A maintained home for your STAN models

Who STAN is for

STAN ("subSTance flow ANalysis") was developed at TU Wien by Oliver Cencic and Helmut Rechberger. It is the reference tool of the academic material and substance flow analysis community: research, teaching, waste management. It implements the Austrian standard ÖNORM S 2096 and is cited across much of the literature in the field. It is free.

What we learned from STAN

Statistical rigour, and the way it displays it. STAN checks the plausibility of the model, reconciles data by weighted least squares, propagates error analytically, then returns diagnostics that publications in the field cite by name: degree of over-determination, objective function value, reconciliation quality index between 0 and 1, gross error detection through a test on the residuals, and a histogram of standardised scores. It couples goods and substance layers through concentration equations, chains stocks from one period to the next, rounds numbers according to the uncertainty attached to them, and lets you attach a literature reference and a remark to every value, layer by layer and period by period. On the traceability of a number inside a scientific paper, it is a model of its kind.

What we chose to do differently

STAN is a desktop calculation tool in which the diagram serves the mathematical model. That produces two structural limits.

The first is the data model: STAN organises values into layers (goods, substances, energy) crossed with periods of equal length. That is exactly what a substance analysis over time requires, and nothing else. A third dimension — a scenario, a region, a use, a supply chain — has no slot in the file: it becomes another file, to be maintained in parallel.

The second is the output: the result of a STAN study is a file or an image. The stan2web portal lets you deposit and search files, read-only, but a file is not a deliverable for a client who wants to explore. To that adds a detail that weighs in daily use: undo only covers moves and resizes, not changes to the model.

Finally, a point of context that is not a criticism of the software: its last published release dates from February 2022, and it relies on an ageing Windows runtime. For a lab starting a three-year PhD, that is a risk to weigh.

What we added

The same discipline, on a web foundation, with visualisation and sharing as first-class citizens: styles, crossed tags, multiple views of one model, legends, auto-layout, and the publication of an interactive diagram you share by URL and embed in a website. On the data side, a structured Excel workbook makes the full round-trip, where STAN goes through the clipboard. On the calculation side, our solver is free — no separately licensed third-party module is needed to handle large systems — it accepts minimum/maximum constraints, returns intervals when the information is insufficient, and uses Monte-Carlo sampling for uncertainty analysis. And dimensions are free and can be crossed: years, scenarios, regions, products, with hierarchical aggregation and disaggregation of nodes. See how, on a concrete example →

When to stay on STAN

Stay on STAN if your work relies on what we do not have yet: the concentration equations that couple goods and substances within a single resolution — the core of substance flow analysis —, the full statistical report (over-determination, quality index, standardised scores), uncertainty-driven rounding, or assisted naming from a nomenclature such as the European Waste Catalogue. Stay there too if your publication has to claim conformity with ÖNORM S 2096 as STAN implements it. The table below says, row by row, where we stand.

And you do not have to choose upfront: our importer reads your files, and your originals stay valid STAN files.

Teaching material flow analysis? Write to us, educational use is open to discussion: see pricing.

The door only opens one way: STAN cannot read our files, but we read both STAN formats, including the embedded drawing. Migrating loses neither your model nor your layout, and your originals stay valid STAN files.