SMI-Lab-Inha · Nov 9, 2026 · Nov 8, 2026 · Nov 8, 2026 · Nov 8, 2026 · Nov 8, 2026
diff --git a/ARCHITECTURE.md b/ARCHITECTURE.md
@@ -322,14 +322,16 @@ analytic Jacobians.
 about its static equilibrium, `K φ = ω² M φ`, over the free DOFs, with `K` the
 static tangent (element stiffness and linearised seabed contact) and `M` the
 consistent mass plus Morison added mass. `K` and `M` are assembled in symmetric
-band storage; LAPACK `DSBGVX` gives the eigenvalues and inverse iteration on the
-banded pencil gives each mode shape, so the cost grows with line length rather
-than with its square. Dense `DSYGV` is kept as the test reference. Both line paths
-are supported (`nModes`). A finite-EI line's modes (`CD_Modal_Hermite_Line`) take
-its end connections and attachments from the dynamic model, and a line restrained
-in torsion adds the twist angle of every node: its lumped polar inertia and the
-second variation of the nodal-twist energy (`CD_HermiteCable_Dyn_Twist_Hessian`)
-in a band of 7 DOFs per node.
+band storage. A shift-invert block Lanczos iteration on the banded Cholesky factor
+of `K` gives the eigenvalues, certified by a Sturm count of `K - σM` (LAPACK
+`DSBGVX` takes short lines, an indefinite or singular `K`, and an uncertified run),
+and inverse iteration on the banded pencil gives each mode shape, so the cost
+grows with line length rather than with its square. Dense `DSYGV` is kept as the
+test reference. Both line paths are supported (`nModes`). A finite-EI line's
+modes (`CD_Modal_Hermite_Line`) take its end connections and attachments from the
+dynamic model, and a line restrained in torsion adds the twist angle of every
+node: its lumped polar inertia and the second variation of the nodal-twist energy
+(`CD_HermiteCable_Dyn_Twist_Hessian`) in a band of 7 DOFs per node.
 
 ### Outputs
 

diff --git a/CHANGELOG.md b/CHANGELOG.md
diff --git a/CITATION.cff b/CITATION.cff
@@ -23,8 +23,8 @@ authors:
 repository-code: "https://github.com/SMI-Lab-Inha/CableDyn"
 url: "https://github.com/SMI-Lab-Inha/CableDyn"
 license: Apache-2.0
-version: 0.4.1
-date-released: 2026-11-08
+version: 0.4.2
+date-released: 2026-11-09
 keywords:
   - mooring dynamics
   - cable dynamics

diff --git a/CMakeLists.txt b/CMakeLists.txt
diff --git a/README.md b/README.md
@@ -9,38 +9,76 @@
 [![License: Apache-2.0](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
 [![DOI](https://img.shields.io/badge/DOI-10.1016%2Fj.oceaneng.2026.128332-blue.svg)](https://doi.org/10.1016/j.oceaneng.2026.128332)
 
-CableDyn is an open-source, geometrically exact finite-element solver for the
-static and dynamic analysis of offshore mooring lines and dynamic power cables.
-Its strain and curvature are the exact large-deformation measures, never
-linearised, so large displacements and rotations, slack and snap, and a line
-lifting off the seabed lie inside the formulation. It is written in
-Fortran 2018 and runs as a standalone program, as a mooring module of OpenFAST,
-maintained by NLR (National Laboratory of the Rockies, formerly NREL)
-(`CompMooring = 5`), through a C interface, or from Python. It reads input decks
-in the MoorDyn v2 format and solves them with two line formulations:
-
-- **Tension-only lines** (`EI = 0`): chains, wire ropes, and synthetic-fibre
-  moorings.
-- **Finite-bending lines** (finite-EI): cubic-Hermite elements for dynamic power
-  cables, including lazy-wave configurations, with continuous centreline curvature.
-
-The current stable release is `v0.4.1`. Interfaces are
-versioned and may change before `1.0.0`.
-
-## Capabilities
-
-| Area | Features |
-|---|---|
-| Statics | Catenary seeding, Newton solution, load continuation, mesh sequencing, adaptive meshing |
-| Dynamics | Implicit generalised-alpha integration, prescribed support motion, consistent mass, automatic step subdivision, torsion with a twist angle and rotary inertia at every node, cross-flow vortex-induced vibration |
-| Environment | Buoyancy, Morison drag, added mass, Froude–Krylov loading, regular and irregular waves, current profiles, seabed contact with friction |
-| Materials | Linear axial stiffness, Kelvin–Voigt damping, viscoelastic (series-Kelvin; MoorDyn `ElasticMod` 2/3), Syrope polyester model, torsional stiffness with torque–tension coupling, tapered bend stiffeners |
-| Topology | Fixed, coupled, free, and connected points; clump weights; buoyancy modules along cables; rigid bodies and rods; end rotational stiffness; cables between two moving ends; active line-length control; line failure |
-| Outputs | Effective tension, position, velocity, acceleration, curvature (also on the section's own axes), bending moment, torque and twist, end angles, contact response, natural modes, solver diagnostics |
-| Interfaces | Standalone driver, OpenFAST and FAST.Farm (with linearisation of the line dynamics), C ABI, Python package |
-
-Section and keyword names follow MoorDyn v2; unsupported combinations stop with
-a named error rather than running approximately.
+CableDyn is an open-source, geometrically exact finite-element solver for the static and
+dynamic analysis of mooring lines and dynamic power cables. It resolves what a power cable is
+designed against: the curvature along the line, under a bend stiffener, through a lazy wave and
+across a moving touchdown, together with tension, torque and twist. Strain and curvature are the
+exact large-deformation measures, never linearised, so slack and snap, large rotations and a line
+lifting off the seabed lie inside the formulation.
+
+It runs as a standalone program, as a mooring module of OpenFAST and FAST.Farm
+(`CompMooring = 5`), through a C interface, or from Python.
+The current stable release is `v0.4.2`; interfaces are versioned and may change before `1.0.0`.
+
+## What CableDyn does
+
+- **Finite-element cables with real bending stiffness.** Cubic-Hermite elements carry position
+  and tangent at every node, so the slope is continuous along the line, and the deflection of a
+  beam under its weight converges at fourth order with the mesh (measured slope 3.99).
+  Tension-only `EI = 0` elements serve chains, wire and fibre ropes in the same model.
+- **An implicit solver.** Generalised-alpha integration with consistent mass and automatic step
+  subdivision: a snap load stays within 1.5 % of a 0.5 ms reference at a 20 ms step, where an
+  explicit lumped-mass step must stay near 1 ms.
+- **Bend stiffeners and section curvature.** Tapered stiffener sleeves with their own mass,
+  buoyancy and drag; the curvature of the section, its two components on the section's own axes
+  (`CurvX`, `CurvY`) and the moments of the cable and of the sleeve, for fatigue.
+- **Torsion.** A twist angle and polar inertia at every node, torque–tension coupling for
+  armoured cables, the buckling onset of a twisted line, and end restraints that are rigid,
+  elastic or free in twist.
+- **Statics that start from the geometry alone.** No seed files: a catenary start, continuation in
+  bending stiffness and buoyancy, mesh sequencing and refinement to the bending length solve hard
+  lazy waves, long grounded runs, bathymetry and friction held as laid; a deck that has no
+  equilibrium stops with a named reason.
+- **Coupled analysis.** OpenFAST and FAST.Farm coupling with checkpoint restart at any step, the
+  host's waves and current at every node, active line control, line failure, and linearisation
+  that carries the line dynamics as states.
+- **More analyses.** Natural modes and shapes, cross-flow vortex-induced vibration of chains and
+  cables, viscoelastic and Syrope fibre ropes, buoyancy modules, clumps, rigid bodies and rods,
+  range graphs and seabed contact channels.
+- **In-process hosts.** The C API and Python step a deck from a CFD code or a script, with the
+  host's fluid field at every node, saved and restored states and whole-line reads in one call.
+
+## Evidence
+
+Every capability above is tested, and the comparisons are published with their data and scripts
+in the [validation record](https://cabledyn.readthedocs.io/en/latest/validation.html). A
+selection:
+
+| Case | Reference | Result |
+|---|---|---|
+| Large-deflection elastica | Bisshopp–Drucker closed form | error 1.9e-6 |
+| Lazy-wave cables at 80, 200 and 800 m, static | OrcaFlex 11.6d | peak curvature within 0.08, 0.26 and 4.8 % |
+| Installed 800 m lazy-wave cable | OrcaFlex 11.6d | peak curvature 0.06 %, hang-off tension 0.15 % |
+| Lazy wave in an irregular sea | OrcaFlex 11.6d, same 80-component sea | curvature 0.21 %, tension 0.20 % |
+| Moving touchdown under heave | OrcaFlex 11.6d | curvature within 0.6 %, touchdown point within 0.08 m |
+| Tapered bend stiffener at a clamped hang-off | OrcaFlex 11.6d | curvature 0.45 % of its root value, moments 0.01 % |
+| Lazy-wave cable twisted five turns | OrcaFlex 11.6d | torque 0.019 %, peak curvature 0.049 % |
+| One-hour VolturnUS-S chain mooring | OrcaFlex 11.6d | fairlead tension mean 0.27 %, standard deviation 0.26 % |
+| Holcombe lazy-wave model test | measured geometry | position error 1.1–1.9 % of the suspended length |
+| Bergdahl dynamic chain test, 30 conditions | measured end force | correlation 0.993 |
+| IEA 15 MW VolturnUS-S in OpenFAST with a power cable | OpenFAST with MoorDyn | platform motions within 0.6 % |
+
+Run times on one processor thread, same host and protocol, with the timing scripts in
+[`validation/performance`](https://cabledyn.readthedocs.io/en/latest/validation.html):
+
+| Case | CableDyn 0.4.2 | OrcaFlex 11.6d |
+|---|---|---|
+| 80 / 200 / 800 m lazy-wave spans, 60 s of heave | 1.3 / 2.4 / 5.0 s | 2.7 / 5.3 / 27.8 s |
+| One-hour VolturnUS-S chain mooring | 4.3 s | 19.7 s |
+
+The method and the lazy-wave comparisons are peer-reviewed
+([Ocean Engineering 368, 128332, 2026](https://doi.org/10.1016/j.oceaneng.2026.128332)).
+Unsupported combinations stop with a named error rather than running approximately.
 
 ## Quick start
 
@@ -51,8 +89,8 @@ a named error rather than running approximately.
    |---|---|
    | `CableDyn_driver.exe` | Standalone solver for Windows x64 (AVX2 processor) |
    | `openfast.exe` | OpenFAST with CableDyn available as `CompMooring = 5` |
-   | `cabledyn-0.4.1-py3-none-any.whl` | Python package (pre- and post-processing, batch runs) |
-   | `cabledyn-0.4.1.tar.gz` | Python source distribution |
+   | `cabledyn-0.4.2-py3-none-any.whl` | Python package (pre- and post-processing, batch runs) |
+   | `cabledyn-0.4.2.tar.gz` | Python source distribution |
    | `SHA256SUMS.txt` | Checksums for the assets above |
 
    Both executables are statically linked, need no installer, compiler, or runtime
@@ -90,7 +128,7 @@ a named error rather than running approximately.
    cabledyn-run examples\wd0050_chain.dat wd0050
    ```
 
-   The pure Python wheel on the release page (`cabledyn-0.4.1-py3-none-any.whl`)
+   The pure Python wheel on the release page (`cabledyn-0.4.2-py3-none-any.whl`)
    carries no driver: `cabledyn-run` then needs `--executable .\CableDyn_driver.exe`,
    the `CABLEDYN_DRIVER` environment variable, or the driver on `PATH`. A source
    build names the driver `cabledyn` (see [Build from source](#build-from-source));
@@ -101,6 +139,11 @@ a CableDyn deck; ready-to-run cases are in
 [`examples/openfast`](https://cabledyn.readthedocs.io/en/latest/tutorial_openfast.html). Every shipped deck is listed
 in [`examples/`](https://cabledyn.readthedocs.io/en/latest/examples.html).
 
+Coming from MoorDyn: CableDyn reads decks written in the MoorDyn v2 format (its section and
+keyword names), so an existing mooring model runs as it is and can then take finite-EI sections,
+stiffeners and the other features above; the
+[migration guide](https://cabledyn.readthedocs.io/en/latest/migrating.html) lists the differences.
+
 ## Documentation
 
 The manual at <https://cabledyn.readthedocs.io> covers installation, tutorials,
@@ -157,7 +200,7 @@ If you use CableDyn in academic work, please cite:
 > offshore wind. *Ocean Engineering* 368 (Part 2), 128332, 2026.
 > <https://doi.org/10.1016/j.oceaneng.2026.128332>
 
-To identify the software version, also cite the release (CableDyn 0.4.1).
+To identify the software version, also cite the release (CableDyn 0.4.2).
 The [citation file](https://cabledyn.readthedocs.io/en/latest/citing.html) contains both entries (GitHub **Cite this
 repository**); BibTeX is on the manual's
 [How to cite](https://cabledyn.readthedocs.io/en/latest/citing.html) page.

diff --git a/RELEASE.md b/RELEASE.md
@@ -104,7 +104,7 @@ The build fails if either executable imports a non-system DLL or lacks the UTF-8
 active-code-page manifest (`release/check_static_windows_binary.ps1`), or if the
 standalone and `CompMooring = 5` smoke cases fail with a system-only `PATH` or from
 a folder named in mixed scripts. It also fails unless `openfast.exe -v` names the
-executable `OpenFAST-v5.0.0-CableDyn-0.4.1`: the OpenFAST base and the CableDyn version,
+executable `OpenFAST-v5.0.0-CableDyn-0.4.2`: the OpenFAST base and the CableDyn version,
 never a `-dirty` describe of the patched tree.
 
 The release workflow (`.github/workflows/release-windows-static.yml`) produces
@@ -140,7 +140,7 @@ python release/smoke_test_wheel.py examples/dynamic_chain_current.dat --expect-b
 Tag only the commit that passed every check:
 
 ```powershell
-$ReleaseVersion = "0.4.0"
+$ReleaseVersion = "MAJOR.MINOR.PATCH"
 git tag -s "v$ReleaseVersion" -m "CableDyn v$ReleaseVersion"
 git push origin "v$ReleaseVersion"
 ```