**Open Source Software Comparison for Seastead Wave Simulation** **High-Accuracy, Visual, Non-Linear CFD/SPH Options for Large Motion Foil-Leg Platforms** ```html Seastead Simulation Software Comparison (2025)

Seastead Simulation Software Comparison (2025)

Target: Accurate non-linear simulation of a triangular truss platform with three 19 ft NACA-foil legs (large heave/pitch/roll excursions, partial emergence/submergence of legs). High-fidelity visualization of motions, accelerations at any point, and failure modes in increasingly severe waves. Open-source only. GPU acceleration preferred.

Comparison Table

Software	Method	Handles Large Non-Linear Motions?	Visualization Quality	GPU Acceleration	Est. Time to First Working Simulation (with Claude Code/Cursor)	Time to Change to New Design	Engineering Accuracy for Your Case	Learning Curve / Notes
DualSPHysics 6.0+	Smoothed Particle Hydrodynamics (SPH)	Excellent (fully non-linear, 6-DoF rigid bodies, wetting/drying)	Very Good (ParaView or built-in viewer → easy to make videos like the one you linked)	Excellent (CUDA/OpenCL – uses your good GPU very well)	2–4 weeks	1–3 days	Very High for your needs (captures leg emergence, slamming, damping from legs very well)	Python + XML case setup. Very good documentation and many marine examples. Claude Code excels here.
Project Chrono + Chrono::FSI-SPH	SPH (or DEM) coupled with multibody dynamics	Excellent	Good to Very Good (ParaView or Chrono::Viz)	Good (GPU SPH solver available)	3–5 weeks	2–4 days	Very High – you correctly noted that Chrono::FSI-SPH does NOT use BEM. It is a full SPH solver.	Strong multibody & control capabilities. Good if you later want to add the stabilizers and actuators.
OpenFOAM (v2312 or later) interDyMFoam / overInterDyMFoam	Volume-of-Fluid (VOF) + overset or dynamic mesh	Very Good (with overset mesh can handle large 6-DoF motion)	Excellent (ParaView – best post-processing in the list)	Moderate (some GPU solvers exist but CPU still dominant for most marine cases)	6–10 weeks	3–7 days	Highest possible (full viscous + turbulence + wave generation)	Most accurate but slowest to set up and run. Steepest learning curve. Best for final validation.
Capytaine + MoorDyn + Python	Boundary Element Method (linear + some weakly non-linear)	Poor – assumes small motions	Limited (requires extra visualization work)	None	1–2 weeks	1 day	Not suitable – your 19 ft legs go from 50 % to nearly 100 % submergence. Linear BEM will be inaccurate.	Fast for conventional WECs but wrong tool for your large-motion foil-leg platform.
Blender + Add-ons (Flip Fluids, Mantaflow, or custom rigid-body + ocean modifier)	Mostly artistic FLIP / Lattice Boltzmann or procedural waves	Moderate	Excellent artistic rendering	Good	2–3 weeks	1–2 days	Low engineering accuracy – not acceptable for quantitative acceleration & failure analysis.	Still not accurate enough for engineering decisions. Good only for pretty videos, not for "high level of engineering accuracy".
WEC-Sim + MoorDyn	Mostly linear + some non-linear hydro coeffs (BEM-based)	Moderate	MATLAB plots → needs extra work for nice videos	None	N/A (not open-source)	N/A	Not suitable for your large leg excursions.	Requires MATLAB + Simulink + Simscape Multibody. For a non-student in Anguilla this easily exceeds $2,000–$4,000 USD per year. Not recommended.
PreCICE + OpenFOAM / SU2 / CalculiX coupling	Partitioned FSI (VOF + FEM or rigid body)	Very Good	Very Good	Moderate	8–12 weeks	4–8 days	Very High	Extremely powerful but very complex. Overkill at this stage.
caisson / OceanWave3D + HOS (high-order spectral waves)	Potential flow + fully non-linear waves	Only small body motions	Limited	Some GPU versions	4–6 weeks	2–4 days	Good for wave generation, poor for your large-motion legs.	Not ideal.

Detailed Recommendations & Realistic Timelines (with Claude Code assistance)

1. DualSPHysics (Top Recommendation for you right now)

Fully non-linear, mesh-free, excellent at capturing slamming, emergence, and viscous damping of your foil legs.
GPU acceleration is mature and fast on your hardware.
Easy to output accelerations at any point on the triangular truss.
Very good community examples for floating bodies and wave generation (Piston / relaxation zone).
Visualization: export to ParaView → you can make videos almost identical to the one you showed.

Estimated time with Claude Code: 2–4 weeks to first realistic simulation, then 1–3 days per new design variant.

2. Project Chrono::FSI-SPH

Answers your question directly: Chrono::FSI-SPH does NOT use BEM. It is a full SPH solver coupled to Chrono’s rigid-body / multibody engine.
Very natural for your truss + three legs model (define geometry as meshes or primitives).
Built-in support for adding control (you can later add your stabilizer actuators easily).

Estimated time: 3–5 weeks to first working case.

3. OpenFOAM (for highest accuracy validation)

Use overInterDyMFoam with overset mesh or interDyMFoam with morphing mesh.
Can give you the most accurate drag, slamming loads, and green-water effects.
Run times are longer; you will use your GPU less effectively than with SPH codes.

Estimated time: 6–10 weeks for a robust setup. Best used after you have screened designs with DualSPHysics.

Why the others are not suitable

Capytaine / NEMOH / WAMIT-style BEM: Linear or weakly non-linear assumptions break down when your legs move from 50 % to nearly 100 % submergence and you have significant wave slamming.
Blender physics: Still not engineering-grade. The FLIP solver in Flip Fluids is improved but not validated for quantitative acceleration predictions on offshore structures. Use only for marketing videos.
WEC-Sim: Not open source, expensive license stack, and still mostly based on linear hydro coefficients.

Recommended Path (Most Efficient)

Start with DualSPHysics 6.0 (or latest) on Linux or Windows. Use Claude Code to generate the XML case files and Python pre/post-processing scripts.
Model the triangular truss as a single rigid body with correct mass, inertia, and center of gravity. Model the three NACA legs accurately (you can import STL meshes).
Use a relaxation zone or piston wavemaker to generate regular and irregular waves. Gradually increase wave height until you see instability or excessive accelerations.
Extract heave, pitch, roll, and point accelerations at the living-area corners.
Once you are comfortable, validate the most interesting cases with OpenFOAM or Chrono::FSI-SPH.
After the first model works, changing geometry (different leg length, different triangle aspect ratio, different truss height, etc.) is relatively fast — usually 1–3 days per new configuration with good scripting.

Additional Free Tools You Will Use

ParaView – best open-source visualization and video export.
Salome or Gmsh – for clean STL mesh generation of your NACA legs and truss (if needed).
Python (with numpy, pandas, matplotlib, pyvista) – scripting everything. Claude Code is excellent at this.

Bottom line (2025): DualSPHysics gives you the best combination of accuracy, visualization quality, GPU usage, and speed-to-first-result for your unusual three-foil-leg seastead. Chrono::FSI-SPH is a very close second and worth testing in parallel. OpenFOAM should be used for final high-fidelity validation runs.

Would you like me to generate the initial DualSPHysics case template (XML + Python scripts) for your exact 80 ft × 40 ft triangular platform with three 19 ft NACA legs?

```