Apppublic

NaveenXSR17/Glider_Flap_Optimizer_Model_1

sourceHugging Facemitupdated 6mo agoView on Hugging Face

0likes

App README

✈ Glider Flap Optimizer

ML-powered aerodynamic surrogate model for predicting CL, Cm, CD and finding the optimal flap configuration for maximum glide ratio under full pitch-stability constraints.

What This Is

A gradient boosting surrogate model trained on 409 CFD simulation runs of a glider across 24 flap configurations. Instead of running a new CFD simulation for every candidate design, the surrogate predicts aerodynamic coefficients in milliseconds — then an optimizer sweeps the full configuration space to find the best flap setup.

Three inputs → Three outputs:

Input	Description	Range
Flap Position	Chord-wise hinge location	0.0, 0.65, 0.70, 0.75
Flap Angle	Deflection angle	0°, 2°, 4°, 6°, 8°, 10°
α (alpha)	Angle of attack	−3.5° to 9.5°

Output	Description	Test R²
CL	Lift coefficient	0.9988
Cm	Pitching moment coefficient	0.9953
CD	Drag coefficient	0.9972

Optimization Result

The optimizer finds the configuration that maximises CL/CD subject to three pitch-stability constraints:

1.Cm < 0 — trim condition (nose-down restoring moment)
2.dCm/dα < 0 — pitch stiffness (static stability)
3.Static Margin > 0 — CG ahead of neutral point

Parameter	Optimal Value
Flap Position	0.65
Flap Angle	2°
Trim Alpha	1.4°
CL/CD (glide ratio)	25.39
Static Margin	0.136 (13.6% MAC) ✓

Run Locally

bash

git clone https://huggingface.co/spaces/YOUR_USERNAME/glider-flap-optimizer
cd glider-flap-optimizer
pip install -r requirements.txt
streamlit run app.py

Tech Stack

Python · scikit-learn · Streamlit · NumPy · pandas · Matplotlib · joblib

Built as part of a glider aerodynamics design project. The surrogate model replaces expensive CFD calls during optimization, enabling real-time interactive exploration of the flap design space.