux surface averaging
  • To support P and J as function of rho_tor_norm (for StationaryPlasma)
  • To support entropy and q as function of psi_norm (for DynamicPlasma, enable Grad Hogan solver)
• Meeting about how to couple plasma and passive/active structures
  • Now QED can take a coupling matrix (coming from VacuumFields)
  • Benchmarked with work that Luke has done
  • We need to code the plasma to conductors coupling
  • We'll see the currents that changes in the plasma current/shape induce in the structures
  • We'll be able to ramp the OH coil to drive current in the plasma

Adriana

• Optimization results with risk metric used as an objective
  • 500 viable solutions

Adriana

• Optimization results with risk metric used as an objective
  • 500 viable solutions
• Will start woking on the poster

Tom

• MXH now in TJLF (and TGLF + ProfilesGen)
  • Must hook it up in FluxMatcher
  • Met with Gary, Sophia
• GKNN is still noisy
  • training with more data should improve things
  • TGLF is not as accurate in the core (use GKNN only rho<0.5)
• UCSD are doing a grid search to find optimal training parameters

Nan

• Tested ion evolution in FUSE
  • FluxMatcher can work on DIII-D default data
  • Some issue with my own DIII-D ODS... something weird
  • Cannot load my own ODS in FUSE

Galina

• Found original paper for pelled deposition. Looking at projection of pellet trajectory on the flux-surface to figure out ablation cloud. Will read more, look at the Python, ask Joey.