What this runs, and how far to trust it
The same CHO model as the Python simulator, ported line for line and checked against it in CI: the two agree to better than one part in a hundred thousand on every state, every hour, across five test processes.
It runs the preview case only. DO is held at 40 %, pH at 7.10 and temperature at its setpoint — ideal control, not the controllers, gas lines and sensors the full simulator models. The glucose assay is exact and sampled on the chosen interval; the top-up is 50 % glucose, delivered over half an hour, from 24 h on, when glucose reads below 97 % of target. The run stops if the volume reaches 1.6× its starting value.
Against the full physical simulation of the default process the preview reads about 1 point high on final viability and about 3–4 % high on titer. Use it to see which way and roughly how far a change moves a run; confirm a decision with the Python simulator.
Parameter defaults are calibrated to one GS CHO process (peak VCD around day 4–6, ammonia rising in the last week). Tags: measured reasonably well established across CHO lines, typical a representative value that varies several-fold between lines, shape sets the form of a response rather than its size. Hover a parameter name for its justification.
Nothing you enter leaves this page: the model runs entirely in your browser. A copied link carries only the settings you changed.