PyMRM Gallery

Published chemical engineering models, reproduced against their original data

What this is

A growing, open collection of the important phenomenological models of chemical engineering — packed beds, fluidised beds, bubble columns, membrane reactors, catalyst particles, and the unit operations next door — each one reproduced with pymrm and shown against the experimental data it was originally validated on.

Every page carries the same four things:

the model as originally published, in the original notation;
the core experimental data, as a plain CSV with full provenance;
the pymrm reproduction, overlaid on that data with a quantitative agreement metric;
what pymrm adds — a relaxed assumption, a finer discretisation, a limit the original could not reach. Or an honest statement that the page is a faithful reproduction and adds nothing.

Every page is a Jupyter notebook you can download and run, unmodified, in Google Colab.

Browse the gallery

Three ways to use it

Find a model and adapt it. Filter the gallery by physical domain (sec:) to find a model of your system, then edit the parameters. Each page ends with a Reuse section listing exactly what to change.

Find a structure and adapt it. More powerful, and less obvious: filter by struct: instead. A diffusion-limited enzyme pellet, a catalyst pellet, and the solid phase of a lithium-ion battery electrode are the same mathematical problem, so the same pymrm code solves all three. If nothing in your field matches, something in someone else’s probably does.

Let a coding agent do it. The catalog is published as machine-readable models.yaml and the repository carries an AGENTS.md with instructions. Point your agent at the repository and ask it to build a model of your system; it is designed to find the structurally nearest page and adapt it rather than assemble something from API documentation.

Status

Early. The literature survey catalogues 266 models across ten sections; pages are being built in priority order, sequenced by data availability rather than by model importance.

Contributions are welcome and individually credited — see Contribute.

--- title: "PyMRM Gallery" subtitle: "Published chemical engineering models, reproduced against their original data" --- ## What this is A growing, open collection of the important phenomenological models of chemical engineering — packed beds, fluidised beds, bubble columns, membrane reactors, catalyst particles, and the unit operations next door — each one reproduced with [`pymrm`](https://github.com/computational-chemical-engineering/pymrm) and shown **against the experimental data it was originally validated on**. Every page carries the same four things: 1. the model **as originally published**, in the original notation; 2. the **core experimental data**, as a plain CSV with full provenance; 3. the **pymrm reproduction**, overlaid on that data with a quantitative agreement metric; 4. **what pymrm adds** — a relaxed assumption, a finer discretisation, a limit the original could not reach. Or an honest statement that the page is a faithful reproduction and adds nothing. Every page is a Jupyter notebook you can download and run, unmodified, in Google Colab. [Browse the gallery](gallery.qmd){.btn .btn-primary} ## Three ways to use it **Find a model and adapt it.** Filter the [gallery](gallery.qmd) by physical domain (`sec:`) to find a model of your system, then edit the parameters. Each page ends with a *Reuse* section listing exactly what to change. **Find a *structure* and adapt it.** More powerful, and less obvious: filter by `struct:` instead. A diffusion-limited enzyme pellet, a catalyst pellet, and the solid phase of a lithium-ion battery electrode are the *same mathematical problem*, so the same pymrm code solves all three. If nothing in your field matches, something in someone else's probably does. **Let a coding agent do it.** The catalog is published as machine-readable [`models.yaml`](https://github.com/computational-chemical-engineering/pymrm-gallery/blob/main/models.yaml) and the repository carries an [`AGENTS.md`](https://github.com/computational-chemical-engineering/pymrm-gallery/blob/main/AGENTS.md) with instructions. Point your agent at the repository and ask it to build a model of your system; it is designed to find the structurally nearest page and adapt it rather than assemble something from API documentation. ## Status Early. The [literature survey](https://github.com/computational-chemical-engineering/pymrm-gallery/tree/main/docs) catalogues **266 models** across ten sections; pages are being built in priority order, sequenced by data availability rather than by model importance. Contributions are welcome and individually credited — see [Contribute](contributing.qmd).