# Hmt Monte Carlo

> Run Monte Carlo uncertainty analysis on a hydraulic model by sampling parameters from statistical distributions and computing exceedance probabilities. Use this when the user wants uncertainty analysis, Monte Carlo simulation, or probabilistic results. Use when this capability is needed.

- Skill: `tomevault-io/hmt-monte-carlo` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/hmt-monte-carlo`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/hmt-monte-carlo/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/hmt-monte-carlo

---


Run Monte Carlo uncertainty analysis on a hydraulic model.

## Prerequisites
- Project in a `base_case/` subdirectory
- Parameter uncertainty distributions defined by user
- Solver paths configured in `hmt_config.json`

## Steps

1. **Check that a project is open.**
   If not, ask the user to run `/hmt-open` pointing to `base_case/<project_file>`.

2. **Show available materials.**
   ```bash
   hmt-cli get_materials
   ```

3. **Ask the user for uncertain parameters.** For each:
   - Material name or BC ID
   - Distribution: `"truncated_normal"` (default) or `"uniform"`
   - For `truncated_normal`: mean, std, min, max
   - For `uniform`: min, max

4. **Build validated parameter specifications.**
   ```bash
   hmt-cli build_param_specs --args '{"specs": [
     {"type": "manning_n", "material_name": "<name>", "distribution": "truncated_normal",
      "mean": <mean>, "std": <std>, "min": <min>, "max": <max>},
     ...
   ]}'
   ```
   Confirm the distributions with the user.

5. **Ask for MC settings:** number of samples (50–200), random seed (default 42),
   output directory (default `./mc_runs`), parallel processes (default 1).

6. **Generate samples.**
   ```bash
   hmt-cli generate_mc_samples --args '{"param_specs": <specs>, "n_samples": <N>, "random_seed": <seed>, "output_csv": "mc_samples.csv"}'
   ```
   Show a preview of the first 5 sample rows.

7. **Run Monte Carlo simulations.**
   ```bash
   hmt-cli run_monte_carlo --args '{"base_case_dir": "./base_case", "param_specs": <specs>, "n_samples": <N>, "n_processes": <procs>, "random_seed": <seed>, "sample_csv": "mc_samples.csv", "delete_cases": true, "output_dir": "<output_dir>"}'
   ```
   Periodically show progress while waiting:
   ```bash
   tail -20 mc_progress.log
   ```

8. **Report:** successful/failed runs, results JSON path.

9. **Compute statistics.** Ask for observation point coordinates if the user wants point statistics.
   ```bash
   hmt-cli get_mc_statistics --args '{"results_json": "<path>", "observation_points": [{"name": "<name>", "x": <x>, "y": <y>}], "exceedance_probabilities": [99, 90, 50, 10, 1]}'
   ```
   Present exceedance table: Point | P99 | P90 | P50 | P10 | P1

10. **Report** the spatial exceedance VTK path for ParaView visualization.

**Troubleshooting:**
- Many failed cases → check `pyHMT2D.log`
- Variable detection fails → pass `"variable": "<name>"` explicitly

---
> Source: [psu-efd/pyHMT2D](https://github.com/psu-efd/pyHMT2D) — distributed by [TomeVault](https://tomevault.io).
<!-- tomevault:4.0:skill_md:2026-06-19 -->