feat: WindIO material overrides (#133) and per-mode modal mass/stiffness (#134)#135
Conversation
…ess (#134) #133: Tower.from_windio and from_windio_with_monopile gain optional E, rho, nu and outfitting_factor keywords. Each defaults to None (use the ontology value); a number overrides it so a material / outfitting sensitivity sweep runs straight off a WindIO file without editing the yaml. On the combined monopile+tower path the override applies to both segments. Existing callers are unaffected. #134: a solve attaches per-mode generalised_mass (kg) and generalised_stiffness (N/m) to ModalResult, normalised to unit tip lateral displacement, so a mode's modal mass and stiffness can be read off and lined up against another tool's modal report. Re-dimensionalised from the FEM generalised form phi^T gm phi via ref_mr = RM*radius over the squared tip lateral displacement DOF; stiffness = omega^2 * mass. Rigid-body platform modes (tip barely moves) yield NaN.
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 929be1375e
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".
| if tip2 <= 0.0 or not np.isfinite(gm_form) or gm_form <= 0.0: | ||
| continue | ||
| m = ref_mr * gm_form / tip2 |
There was a problem hiding this comment.
For modes whose tip has only numerical lateral motion, such as a floating heave/yaw or torsion-dominated mode, tip2 can be tiny but still positive, so this exact-zero check skips the intended undefined path and then divides by that tiny value. That can report enormous or infinite generalized_mass / generalized_stiffness instead of NaN; use a scale-aware tolerance for tip2 before normalizing.
Useful? React with 👍 / 👎.
The generalised-mass normalisation divided by the squared tip lateral displacement with only an exact-zero guard. A torsion / heave / yaw -dominated mode has a tiny but non-zero tip, which slipped through and produced a huge or infinite modal mass instead of NaN. Gate on the tip motion being a real fraction (>1e-8) of the mode's largest DOF amplitude, so only genuine lateral modes are normalised and the rest report NaN.
|
@codex review |
|
Codex Review: Didn't find any major issues. Can't wait for the next one! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
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". |
1 participant
Ships the two features that came out of the recent tower-comparison threads.
#133 — material / outfitting overrides on the WindIO paths
Tower.from_windioandTower.from_windio_with_monopileaccept optionalE,rho,nuandoutfitting_factorkeywords. Each defaults toNone(use the ontology value); pass a number to override it, so a sensitivity sweep runs straight off a WindIO file without editing the yaml. On the combined path the override applies to both the monopile and the tower segment. Existing callers are unaffected.#134 — per-mode generalised mass and stiffness on the result
A solve attaches
generalized_mass(kg) andgeneralized_stiffness(N/m) arrays toModalResult, one entry per mode, normalised to unit tip lateral displacement, so a mode's modal mass and stiffness can be read off and compared against another tool's modal output. Re-dimensionalised from the FEM generalised formphi^T gm phiviaref_mr = RM*radiusover the squared tip lateral-displacement DOF;stiffness = (2*pi*f)^2 * massby construction. A rigid-body platform mode (tip barely moves) yieldsNaN.Notes
ModalResultis preserved (new fields appended at the end,compare=False).tests/test_material_override_and_modal_mass.pycovers the override scalings (E, rho, outfitting, monopile+tower, None-is-ontology) and the modal mass/stiffness (populated, andsqrt(K/M)/2pireproduces each frequency).