Skip to content

EXAMPLE_WORKFLOWS.md

Latest commit

 

History

History
89 lines (68 loc) · 3.65 KB

EXAMPLE_WORKFLOWS.md

File metadata and controls

89 lines (68 loc) · 3.65 KB
Raw

Example workflows

Pointers to runnable examples in the SEA-Stack tree: standalone libraries, the full simulation app, and packaged demos. For a separate CMake project that links against an installed SDK, see DOWNSTREAM_PROJECT.md.

Standalone hydrodynamics

Evaluate hydrodynamic forces without Chrono — see examples/standalone_hydro.

Standalone PTO and control

Run PTO and control models without Chrono or Eigen — see examples/standalone_controller.

Chrono-integrated simulation

Full time-domain multibody dynamics with run_seastack and YAML configuration. After a build or package install, start from the Quick start section of the main README; case YAML and assets live under data/demos/run_seastack/.

Demo-driven cases

Preconfigured models (RM3, 5SA, OSWEC, IEA sphere, F3OF, and others) with documented command lines: demo suite README.

Regular waves: constant-coefficient radiation

For a regular wave, radiation.method: constant_coefficients applies the BEM coefficients at the wave frequency directly:

F_rad = -A(w) xddot - B(w) xdot

A(w) and B(w) are the full coupled 6N x 6N matrices from the .h5 file (added_mass/all, radiation_damping/all). A(w) replaces the infinite-frequency added mass A_inf in the Chrono added-mass load, so added mass is never counted twice. B(w) is applied as the radiation force, in place of the RIRF convolution or state-space model. Example: rm3/regular_waves_cc.

hydrodynamics:
  waves:
    type: regular
    height: 2.0
    period: 7.853984            # must be a BEM grid period
  radiation:
    method: constant_coefficients
    frequency_tolerance: 1.0e-4  # optional; relative |w_k - w| / w_k

Use it when the BEM frequency grid is coarse (for example design-screening runs with a dozen periods). On such grids the RIRF is built from few frequencies and can misrepresent A(w) and B(w) even at the grid points. On dense grids both methods agree closely.

Rules (checked at startup; violations stop the run with an explicit error):

  • Regular waves only. Irregular, bimodal, spread, eta_file, still water and free decay are rejected: a single frequency cannot represent them.
  • The period must be on the BEM grid. No interpolation and no extrapolation. The match uses frequency_tolerance (default 1e-4, relative). On failure the error lists the nearest grid period and all available periods.
  • RIRF-only options are rejected: smoothing, taper, and radiation truncation_time.
  • The .h5 must contain added_mass/all and radiation_damping/all for every body (standard BEMIO output).

Startup log: the requested and matched period and frequency, the frequency difference, the A(w), B(w) and A_inf diagonals per body, and a line confirming that A(w) replaces A_inf.

Limitations:

  • Steady state only. The start-up transient is wrong: it decays at the rate set by B(w) at the wave frequency, not by the true radiation damping at the body's natural frequency. If B(w) is small, the transient can take minutes of simulated time to decay. Use an excitation ramp, run long enough, and discard the transient before post-processing.
  • Radiation damping is evaluated once per time step from the current velocity (as for the other radiation methods). The resulting error is first order in the time step: about 0.6% amplitude near heave resonance at dt = 0.01 s for the IEA sphere, halving at dt = 0.005 s.