# Open Source Software for Seastead Wave Simulation

## Executive Summary

Given your specific requirements—partially submerged NACA foil legs, high accuracy needs, GPU availability, and desire for visualization—here's my recommendation:

**Primary Choice: DualSPHysics**
**Secondary Choice: Chrono::FSI with SPH**

Capytain/MoorDyn and WEC-Sim should be deprioritized due to BEM limitations with your partially submerged foils.

---

## Software Comparison Table

| Software | Method | GPU Support | Learning Curve | Accuracy for Your Case | Setup Time |
|----------|--------|-------------|----------------|------------------------|-------------|
| **DualSPHysics** | SPH | Excellent (CUDA/OpenCL) | Medium | ✅ Excellent | 1-2 weeks |
| **Chrono::FSI** | SPH + Multi-body | Moderate (CUDA) | High | ✅ Excellent | 3-4 weeks |
| **OpenFOAM** | CFD (VOF) | Moderate | Very High | ✅ Excellent | 4-8 weeks |
| **Capytain** | BEM | None | Low | ⚠️ Limited | 1-2 weeks |
| **MoorDyn** | Mooring only | None | Low | ✅ Good (with BEM) | 1-2 weeks |
| **WEC-Sim** | BEM + Multi-body | Moderate | Medium | ⚠️ Limited | 2-3 weeks |
| **HAMS** | BEM | None | Low | ⚠️ Limited | 1 week |
| **Blender** | N/A | GPU | Low | ❌ Not engineering | N/A |

---

## Detailed Analysis

### 1. DualSPHysics

**Best for your needs** due to SPH method handling partially submerged bodies well.

**Pros:**
- True Smoothed Particle Hydrodynamics—excellent for free surface flows
- Native GPU acceleration (CUDA/OpenCL)—your GPU will be fully utilized
- Handles complex geometries without mesh generation issues
- Excellent for bodies that emerge/submerge dynamically
- Open source (GPL v3)
- Active development community
- ParaView integration for visualization

**Cons:**
- Particle methods require many particles for accuracy (computational cost)
- Setting up fluid-structure coupling requires work
- Less mature than some alternatives

**Accuracy for your case:** ✅ **Excellent**
- SPH naturally handles the free surface boundary conditions you need
- No linearization assumptions—captures nonlinear wave effects
- Can model wave breaking and splash-up

**Estimated setup time with Claude Code:** 1-2 weeks to first meaningful simulation

**Scriptability:** Python API available, JSON configuration files

---

### 2. Chrono::FSI

**Strong option if you want integrated multi-body dynamics.**

**Pros:**
- Combines SPH fluid solver with Chrono's multi-body dynamics
- Can model your entire structure including joints/constraints
- GPU support via CUDA
- Well-documented multi-body dynamics engine
- Can couple with other physics (ropes, impacts)

**Cons:**
- More complex overall system
- SPH implementation newer than DualSPHysics
- Documentation can be sparse for FSI specifically
- Steeper learning curve for fluid-structure coupling

**Accuracy for your case:** ✅ **Excellent**

**Estimated setup time with Claude Code:** 3-4 weeks to first meaningful simulation

---

### 3. OpenFOAM

**Industry-standard CFD but overkill for your needs.**

**Pros:**
- Industry-validated accuracy
- interFoam/multiphaseEulerFoam for free surface
- Extremely flexible—can model anything
- Mature community and documentation

**Cons:**
- Mesh generation is complex for your geometry
- Primarily CPU-based (some GPU support emerging)
- Extremely steep learning curve
- May require 4-8 weeks to get reliable results

**Accuracy for your case:** ✅ **Excellent** (if set up correctly)

**Recommendation:** Only if you need to model propulsion, detailed wake effects, or have specific CFD validation requirements.

---

### 4. Capytain + MoorDyn

**Fast for linear analysis, but limited for your geometry.**

**Pros:**
- Fast frequency-domain calculations
- Good for mooring analysis when coupled with MoorDyn
- Python interface is clean
- Good for parametric studies

**Cons:**
- **BEM limitations are real for your case**: Your partially submerged foils violate BEM assumptions
- Linearized wave theory
- Cannot capture strongly nonlinear effects
- May significantly underestimate motions when foils emerge

**Accuracy for your case:** ⚠️ **Limited**—not recommended as primary tool

**If you still want to try:**
```python
# Example Capytain setup (simplified)
import capytain
from capytain import Float, Stationary

# Your geometry would need careful meshing
# Capytain uses PANEL method meshes
```

---

### 5. WEC-Sim

**Not open source despite being free to download.**

**Pros:**
- Designed for wave-structure interaction
- Good visualization tools
- Active marine renewables community

**Cons:**
- **Requires MATLAB** (~$1,000-2,000 for non-students)
- Requires Simulink and Simscape Multibody
- Uses BEM under the hood (same limitations as Capytain)
- Not truly open source

**Cost:** MATLAB + Toolboxes ≈ $1,500-3,000+

**Not recommended** given your open-source requirement.

---

### 6. HAMS (Hydrostatic/Hydrodynamic Analysis of Marine Structures)

**Specialized BEM solver from Shanghai Jiao Tong University.**

**Pros:**
- Very fast
- Good for traditional ship/platform shapes
- Python interface
- Well-validated for hull forms

**Cons:**
- BEM limitations apply
- Designed for conventional hulls, not foil shapes
- May struggle with your unusual geometry

**Accuracy for your case:** ⚠️ **Limited**

---

### 7. Blender Physics

**Not suitable for engineering analysis.**

Blender's physics engine is designed for visual effects and games, not engineering analysis. It lacks:
- Proper fluid-structure coupling
- Wave theory foundations
- Engineering-grade accuracy

**Exception:** If you want to create animated visualizations of results from another simulation tool, Blender is excellent for that.

---

## Why BEM is Problematic for Your Design

Your 3-foot-wide NACA foil legs that submerge from 50% to nearly 100% create several issues for BEM:

1. **Wetted surface changes**: BEM assumes small changes in wetted area
2. **Nonlinear effects**: BEM linearizes about the mean waterline
3. **Foils near free surface**: BEM assumptions break down when foils operate near the free surface
4. **Tip effects**: Your foils can emerge entirely in some conditions

**SPH and VOF/CFD methods don't have these limitations**—they solve the full Navier-Stokes equations at the free surface.

---

## My Recommendation: DualSPHysics

Given your priorities:

1. **GPU utilization**: DualSPHysics is built for GPU from the ground up
2. **Free surface handling**: SPH excels at tracking complex free surfaces
3. **Partially submerged bodies**: No linearization needed
4. **Visualization**: Output to VTK/CSV, visualize in ParaView
5. **Open source**: Fully open, no licensing costs
6. **Python interface**: You can script everything

### Implementation Path

1. **Week 1**: Install DualSPHysics, learn basics, test 2D wave tank
2. **Week 2**: Create your 3D geometry (STL import), test static buoyancy
3. **Week 3**: Add wave generation, validate results against theory
4. **Week 4**: Run your first dynamic simulation, visualize in ParaView
5. **Week 5+**: Parametric studies, compare designs

### Alternative: Chrono::FSI

If you later want to add:
- Rope dynamics (your dinghy ropes)
- Joint constraints
- Contact/collision

Chrono::FSI would be better suited. Consider it for Phase 2.

---

## Visualization Recommendations

Regardless of which simulation tool you choose:

**ParaView** (free, excellent)
- Industry-standard scientific visualization
- Handles large particle/mesh datasets
- Can create animations of wave-structure interaction
- Python scripting available

**Blender** (for final renders)
- Import simulation data
- Create photorealistic visualizations
- Useful for presentations/website

---

## Design Iteration Workflow

```
Design Change → Python Script → Run Simulation → Visualize Results
                    ↓
              Claude Code helps write scripts
                    ↓
              Batch parametric studies
```

Once you have one design working, running variations is straightforward:
1. Modify geometry parameters in Python
2. Re-run simulation
3. Compare motion responses
4. Iterate

---

## Questions to Consider

Before committing to a tool, consider:

1. **What wave conditions do you need to test?** Regular waves, irregular seas (JONSWAP spectrum), extreme events?

2. **Do you need to model mooring?** If so, you'll want MoorDyn integration regardless of your main solver.

3. **What's your turnaround time?** Daily iterations vs. weekly deep dives affects tool choice.

4. **Do you need to model wind effects** in addition to waves? This would add complexity.

---

## Summary Table

| Priority | Recommendation |
|----------|----------------|
| **Best overall** | DualSPHysics |
| **If you need multi-body later** | Chrono::FSI |
| **If you only need rough estimates** | Capytain |
| **If you want industry CFD** | OpenFOAM |
| **For pretty pictures** | Blender (post-simulation) |
| **Not recommended** | WEC-Sim (cost), Blender physics (accuracy) |

---

Shall I help you get started with a specific software package? I can provide more detailed setup instructions, sample configuration files, or help you create your first simulation geometry.