Skip to content

feat(outputs): equilibrium tension recovery, continuous touchdown, anchor uplift channels - #21

Merged
SMI-Lab-Inha merged 5 commits into
devfrom
feat/output-recovery
Oct 16, 2026
Merged

SMI-Lab-Inha merged 5 commits into
devfrom
feat/output-recovery

Conversation

@SMI-Lab-Inha

Copy link
Copy Markdown
Owner

Summary

Finite-EI line outputs that engineers read for design, node tension, curvature, bend moment and touchdown, were wrong or jumpy on practical meshes. They are now recovered from nodal equilibrium. The PR also adds the anchor uplift channels used for drag-anchor checks.

Equilibrium-recovered node values (default)

  • Old definition: node tension used the element-mean axial force EA(|r'| − 1). That value oscillates between elements wherever the line bends over an element, and exceeds the tension by EI κ² on any mesh. On the 87-element lazy wave of the IEA-15 mixed example, the interior node tension dropped to −173 N against a true 1.2 kN.
  • New definition: each element's end forces (with its own distributed loads, seabed contact including bathymetry crease rounding, and inertia) give:
    • the effective tension, i.e. the axial component of the section force;
    • the bend moment;
    • the curvature, as that moment over EI.
  • Accuracy: the largest interior tension error against the converged equilibrium falls from 1552 N to 60 N on the example mesh, and from 516 N to 0.3 N on a 1 m mesh.
  • Unchanged: line-end values, positions and the material-strain checks.
  • Held ends on the seabed: a held end laid on the seabed carries the floor reaction of its half element. A laid anchor reported a spurious 903 N·m before; it now reports 0.02 N·m.
  • Low EI: a section with EI below |w| l0² /12 × 1 m reports the exact geometric curvature, because BendMom/EI would amplify the recovery error there.
  • Opting out: OutputRecovery element restores the 0.1.1 definitions. The exact geometric curvature stays available as CD_HermiteCable_Dyn_Curvature_Exact and CD_HFMF_Curvature_Exact.
  • Cost: the recovery is cached per element and keyed on that element's own 12 DOFs, the time and the load configuration. The Ten/Curv/BendMom channels of one output row share one evaluation, and per-node loops (static profile, C API line values) cost O(n). Range graphs and .t.out files written every step cost about 4 % more run time; plain channels cost under 2 %.

Continuous touchdown point

The touchdown point is located inside the element by the parabola through the last grounded node and the next two. It now moves continuously instead of in one-element steps: under a surge, the largest step between output rows falls from 4.64 m to 0.08 m with 5 m elements. The Fortran and Python locators agree.

Anchor uplift channels

  • AnchFz<L> / AnchFH<L>: the vertical (positive up) and horizontal components of the AnchTen<L> end force. They equal Point<P>Fz/Point<P>FH for an anchor that holds one line.
  • TDP<L>Grd: the grounded length, continuous; TDP<L>s + TDP<L>Grd is the line length.
  • Routes: static, EI = 0, finite-EI, mixed and multibody, the C API, and coupled OpenFAST/FAST.Farm. The Python deck reader and builder accept the names.
  • Docs: the manual names TDP<L>Grd (or AnchIncl) as the no-uplift check. On an EI = 0 line, AnchFz holds the anchor node's lumped weight, so its sign change lags lift-off.

Records

  • T-6 static-profile curvature (VALIDATION.md): 0.056 % → 0.049 %.
  • One-hour coupled full-turbine record: its curvatures cannot be re-run from the repository, so they are marked as the geometric definition.
  • Hs = 11 power-cable record and the OpenFAST tutorial: these give the recovered static values, Curv4N38 0.0902 1/m and BendMom4N38 1795 N·m.
  • Not affected: EI = 0 outputs, and every line-end output.

Tests

  • tests/test_anchor_uplift.py: the uplift channels on every route, a laid-anchor bend moment bounded from the deck's own properties, and the AnchFz lag.
  • Quadratic touchdown locator regressions (Fortran and Python).
  • The recovered tension, curvature and moment against converged equilibria.
  • Pinned hang-off and kink checks now read the geometric curvature.
  • The OpenFAST CompMooring = 5 module builds, and a 10 s coupled run terminates normally.
  • Full suite, Windows, gfortran, Release, slow tier included: 264/264. Python: 1999 passed. pre-commit run --all-files and the Sphinx build with -W pass.

Checklist

  • pre-commit run --all-files passes
  • ctest --test-dir build --output-on-failure passes
  • python -m pytest tests python/tests passes (if Python or validation records changed)
  • VALIDATION.md updated if a validation result changed
  • CHANGELOG.md updated under [Unreleased] (heading added if missing) for user-visible changes
  • Documentation in doc/ updated for changed inputs, options, or outputs
  • Conventional Commit title (feat:, fix:, docs:, ...)
  • Commits are signed off (git commit -s) under the
    Developer Certificate of Origin

…plift channels

Finite-EI node tension, curvature and bend moment are recovered from
nodal equilibrium. They had used the element-mean axial force
EA(|r'| - 1), which oscillates between elements wherever the line bends
over an element and exceeds the tension by EI kappa^2 on any mesh: on
the 87-element lazy wave of the IEA-15 mixed example the interior node
tension dropped to -173 N against a true 1.2 kN. Each element's end
forces, with its own distributed loads, contact and inertia, now give
the effective tension (the axial component of the section force, the
force that balances the loads), the bend moment and the curvature as
that moment over EI. The largest interior tension error against the
converged equilibrium falls from 1552 N to 60 N on the example mesh and
from 516 N to 0.3 N on a 1 m mesh. Line-end values, positions and the
material-strain checks are unchanged. OutputRecovery element restores
the previous definitions; the exact geometric curvature stays available
as CD_HermiteCable_Dyn_Curvature_Exact and CD_HFMF_Curvature_Exact.

The touchdown point is located inside the element by the parabola
through the last grounded node and the next two, so it moves
continuously instead of in steps of one element: under a surge the
largest step between output rows falls from 4.64 m to 0.08 m with 5 m
elements.

Adds the anchor uplift channels AnchFz<L> and AnchFH<L> (vertical and
horizontal components of the AnchTen<L> end force) and TDP<L>Grd (the
grounded length) for the drag-anchor check, on every route.

Range graphs and .t.out files written every step cost about 4 % more
run time, plain channels under 2 %. The T-6 record of VALIDATION.md,
which reads the static-profile curvature, moves from 0.056 % to
0.049 %.

Signed-off-by: Jae Hoon Seo <jaehoon.seo@inha.ac.kr>
…y cache

- The recovered end moment of a held end laid on the seabed missed the
  floor reaction of the end node's half element, so a laid anchor
  reported a spurious bend moment (903 N m at the cable anchor of the
  IEA-15 mixed example, 205 N m at the next node). The end element now
  carries that reaction: 0.018 and 0.10 N m. Test in
  tests/test_anchor_uplift.py.
- TDP<L>Grd was reset to 0 while only the end node was grounded and then
  jumped by a full chord; it is now the deformed chord length from the
  grounded end to the touchdown point throughout, and TDP<L>s +
  TDP<L>Grd is the line length.
- The node recovery is cached per element on the state (t, q, v, a and
  the load configuration) and reused across the Ten, Curv and BendMom
  channels of an output row (150 channels: 0.61 s -> 0.55 s on a 64
  element cable); outputs are bit-identical. Channel evaluation takes
  the model INTENT(INOUT) (CD_AGG_EvalChannel and related).
- A section with EI below max(|w| l0^2)/12 times 1 m reports the exact
  geometric curvature: the recovered moment carries an error
  independent of EI that Curv = BendMom/EI amplifies (6.9 1/m against
  0.12 1/m at EI = 100 N m^2).
- Regression tests pin the quadratic touchdown locator (Fortran and
  Python). Docs state that end values are the end reactions while
  interior values include inertia, that the recovery balance in
  dynamics is the one at t_{n+1}, and that AnchFz is exactly 0 at a
  grounded finite-EI anchor.

Signed-off-by: Jae Hoon Seo <jaehoon.seo@inha.ac.kr>
…covery, O(n) node queries

The OpenFAST CompMooring = 5 module no longer compiled: CD_AGG_ChannelHeader
had become INTENT(INOUT) while CD_InitVars passes it part of an INTENT(IN)
instance. CD_AGG_ChannelHeader and CD_AGG_GetInitLine write nothing and are
INTENT(IN) again, as are the standalone driver's mixed-route header and
initialisation printers; only the channel evaluator and the static profile,
which reach the recovery cache, stay INTENT(INOUT), and every glue call of
those passes the module-level instance or a local aggregate.

The seabed force spread over the elements by the node recovery now matches
the residual in a rounded bathymetry crease: it takes the crease curvature
and adds the elastic increment of the penalty potential, as assemble_contact
and the end force do, so interior tension, curvature and bend moment over a
crease use the force the solve applied.

A node query compared the whole line state with the cached one, so a loop
over the nodes of a line (static profile, C API line values, a row of node
channels) cost O(n^2). The cache is now keyed per element by its own 12
position, velocity and acceleration DOFs (the only state its end residual
depends on) besides the time and the load configuration, so a query
compares two elements and needs no list of the places that change the
state.

Tests: the pinned hang-off check of test_pinned_touchdown_stability and
the kink checks of test_deck_hermite_cable read the geometric curvature
(CD_HFMF_Curvature_Exact): the recovered curvature is zero at a pinned end
by construction and is not a shape measure. test_anchor_uplift reads the
anchor node and element length of the mixed example from its deck (the
example is meshed with 248 elements since the rebase) and allows the bend
of the laid run sinking into the penalty seabed beside the anchor.

Documentation: the anchor uplift section no longer claims that AnchFz and
TDP<L>Grd agree. AnchFz is the anchor reaction; on an EI = 0 line it holds
the anchor node's lumped weight, so its sign change lags the lift-off of
the anchor segment, and TDP<L>Grd (or AnchIncl) is the no-uplift check. The
curvatures of the one-hour coupled full-turbine record, which cannot be
re-run from the repository, are marked as the geometric definition; the
Hs = 11 power-cable record and the OpenFAST tutorial give the recovered
static values (Curv4N38 0.0902 1/m, BendMom4N38 1795 N m); the CHANGELOG
qualifies the numbers measured on the example's former 87-element mesh and
no longer says every other record is unchanged. Stale comments on the
finite-EI segment tension and the TDP channel list are corrected.

Signed-off-by: Jae Hoon Seo <jaehoon.seo@inha.ac.kr>
- The static-profile Tension column is described by its channel
  definition only; "the segment tension" was wrong for a finite-EI line,
  whose profile tension is the recovered nodal F.t.
- test_anchor_uplift reads the bare type's diameter, mass per length
  and EI and the kBot option from the mixed example instead of
  hard-coding them, so the laid-anchor moment bound follows the deck.
- The recovery-cache comment lists axial damping among the loads that
  advance the epoch, as Set_Axial_Damping does.

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

Copy link
Copy Markdown
Owner Author

@codex review

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 16, 2026 •

Copy link
Copy Markdown

Codex Review Summary

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

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-16T22:00:59.673784Z f9f2fd7 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 Oct 16, 2026 •

Copy link
Copy Markdown

Documentation build overview

📚 CableDyn | 🛠️ Build #35201747 | 📁 Comparing f9f2fd7 against latest (f27c5e3)

  🔍 Preview build  

28 files changed · ± 28 modified

± Modified

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 939600106d

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".

Comment thread python/cabledyn/deck_file.py Resolved
The native aggregate builds every finite-EI line of a coupled deck
(OpenFAST, FAST.Farm, the C API, the standalone mixed deck) held at a
Fixed End B and refuses any other. The Python validator accepted such a
line, for example a finite-EI cable shared by two turbines, together
with its AnchFz and AnchFH channels, so a deck it reported valid failed
at initialisation. It now refuses the line with the native message. An
EI = 0 line shared by two turbines keeps its end-force channels.

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

Copy link
Copy Markdown
Owner Author

@codex review

@chatgpt-codex-connector

Copy link
Copy Markdown

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

Reviewed commit: f9f2fd7a57

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 71158d5 into dev Oct 16, 2026
15 checks passed
@SMI-Lab-Inha
SMI-Lab-Inha deleted the feat/output-recovery branch October 17, 2026 09:04
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