Cable Sag Predictor

In this project, I built a simple data-driven model that predicts cable sag from a few design parameters. In effect, it's a lightweight surrogate model that approximates the Kangaroo physics simulation — once fitted, it estimates sag instantly, without re-running the simulation. 💡

The workflow was pretty straightforward:

First, I generated about 25 training samples using the Kangaroo plugin.

Here are the variables: 🔸 Inputs: Span, Cable Length, and Gravity (Load)
🔹 Output: Cable Sag

Then, I used the Galapagos evolutionary algorithm to fit the model and minimize the prediction error against the dataset. 🎯

Finally, I tested it with entirely new data. Even with just 25 training samples, the results for this simple model were surprisingly accurate and impressive!

---

🛠️ Tools Used:

• Rhinoceros 3D / Grasshopper
• Kangaroo Physics (for data generation)
• Galapagos (for model optimization)