Explore examples from ANULUM’s own research and software practice, with the source material and limitations available beside each result.
These examples describe ANULUM’s own public work. They are not customer testimonials.
Research communication
Turn a scientific relationship into an experiment
Problem: a formula is difficult to explore without seeing how its inputs affect the result.
Implementation: the compression explorer connects declared inputs to the same closed-form relationships documented by the reactor family. Published presets and source links sit beside the calculations.
Observable result: changing an input updates the quantities and plots. The lab adds comparison, settings links and data exports.
Boundary: the ideal limits do not predict yield or validate a machine.
Problem: a result without its assumptions and source is difficult to assess.
Implementation: the reactor comparison links individual numerical anchors to device pages and references, while distinguishing reproduced, consistent and unresolved cases.
Observable result: a reader can follow a comparison row back to its stated source.
Boundary: a reproduced anchor does not by itself establish whole-device validation.
Problem: visitors need to understand which project serves their goal and where its evidence lives.
Implementation: the public portfolio records purpose, status and source links. Universal search, the project finder and the atlas reuse those public records.
Observable result: visitors can search a concept, compare project facts and follow a publication or software link.
Boundary: displayed metadata is a dated snapshot, and a relationship link does not prove integration.
These are software demonstrations and reading copies of executed notebooks. They do not reproduce physical experiments or qualify hardware. Retain the source version, environment and your outputs when running a notebook.
Synchronisation
Input: the coupling and oscillator settings exposed by the demo.
Model and output: run the simulation and inspect the displayed phase evolution and synchronisation measures; export the displayed results to retain the run.
Limit: a settings link starts a fresh trajectory. This teaching model does not reproduce a published hardware experiment.
Input: the example network and input arrays defined in the quick-start notebook.
Model and output: follow the Python cells and inspect the layer outputs and the example stochastic computation.
Limit: example and random inputs are computational illustrations; fresh runs may produce different numbers. They are not recordings of biological neural activity or measurements of a physical chip.
Input: the coil configuration, target boundary and regularisation parameter declared by notebook 09.
Model and output: follow the Grad–Shafranov and coil-current optimisation cells; inspect the modelled flux, boundary and current summaries.
Limit: this reading copy identifies SCPN-Fusion-Core v3.5.0. Use its declared setup when reproducing it; a modelled equilibrium does not validate an operating tokamak or a real-time controller.