Taxonomy

What the category codes mean

Every gallery page carries tags on five axes at once. The Categories panel on the Gallery groups them by axis — click a tag to filter, click Categories to clear. This page defines every code those tags use.

Axis	Prefix	Meaning	Example
Section	sec:	Catalog section — the physical domain	A transport closures, D fixed bed
Structure	struct:	Mathematical structure — the axis to filter on if you want to reuse code	S8 nested reactor–particle coupling
Tier	tier:	Standing in the literature	T0 canonical
Data	data:	Provenance tier of the dataset	tier1 open repository, tier6 reference solution
Phase	phase:	Phase system	gas, gas-solid

Two pages sharing a structure code are solved by nearly the same pymrm code even when their physics is unrelated — that is the intended route to reuse.

Section — physical domain (sec:)

The catalog is divided into ten sections by physical domain. This is what a visitor browses. Only the sections that already have a page appear in the filter panel.

Code	Section	Scope
A	Transport & closure relations	Dispersion, mass/heat transfer coefficients, pressure drop, effective properties, diffusion models
B	Catalyst particle & intraparticle transport	Thiele/effectiveness, pore models, deactivation, non-catalytic conversion
C	Reaction kinetics	LHHW, microkinetics, named industrial rate laws
D	Fixed / packed bed reactors	1D–2D, pseudo-homogeneous → heterogeneous, runaway, dynamics
E	Fluidised beds	Bubbling, turbulent, circulating, spouted
F	Gas–liquid & slurry reactors	Bubble columns, stirred tanks, airlift, slurry FT
G	Three-phase packed beds	Trickle bed, packed bubble, monolith slurry
H	Membrane reactors & separations	Pd, zeolite, polymeric, perovskite, membrane-assisted
I	Structured & intensified reactors	Monolith, foam, micro, reverse-flow, chemical looping
J	Adjacent unit operations	Adsorption/PSA, absorption/reactive separation, crystallisation/PBE, electrochemical, bio, polymer

Structure — mathematical structure (struct:)

This is the axis that makes the gallery useful to a coding agent. Every entry is tagged with one or more structure codes. An agent asked to “build me a model of X” first identifies the structure, then copies the nearest gallery page.

Code	Structure	pymrm ingredients
S1	ODE-IVP in time (batch, CSTR dynamics)	solve_ivp, or backward Euler + newton
S2	ODE-IVP in space (ideal PFR marching)	marching loop + newton per step
S3	1D steady BVP (particle, film)	construct_grad + construct_div + spsolve/newton
S4	1D transient PDE (axial dispersion, breakthrough)	+ accumulation matrix, time loop
S5	1D convection-dominated (sharp fronts)	construct_convflux_upwind + interp_cntr_to_stagg_tvd deferred correction
S6	2D PDE (radial dispersion, Graetz, monolith)	per-axis operators on (n_z, n_r, n_c), nu=1 radial
S7	Multi-domain / multi-phase coupling	(n_z, n_phase, n_c) layout, or update_csc_array_indices monolithic assembly
S8	Nested-scale coupling (reactor ↔︎ particle)	Schur complement via splu, shapes_d boundary unknowns
S9	Implicit multicomponent flux (Maxwell–Stefan, dusty gas)	per-face linear solve for fluxes, NumJac(axes_blocks=[-1])
S10	DAE / constrained (pressure–velocity, electroneutrality, equilibrium)	monolithic Jacobian with algebraic rows
S11	Population balance (internal coordinate)	treat size/age as an extra axis; growth = convection, so S5 machinery applies
S12	Moving boundary / shrinking core	front tracking, or fixed-grid with a phase indicator
S13	Non-standard geometry	construct_div(nu=callable) for arbitrary area profiles

Structure codes cluster the catalog in a way the physical taxonomy does not: a chromatography breakthrough curve (S4+S5) and a fixed-bed tracer RTD (S4+S5) are the same pymrm model with different closures. Cross-linking on S codes is what lets a researcher in one field reuse a page from another.

Tier — standing in the literature (tier:)

Rather than fake precise citation counts, entries are ranked in four tiers by their role in the literature. Tier is a judgement call and is revisited per entry when the page is written.

Code	Tier	Meaning
T0	Canonical	Named after its authors, taught in every reaction engineering course, still cited as the reference formulation decades later. (Thiele modulus, Ergun, Danckwerts BCs, Kunii–Levenspiel, Monod.)
T1	Field standard	The default model in its subfield; a paper there is expected to use it or justify not using it. (Xu–Froment, Doyle–Fuller–Newman, ASM1, Wakao–Funazkri.)
T2	Established alternative	Well-cited, competes with the T1 model, worth showing as a comparison. (Numaguchi–Kikuchi vs Xu–Froment; Syamlal–O’Brien vs Gidaspow drag.)
T3	Specialised / emerging	High quality, narrower use, or recent. (Electrified reforming, CO₂ electroreduction models, EMMS.)

Data — dataset provenance (data:)

Each page states where its validation data came from. The tier records the provenance, not the quality: a reference solution (tier6) is exact but is not experiment, while a figure digitisation (tier4) is real data recovered with some extraction error.

Code	Provenance
data:tier1	Open repository
data:tier2	Table printed in the paper
data:tier3	Supplementary material
data:tier4	Digitised from a figure
data:tier5	Obtained from the authors
data:tier6	Reference or exact solution, not experiment

See About for the legal basis on which digitised datasets are redistributed.

Phase — phase system (phase:)

The phase system present in the model, used to narrow the gallery to the kind of contactor you are working with.

Code	Phases
phase:gas	Single gas phase
phase:liquid	Single liquid phase
phase:gas-solid	Gas with a solid (catalyst, sorbent, bed)
phase:liquid-solid	Liquid with a solid
phase:gas-liquid	Gas and liquid
phase:gas-liquid-solid	Three-phase (e.g. trickle bed, slurry)

The page contract

Every gallery page, regardless of section, shows the same things in the same order, which is what makes the gallery scrapeable by an agent. See About for the full contract, and Contribute for how to add a page.