About

Purpose

Chemical engineering has an unusually rich stock of phenomenological models — Thiele’s effectiveness factor, Kunii–Levenspiel’s bubbling bed, the Doyle–Fuller–Newman battery model — that are taught everywhere and reimplemented constantly, usually from scratch and usually without the original data at hand.

This gallery collects them in one place, each reproduced with pymrm and shown against the experimental data it was validated on, as a runnable notebook.

The page contract

Every page shows the same four things, in the same order:

the model as originally published, in the original notation, with a table mapping symbols to code variables;
the core experimental data, as a plain CSV with a machine-readable provenance sidecar;
the pymrm reproduction overlaid on that data, with a quantitative agreement metric that CI tracks for regressions;
what pymrm adds — and if it adds nothing, the page says so. A faithful reproduction is a legitimate result; a manufactured improvement is not.

Every page also carries at least one validation that does not use the experimental data: an analytical limit, a conservation check, or a grid/time-step independence check.

Classification

Two axes. Physical domain (sec:) is what you browse. Mathematical structure (struct:) is what you reuse — codes S1–S13 describing the shape of the numerical problem rather than the physics. Pages sharing a structure code are solved by nearly the same pymrm code even when their subject matter is unrelated, which is the intended route to adapting a model from outside your own field.

Full scheme: Taxonomy.

Licensing

Component	Licence
Code — notebooks, `shared/`, `scripts/`	MIT
Prose and figures generated here	CC-BY-4.0
Data we generate ourselves	CC0
Data digitised from third-party publications	not relicensed — see below

Digitised datasets are distributed as factual data with attribution. Numerical values are facts and are not protected by copyright; extraction for scientific research by a research organisation with lawful access is covered by Art. 15n of the Dutch Auteurswet, implementing Art. 3 of the DSM Directive (EU) 2019/790. Source figure images are never reproduced — points are re-plotted from extracted coordinates — and source PDFs are never committed. Each dataset’s .meta.yaml records its origin, extraction method, and basis for redistribution.

If you are an author whose data appears here and you would like a page changed or removed, please open an issue; we will act on it promptly.

Relationship to the rest of PyMRM

pymrm — the library itself.
pymrm-book — the course book: tutorials, exercises, and lecture notes for the Multiphase Reactor Modeling course at TU Eindhoven.
this gallery — validated reproductions of published models, for research reuse rather than teaching.

The book teaches you the method. The gallery gives you a model to start from.

Citing

Cite both the original publication and this gallery. Each page lists its primary reference; the gallery itself has a CITATION.cff and a DOI per release.

--- title: "About" --- ## Purpose Chemical engineering has an unusually rich stock of phenomenological models — Thiele's effectiveness factor, Kunii–Levenspiel's bubbling bed, the Doyle–Fuller–Newman battery model — that are taught everywhere and reimplemented constantly, usually from scratch and usually without the original data at hand. This gallery collects them in one place, each reproduced with `pymrm` and shown against the experimental data it was validated on, as a runnable notebook. ## The page contract Every page shows the same four things, in the same order: 1. the model **as originally published**, in the original notation, with a table mapping symbols to code variables; 2. the **core experimental data**, as a plain CSV with a machine-readable provenance sidecar; 3. the **pymrm reproduction** overlaid on that data, with a quantitative agreement metric that CI tracks for regressions; 4. **what pymrm adds** — and if it adds nothing, the page says so. A faithful reproduction is a legitimate result; a manufactured improvement is not. Every page also carries at least one validation that does not use the experimental data: an analytical limit, a conservation check, or a grid/time-step independence check. ## Classification Two axes. **Physical domain** (`sec:`) is what you browse. **Mathematical structure** (`struct:`) is what you reuse — codes `S1`–`S13` describing the shape of the numerical problem rather than the physics. Pages sharing a structure code are solved by nearly the same pymrm code even when their subject matter is unrelated, which is the intended route to adapting a model from outside your own field. Full scheme: [Taxonomy](taxonomy.qmd). ## Licensing | Component | Licence | |---|---| | Code — notebooks, `shared/`, `scripts/` | MIT | | Prose and figures generated here | CC-BY-4.0 | | Data we generate ourselves | CC0 | | Data digitised from third-party publications | not relicensed — see below | Digitised datasets are distributed as **factual data with attribution**. Numerical values are facts and are not protected by copyright; extraction for scientific research by a research organisation with lawful access is covered by Art. 15n of the Dutch Auteurswet, implementing Art. 3 of the DSM Directive (EU) 2019/790. Source figure images are never reproduced — points are re-plotted from extracted coordinates — and source PDFs are never committed. Each dataset's `.meta.yaml` records its origin, extraction method, and basis for redistribution. If you are an author whose data appears here and you would like a page changed or removed, please open an issue; we will act on it promptly. ## Relationship to the rest of PyMRM - [`pymrm`](https://github.com/computational-chemical-engineering/pymrm) — the library itself. - `pymrm-book` — the course book: tutorials, exercises, and lecture notes for the Multiphase Reactor Modeling course at TU Eindhoven. - **this gallery** — validated reproductions of published models, for research reuse rather than teaching. The book teaches you the method. The gallery gives you a model to start from. ## Citing Cite **both** the original publication and this gallery. Each page lists its primary reference; the gallery itself has a `CITATION.cff` and a DOI per release.