Teaching with OpenFLUME

OpenFLUME is a finite-volume thermo-fluid network solver that runs entirely in the browser. In each of the four labs below, students predict a result by hand, build or open the matching model, run it, and explain the difference. Each lab ends with an answer key you can strip out before handing it to students.

OpenFLUME is alpha software (version 0.3.1), written by one engineer with heavy AI assistance (see "How This Was Built" in the README). The labs follow the project's motto, check it rather than trust it: the hand calculation comes first, and the solver result is something to verify.

Classroom features

• No install. Students open openflume.github.io/OpenFLUME in a current Chromium-, Firefox-, or WebKit-based browser (user manual §2.1), including on school-managed laptops and Chromebooks.
• No license server. There are no seat counts and nothing expires at the end of the semester. The software is MIT-licensed.
• Nothing is uploaded. The solver runs in a Web Worker on the student's machine. Models autosave to browser storage and are saved as .fn files on the student's disk.
• Inspectable. Text view shows the entire model as plain text, the same text as a .fn file. Students can see every number the solver uses, and you can see every number a student changed.
• Verified against references. The labs reuse cases from the validation reports, where each case is compared against closed-form or independent reference solutions.

Limits to tell students about

OpenFLUME is a lumped network code, not a CFD code. The limitations relevant to these labs:

• There is no shock capture. Lab 4 shows the effect at a choked duct exit.
• Wave propagation is not modelled. Water hammer and acoustic waves are outside the model; fluid inertia is lumped per branch.
• Turbomachinery is represented by a pump curve only.
• Unlike streams mix only at reacting junctions. There is no general junction mixing.
• Two-phase flow uses the homogeneous equilibrium model (HEM) only. There is no separated-flow (Lockhart–Martinelli) correlation yet.
• Cavitation is modelled only at venturi closure.
• A steady closed ideal-gas loop needs a pressure anchor.
• User-authored code is trusted, not sandboxed. Loading a file with embedded code triggers a trust prompt. Tell students to accept code only from you.

The complete list is in user manual §1.7.

Suggested courses

Course	Where OpenFLUME fits	Labs and examples
Undergraduate fluid mechanics	Darcy–Weisbach, series/parallel pipes, loop networks, pump operating points	Lab 1; Three-pipe junction, Water distribution network
Heat transfer	Thermal resistance networks, radiation, lumped capacitance, fins, conjugate heat transfer	Lab 3; Heated pipe with radiating wall (conjugate HT), Water-water counterflow heat exchanger
Thermodynamics	Open-system energy balances, adiabatic and isothermal blowdown, fill heating	Lab 2; Tank blowdown
Compressible flow / propulsion	Choked orifices, Fanno flow, feed systems	Labs 2 and 4; LOX/RP-1 thruster (combustor) for a capstone look
Spacecraft thermal	Radiation to deep space, conductor networks, loop heat pipes	Lab 3; Spacecraft radiator panel (ammonia loop heat pipe)
Microfluidics	Capillary-driven filling, stop and trigger valves	Capillary circuit (stop + trigger valve) and the microfluidics validation report

The examples are in the toolbar Examples ▾ menu. Each also has a direct link of the form https://openflume.github.io/OpenFLUME/#example=<slug>, which the labs use.

The labs

Lab	Topic	Learning objectives	Time
Lab 1	Pipe networks	Apply Darcy–Weisbach with an explicit friction factor; solve series, branching, and parallel pipe problems; run a Hardy-Cross loop correction on a multi-loop network; read a hydraulic grade line	60 min pre-lab, 2 h lab
Lab 2	Tank blowdown	Derive the choked adiabatic blowdown solution; explain expansion cooling with an open-system energy balance; show first-order time-step convergence; compare adiabatic and isothermal blowdown	45 min pre-lab, 1.5 h lab
Lab 3	Conduction and thermal networks	Build solid nodes and conductors; solve a composite wall as series resistances; find a nonlinear radiation–convection equilibrium; integrate a lumped cooldown; close the energy balance of a conjugate model; discretize a fin	45 min pre-lab, 2 h lab
Lab 4	Fanno flow	Compute a Fanno choking length from tables; set up a quasi-1-D compressible duct with the Momentum flux and Kinetic energy settings; measure grid convergence; recognize the behaviour of a network solver beyond choking	45 min pre-lab, 1.5 h lab

Lab 1 needs no prior OpenFLUME experience. Labs 2 to 4 assume students have done the Getting started page or Lab 1.

Distributing a starting model

1. Built-in example links. Where a lab starts from a shipped example, the lab page links straight to it (for example https://openflume.github.io/OpenFLUME/#example=tank-blowdown). The link loads a fresh copy of the example. If the student already has a model open, the app asks before replacing it.
2. .fn file. For a model you built or edited, click Save to download it as a .fn text file and post it to your LMS or send it by email. Students open it with Load. Plain text works well with version control and diffs.

Remind students that loading an example or a file replaces the open model, so they should Save their own work first.

How students submit

Ask for:

1. The model: the .fn file from Save.
2. The results: when a model has been run, Save also writes a <model>.runs.json file containing the recorded runs (user manual §6.12). Students can also attach CSV files from the Results tab: Export CSV writes the current plot's channels and Export all writes every channel. Each chart can be exported as PNG or SVG with the run's provenance in the footer.
3. The hand calculation and answers to the lab questions, in your usual format.

Results are also kept in browser storage, but clearing site data deletes them. Students should save the .runs.json file before closing the tab on anything they need to hand in.

Grading tips

• Load the submitted .fn and run it yourself. The status pill should read Converged. If it says stale, the student edited the model after the run their numbers came from.
• Check the inputs in Text view. Most wrong answers come from a wrong input, such as a diameter in millimetres entered as metres or a Darcy friction factor entered where the textbook value was Fanning. Text view and Table view show every value in SI.
• Watch the unit preset. The Units ▾ menu changes what the property panel displays, but .fn files and the Sweep tab are always in SI. Having students set Units ▾ to SI for these labs avoids most unit confusion.
• Grade the comparison as well as the number. The useful skill is explaining why a hand result and a solver result differ: a friction-factor guess, a time step, or grid resolution. Each answer key states the expected size of the difference and its cause.
• Use the Findings. Under each plot, the Results tab lists automatic findings such as the dominant pressure drop, reverse flow, near-sonic flow, and junction mass imbalance (user manual §6.8).
• Change the numbers between terms. The answer keys are public on this site. Changing one or two inputs each term (a pipe diameter, a boundary pressure, a heat load) keeps the method and makes copied keys useless. The hand-calculation procedure in each lab works for any reasonable inputs, and the solver gives you a new key in seconds.

Answer keys

Each lab ends with its answer key inside a collapsible container:

::: details Answer key
...
:::

On the published site the key stays collapsed until a reader clicks it. To make a student handout, copy the lab's Markdown source from the repository (docs/teach/*.md) and delete everything from ::: details Answer key to the closing :::. The rest of the page does not depend on the key.

Every number in a key is either computed independently from the stated inputs (closed forms, or a separate Newton or bisection solve written for this kit) or copied from the cited validation report. Values from OpenFLUME runs are labelled as solver results from version 0.3.1. The last digit can change between versions; please report differences larger than the stated tolerance as an issue.

License for reuse

• Lab text: the text of this educator kit (this page and the four lab pages) is released under the Creative Commons Attribution 4.0 International license (CC BY 4.0). You may copy, adapt, translate, and redistribute it, including in course packs and commercial settings, provided you give credit. A suitable attribution is: "Adapted from the OpenFLUME educator kit, J. Rising, https://openflume.github.io/teach/, CC BY 4.0."
• Software: OpenFLUME itself is MIT-licensed (see LICENSE in the repository).
• Citing the software in a course or paper: use the Zenodo DOI 10.5281/zenodo.22051608; the repository's CITATION.cff has the full record.

Feedback

If you use OpenFLUME in a course, please post in GitHub Discussions with the course, the level, which labs you ran, and what went wrong. Bug reports from lab sessions are especially useful: a confusing interface step, a wrong answer-key number, or a model that would not converge on a student machine. Corrections to these pages are welcome as pull requests.