Skip to content

feat(torsion): torsional inertia and damping, a twist angle at every node - #69

Merged
SMI-Lab-Inha merged 4 commits into
devfrom
feat/nodal-twist
Oct 30, 2026
Merged

SMI-Lab-Inha merged 4 commits into
devfrom
feat/nodal-twist

Conversation

@SMI-Lab-Inha

@SMI-Lab-Inha SMI-Lab-Inha commented Oct 29, 2026 •

Copy link
Copy Markdown
Owner

In the dynamics a finite-EI line restrained in torsion now carries a twist
angle phi at every node, against reference directors that follow the node
tangents by the smallest rotation over each step (time-parallel transport),
so the axial angular velocity of a section is phi_dot. The statics stay
condensed and set the angles from the uniform static torque.

  • Element twist phi_{e+1} - phi_e + beta_e - h_e: beta_e advances by the
    solid angles of the quadrilateral of committed and present tangents
    (CD_HermiteTwist_Reference, local to the element's 12 DOFs, with gradient
    and Hessian), so the turns live in the angles and no 2 pi branch is
    tracked. A rigid end takes its frame's normal carried to the node as the
    node's director and holds the angle; an elastic end adds its spring.
  • Torque M_e = (GJ/l0)(tw + (K/GJ) dl) + (Bt/l0) d/dt(...), with the polar
    inertia Irn (LINE TYPES column, now used) lumped at the nodes and the new
    torsional damping Bt of the GJ column (GJ|K|Bt, negative: -zeta of an
    element's critical l0 sqrt(GJ Irn)).
    A rigid end's element twists through rho as its end frame turns; its
    damping rate adds d rho/dw . omega of the frames over the step.
    The support reaction of a rigid end adds the held end section's polar
    inertia, turned by the frame's angular acceleration and the held angle;
    the polar kinetic energy takes that section's spin with its frame.
  • The angles join the generalised-alpha step; the step factors one band of
    [r, m, phi] per node (half-width 13). Their rows converge on the nodal
    torque balance, scaled by the nodal loads as the force rows are. The
    committed reference advances with the step's own solid angles, so the
    committed state reproduces the converged iterate bit for bit; the twist
    state joins the snapshot, the recovering step and the OpenFAST mirror.
  • TorqN is the node's torque (mean of its elements); end force, end
    moment, energy, element twist and section curvature read the nodal state.
  • Under torque-tension coupling each element's coupling force (K/GJ) M_e
    takes its own torque: in the dynamic tensile_safety audit (the converged
    evaluation's), the element recovery and the node tensions.

Gates: the kernel against direct transports (1e-13) and FD;
nodal_twist_dynamic: the condensed torsion in the limit of no inertia
(<= 6e-7, clamped, bending-pinned, elastic ends), the continuum torsional
wave (0.07 %) and its zero driven-end torque at kL = pi/2 (0.13 %), the
damping decrement (6 %), a parent rolling the end frame at a steady rate
loading the line with (GJ theta + Bt Omega)/L (4e-13) and spinning it with
the continuum's kinetic energy (7e-4), energy band 4.9e-4, snapshot and
mirror restart bit-identical; a zero-allocation step gate. The old
quasi-static gates run at Irn 1e-9/1e-12; a non-stepping coupled write
twists the end element against the frozen interior. Full suite passes;
52 of 53 examples bit-identical, the torsion example within 7e-5 of the
torque.

…node

In the dynamics a finite-EI line restrained in torsion now carries a twist
angle phi at every node, against reference directors that follow the node
tangents by the smallest rotation over each step (time-parallel transport),
so the axial angular velocity of a section is phi_dot. The statics stay
condensed and set the angles from the uniform static torque.

- Element twist phi_{e+1} - phi_e + beta_e - h_e: beta_e advances by the
  solid angles of the quadrilateral of committed and present tangents
  (CD_HermiteTwist_Reference, local to the element's 12 DOFs, with gradient
  and Hessian), so the turns live in the angles and no 2 pi branch is
  tracked. A rigid end takes its frame's normal carried to the node as the
  node's director and holds the angle; an elastic end adds its spring.
- Torque M_e = (GJ/l0)(tw + (K/GJ) dl) + (Bt/l0) d/dt(...), with the polar
  inertia Irn (LINE TYPES column, now used) lumped at the nodes and the new
  torsional damping Bt of the GJ column (GJ|K|Bt, negative: -zeta of an
  element's critical l0 sqrt(GJ Irn)).
- The angles join the generalised-alpha step; the step factors one band of
  [r, m, phi] per node (half-width 13). Their rows converge on the nodal
  torque balance, scaled by the nodal loads as the force rows are. The
  committed reference advances with the step's own solid angles, so the
  committed state reproduces the converged iterate bit for bit; the twist
  state joins the snapshot, the recovering step and the OpenFAST mirror.
- Torq<L>N<J> is the node's torque (mean of its elements); end force, end
  moment, energy, element twist and section curvature read the nodal state.
- Under torque-tension coupling each element's coupling force (K/GJ) M_e
  takes its own torque: in the dynamic tensile_safety audit (the converged
  evaluation's), the element recovery and the node tensions.

Gates: the kernel against direct transports (1e-13) and FD;
nodal_twist_dynamic: the condensed torsion in the limit of no inertia
(<= 6e-7, clamped, bending-pinned, elastic ends), the continuum torsional
wave (0.07 %), the damping decrement (6 %), energy band 4.9e-4, snapshot
and mirror restart bit-identical; a zero-allocation step gate. The old
quasi-static gates run at Irn 1e-9/1e-12; a non-stepping coupled write
twists the end element against the frozen interior. Full suite passes;
52 of 53 examples bit-identical, the torsion example within 7e-5 of the
torque.

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 29, 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 2026-10-30T01:31:47.607878Z 8e334b9 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 29, 2026 •

Copy link
Copy Markdown

Documentation build overview

📚 CableDyn | 🛠️ Build #35513138 | 📁 Comparing 8e334b9 against latest (fd1936d)

  🔍 Preview build  

27 files changed · ± 27 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: e854b207e3

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 src/CableDyn_HermiteCableDynamic.f90 Resolved
An element at a rigid torsional end twists through rho as its end frame
turns (a parent rolling it, CD_HFMF_UpdateStates u_orientation), but the
torsional damping rate only differentiated the nodal angles and the
centreline: a parent rolling the end at a steady rate loaded the line
with GJ theta/L alone, without its Bt Omega/L.

The step now takes each end frame's angular velocity over the step,
log(R_{n+1} R_n^T)/dt (tw_frame_rates), and the end elements' damping
rate adds d rho/dw . omega (end_w of twist_reference). The committed
step's rates (tw_om) serve the evaluations of the committed state and
join the packed twist state, so the snapshot, the recovering step and
the checkpoint mirror carry them.

New gate in nodal_twist_dynamic: a parent rolling the coupled end's frame
at a steady 0.5 rad/s (Irn 1e-9, Bt 40) loads every node with
(GJ theta + Bt Omega)/L to 4e-13 (2.0 of the 12.0 N m is the damping;
the unfixed code read 10.0); the restart gate now runs with Bt and stays
bit-identical through the snapshot and the mirror.

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 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: 389ad6d6f3

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 src/CableDyn_HermiteCableDynamic.f90 Resolved
…ction

The end node of a line rigid in torsion carries half its element's polar
inertia, but its angle is held, so that inertia entered no equation: the
reaction moment on the support took the element's torque alone. A driven
end then loaded its body with too little rotational inertia (a single
element lost all of it).

The held section turns with its end frame and its held angle,
chi_dot = -+ drho/dw . omega + phi_dot (End 1, End 2); the reaction now
adds -I chi_ddot dchi/dw. The frames' angular acceleration at the
committed state, (omega_{n+1} - omega_n)/dt, is recorded at the commit
(tw_oal) and joins the packed twist state with the rates.

New gates in nodal_twist_dynamic: the clamped line driven at kL = pi/2,
where the continuum's torque at the driven end, GJ A k cot(kL), is zero:
the axial reaction there swings by 0.13 % (parent roll) and 0.06 %
(imposed twist) of GJ A k; without the section's inertia both swung by
its I Omega^2 A, 3.9 %.

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 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: 6c4810c8cb

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 src/CableDyn_HermiteCableDynamic.f90 Resolved
…etic energy

The polar kinetic energy summed the nodal rates alone, so a rigid end
driven by its parent frame, whose section spins with the frame while its
held angle stands still, carried none: the energy query and the coupled
energy diagnostics missed the inertia the support reaction now includes.

The held end section's spin is the reaction's chi_dot = -+ drho/dw . omega
+ phi_dot (End 1, End 2), with the committed step's frame rates.

New gate in nodal_twist_dynamic: the clamped line, its end frame rolled by
the parent at a steady rate, holds the continuum's polar kinetic energy
within 7e-4; without the held section's spin it fell 7.5 % short.

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. More of your lovely PRs please.

Reviewed commit: 8e334b93b3

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 66e3e83 into dev Oct 30, 2026
21 checks passed
@SMI-Lab-Inha
SMI-Lab-Inha deleted the feat/nodal-twist branch October 30, 2026 03:07
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