Probabilistic and sensitivity analysis tools built on top of STEM, for studying how uncertain soil, load and damping parameters affect railway-induced vibration predictions.
Start with the tutorial: docs/tutorial_sensitivity.rst. It walks through building a small 3D embankment model, then applies three methods to it:
- Uncertainty quantification -- Monte Carlo / Latin Hypercube sampling and Random Fields, to see the distribution of a predicted response.
- Sensitivity analysis -- the Morris method (and RBD-FAST as an alternative), to screen which parameters matter most.
Each chapter's code blocks are complete and can be pasted into a single script, or run directly from the matching file in docs/ (e.g. tutorial_sensitivity_morris.py).
docs/-- the tutorial and the scripts it's built from.legacy/-- earlier exploratory work, kept for reference but not maintained.
Install dependencies with pip install -r requirements.txt (includes STEM and its Kratos wheels). See the tutorial for STEM-specific setup notes.