Skip to content

fix(statics): refine a graded mesh element by element to the bending length - #125

Merged
SMI-Lab-Inha merged 1 commit into
devfrom
fix/graded-mesh-statics
Nov 8, 2026
Merged

SMI-Lab-Inha merged 1 commit into
devfrom
fix/graded-mesh-statics

Conversation

@SMI-Lab-Inha

Copy link
Copy Markdown
Owner

Six frictional lazy-wave decks with a strongly graded mesh failed their static solve in a 45° or 90° cross current. The mesh had 4 m elements at the sag bend next to a 0.9–1 m buoyant section, and 10 m elements at touchdown. The solver reported "no static equilibrium", yet the same decks solve on 2 m elements.

Cause

The refinement strategy. Half the bending length, (EI/|w|)^(1/3), is 2.5 m there, so both long sections break the resolution rule.

  • On the deck mesh: both static routes fail. The still-water fallback reads a −2.8 kN element-mean "compression" in the 4 m elements beside the short ones, where the effective tension is about +1 kN. This is the long elements' axial oscillation.
  • Uniform refinement does not help: the old refined_restart refined every element by the worst element's factor (8×), which keeps the 11:1 grading. That mesh stalls too.
  • The message was false: an equilibrium exists.

Fix

  • Element-wise refinement, rule_restart. It runs when uniform_restart (the former refined_restart, unchanged) fails. It repeats the route on the deck mesh, splitting each element into the power of two (at most 16) that brings it to half the bending length and leaving the others alone.
    • It is skipped when that mesh equals the uniform one or exceeds 4096 elements.
    • Deck and attachment nodes are preserved.
    • It serves the continuation failure, the polish failure and the stability refine.
  • The message. It now reports a compression read on elements longer than half the bending length, and says to refine those elements (NumSegs) to half the bending length or less. It no longer says "no static equilibrium".

No tolerance or deck-syntax change.

Evidence

  • Robustness sweep, 351 decks:
    • the frictional family goes from 111 OK / 6 FAIL to 117/117;
    • the other two families stay 117/117;
    • no fallback notes anywhere.
  • Grading sweep, 32 decks: buoyant section refined 1–8×, ratios 1.1–23, four currents, 10 and 25 modules. Before: 6 FAIL. After: 32/32.
  • Accuracy:
    • against the same decks on 2 m elements: fairlead tension within 2e-4, every end-force channel within 1.2e-3;
    • against fourfold-refined references: within 1.1e-3 to 1.3e-3.
  • Unchanged results: the 343 decks that converged before are byte-identical, except two that now solve on the rule mesh and land closer to the 4× reference (2.76e-3 → 1.26e-3 and 2.69e-3 → 3.80e-4).
  • Examples: all 54 are byte-identical.

Tests

  • New graded_mesh_statics: the six failing decks, grading controls and two 2 m-mesh comparisons. It fails on all six before this change (exit 2) and passes after it, in 37 s Release and 194 s Debug.
  • Full Release CTest: 367/367 on the default and Prescott kernels.

@SMI-Lab-Inha

Copy link
Copy Markdown
Owner Author

@codex review

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 8, 2026 •

Copy link
Copy Markdown
Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-08T17:50:16.812903Z eaea3c9 Manual request
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@read-the-docs-community

read-the-docs-community Bot commented Nov 8, 2026 •

Copy link
Copy Markdown

Documentation build overview

📚 CableDyn | 🛠️ Build #35737044 | 📁 Comparing eaea3c9 against latest (e8f88f5)

  🔍 Preview build  

11 files changed · ± 11 modified

± Modified

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. 👍

Reviewed commit: 8c958b2e92

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

…length

A finite-EI static route that fails on a deck mesh with elements longer than
half the bending length (EI/|w|)^(1/3) is repeated on a refined mesh. That mesh
was the whole line refined uniformly by the factor of its worst element, which
keeps a graded deck's ratios and refines its short elements with the long
ones. A lazy wave meshed 4 m at its sag bend, 0.9 m over its buoyant section
and 10 m at its touchdown (half its bending length is 2.5 m) failed in an
oblique or cross current: the routes stopped on the deck mesh, the 8x refined
mesh stalled at the start of its stiffness continuation, and the deck reported
"the line is already in axial compression without the current", although the
same layout solves on 2 m elements. That compression (-2.8 kN element mean
where the line carries about +1 kN) is the long elements' axial oscillation
beside the short ones.

When the uniform refinement fails too, the route is now repeated on the deck
mesh with each element split only into the power of two (at most 16) that
brings it to half the bending length. The six failing decks of a sweep over
current speed, heading, module count and refinement now solve without a
fallback, within 1.2e-3 of their fairlead tension of the same decks on 2 m
elements; a grading sweep (buoyant section refined 1, 2, 4 and 8 times in four
currents) converges throughout. Decks that solved before are unchanged, except
two that had failed their uniform refinement and solved on the deck mesh by a
later route: they now solve on the mesh refined to the bending length, closer
to a fourfold refined reference (1.3e-3 and 3.8e-4 against 2.7e-3).

The still-water compression diagnosis of a failed solve in a current now names
a compression read on elements longer than half the bending length as possibly
the mesh's, with the remedy, instead of declaring that no equilibrium exists.

New test graded_mesh_statics: the six decks, two grading ratios that solved
before, and two comparisons with the 2 m mesh.

Signed-off-by: Jae Hoon Seo <jaehoon.seo@inha.ac.kr>
@SMI-Lab-Inha

Copy link
Copy Markdown
Owner Author

@codex review

@SMI-Lab-Inha
SMI-Lab-Inha force-pushed the fix/graded-mesh-statics branch from 8c958b2 to eaea3c9 Compare November 8, 2026 17:48
@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Nice work!

Reviewed commit: eaea3c97de

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@SMI-Lab-Inha
SMI-Lab-Inha merged commit 3bc2450 into dev Nov 8, 2026
22 checks passed
@SMI-Lab-Inha
SMI-Lab-Inha deleted the fix/graded-mesh-statics branch November 9, 2026 08:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant