Skip to content

Re-run the notebooks after the packed-bed fix, and tag the coupled shim - #55

Merged
brosaplanella merged 2 commits into
mainfrom
compat-coupled-tag-rerun-notebooks
Nov 1, 2026
Merged

brosaplanella merged 2 commits into
mainfrom
compat-coupled-tag-rerun-notebooks

Conversation

@brosaplanella

Copy link
Copy Markdown
Contributor

Summary

  • Notebooks: all 18 example notebooks are re-run, so their outputs include the packed-bed fix (Make the storage take exactly the flux law's heat at its wall (F-056) #51, F-056). This was the follow-up to that PR.
    • The default Nallusamy charge now stores 2.318 kWh, up from 2.303 kWh (packed_bed_models). The design study's packed bed delivers 1.27 kWh, down from 1.28 kWh, still 52 % of what it stores.
    • The finned-tube outputs move by at most 0.0004 K.
    • Everything else that differs is solve times, round-off in the conservation errors, or object addresses.
  • Prose: the packed-bed notebook now gives the fit to the measurements at each height. The fluid temperatures are within 2 K RMS: 0.98, 0.68, 1.41 and 1.85 K at a quarter, half, three quarters and the full height. It used to say "1 K to 2 K", which understated the fit after F-056.
  • Other claims checked: the other numbers quoted in the notebook text still hold. The lumped capsule is fully charged after about 105 min, while the resolved bed is not fully charged after 160 min. Jeffreson's correction follows the resolved outlet at first, then charges too slowly. The figures quoted in the advection, validation and design-study notebooks are unchanged.
  • compat/coupled.py: the docstring now names [Bug]: resolve_coupled_variables=True breaks for boundary conditions and modifies the input model pybamm-team/PyBaMM#6001 and its fix, PR #6002. That fix must land before PyBaMM can resolve coupled variables by default. The issue asking for that default is still not filed (see Track upstream PyBaMM fixes, towards an empty compat/ #39).

The new runs saved QuickPlot widget state into four notebooks that had none before. I removed it, so each notebook stores the same kind of metadata as before.

Checks

  • Every notebook ran without errors, one at a time.
  • ruff check and ruff format --check pass on the whole repository, notebooks included.

🤖 Generated with Claude Code

Re-run every example notebook, so the packed-bed outputs include the
wall-conductivity fix (F-056). The stored energy of the default charge
is now 2.32 kWh (was 2.30), and the design study's packed bed delivers
1.27 kWh (was 1.28). Give the fit to Nallusamy et al. (2007) at the
four heights in the packed-bed notebook (0.68 K to 1.85 K RMS).

compat/coupled.py now names pybamm-team/PyBaMM#6001 and its fix, PR
#6002, which must land before the default can change.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
@read-the-docs-community

read-the-docs-community Bot commented Nov 1, 2026 •

Copy link
Copy Markdown

@brosaplanella brosaplanella added the no changelog Skip entry in CHANGELOG label Nov 1, 2026
Add a "Mesh and solver tolerance" section to the packed-bed page: the
default 40 capsule points leave wiggles of a few hundredths of a kelvin
as surface cells melt, gone at 80 points, and the energy error is the
solver's. Explain in the packed-bed example the averaged outlet's
second dip after its peak, which does not change with the mesh.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
@brosaplanella
brosaplanella merged commit 6477e12 into main Nov 1, 2026
20 checks passed
@brosaplanella
brosaplanella deleted the compat-coupled-tag-rerun-notebooks branch November 1, 2026 14:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

no changelog Skip entry in CHANGELOG

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant