A demo repository for a materials qualification.

The demo focuses on tungsten and uses:

• Inputs

  • dpa
  • irradiation_temperature_c
  • impurity_fraction

• Outputs

  • lower_yield_stress_mpa
  • hardness_gpa
  • thermal_diffusivity_mm2_s

This repository contains:repos:

• Streamlit app
• simulator package
• tests
• notebook

The simulator now includes two practical constraints to mimic real experiments:

• Measured DPA noise: the achieved/measured DPA can differ from the requested value. The simulator records and returns the measured DPA, and this value is used in the app tables and plots.
• Discrete impurity stock: impurity fraction is selected from a finite set of available tungsten samples. For each request, the nearest available impurity fraction is used and recorded.

In the Streamlit app these effects are enabled by default (with a fixed random seed for reproducibility). In the core simulator config, DPA noise defaults to 0.0 so unit tests and scripted use remain deterministic unless explicitly enabled.

This simulator is a demonstrator, not a validated materials model.

It should not be used for engineering design decisions.

poetry install
poetry run streamlit run src/material_qualification_demo/app.py

or via the Poetry script entrypoint:

poetry run material-qualification-demo

material-qualification-demo/
├── assets/
│   └── digiLab logo.png
├── notebooks/
│   └── tungsten_simulator_walkthrough.ipynb
├── src/
│   └── material_qualification_demo/
│       ├── __init__.py
│       ├── app.py
│       ├── branding.py
│       └── simulators/
│           ├── __init__.py
│           ├── base.py
│           └── tungsten.py
├── tests/
│   └── test_tungsten_simulator.py
├── pyproject.toml
└── README.md

The app includes a very simple initial design of experiments using two anchor points:

1. the minimum corner of the domain
2. the maximum corner of the domain

So if the user specifies bounds for:

• dpa
• irradiation_temperature_c
• impurity_fraction

the initial DoE is:

[
    [dpa_min, temp_min, impurity_min],
    [dpa_max, temp_max, impurity_max],
]

from material_qualification_demo.simulators.tungsten import (
    TungstenQualificationConfig,
    TungstenQualificationSimulator,
)

config = TungstenQualificationConfig()
sim = TungstenQualificationSimulator(config)

X = [
    [0.0, 300.0, 0.001],
    [2.0, 900.0, 0.010],
]

outputs = sim.forward(X)
print(outputs[0])

• alloy family selector
• helium / hydrogen transmutation variables
• uncertainty model on outputs
• synthetic observation noise
• larger DOE options (LHS, Sobol, random)
• surrogate model fitting inside the app
• export of training dataframes