surrogate-shm-api

Author: Mahdi Mohammadzadeh

A FastAPI service that wraps the trained XGBoost surrogate model from surrogate-shm-model and serves structural response predictions over HTTP.

Given six beam geometry and material parameters, the API returns predicted tip deflection, max bending stress, and first natural frequency in under 10ms.

What it predicts

The surrogate was trained on 10,000 samples generated from Euler-Bernoulli beam theory. It accepts the following inputs:

Parameter	Symbol	Unit
Beam length	L	m
Cross-section width	b	m
Cross-section height	h	m
Young's modulus	E	Pa
Applied load	F	N
Density	rho	kg/m³

And returns:

Response	Symbol	Unit
Tip deflection	delta	m
Max bending stress	sigma	Pa
First natural frequency	f1	Hz

Run with Docker

docker build -t surrogate-shm-api .
docker run -p 8000:8000 surrogate-shm-api

API docs available at http://localhost:8000/docs

Run without Docker

pip install -r requirements.txt
uvicorn app.main:app --reload

Example request

curl -X POST http://localhost:8000/predict \
  -H "Content-Type: application/json" \
  -d '{
    "L": 1.5,
    "b": 0.05,
    "h": 0.04,
    "E": 200000000000,
    "F": 5000,
    "rho": 7850
  }'

Example response

{
  "tip_deflection_m": 0.002341,
  "max_bending_stress_Pa": 175781250.0,
  "natural_frequency_Hz": 13.47
}

Endpoints

Method	Path	Description
GET	/health	Returns service status
POST	/predict	Runs surrogate prediction

Stack

Python, FastAPI, XGBoost, scikit-learn, Docker

Related

surrogate-shm-model rul-prediction-api