The simulator

It answers three questions from one set of dials - how much gas, what it costs to build and pay back, and how many tonnes and acres it eats. Critically it has no model of its own: it runs the same engine over the same assembled graph the canvas does, so the two can never quote different numbers for the same plant. It is a first-pass sizing tool, not a DPR.

Build the plant, price the feedstock, wire the schematic, read the result - with the numbers updating as you go. It mirrors the real decision sequence from feedstock to fuel to margin, which is why the learning hub is organised the same way.

Feedstock type and tonnes per day, the delivered feedstock price in rupees per tonne, the digester volume and temperature, upgrading purity and compressor efficiency, and the offtake route with its price and transport charge. Those are the dials; everything else is computed from them.

Three headline figures - margin, payback and CBG sold - plus three panels: the process envelope, showing whether the digester is inside its safe operating window; economics; and supply, the feedstock calendar. Margin is daily revenue less daily operating cost and excludes capex service; payback is simple and undiscounted.

It is the safe operating window for the digester, read from three numbers: HRT, OLR and blended C:N. The panel flags attention when HRT falls below 18 days, OLR rises above 4 kg VS per cubic metre per day, or C:N leaves the 20-35 band, and it fails at OLR above 5. Those thresholds exist because acidification from overloading is the failure that costs weeks of output.

It shows how feedstock availability moves across the year against the plant's daily demand, exposing the seasonal gap that idles plants - paddy straw arriving in a six-week window, press mud stopping the day the mill stops crushing. It operationalises the sector's dominant failure mode, which is aggregation rather than technology.

No, and the panel says so on its face. The supply calendar is descriptive - it drives copy, calendars and advisory flags - while the engine runs a single steady-state design point at a flat rupees-per-tonne price. Monthly availability curves and storage carry move into the model at a later stage; until they do, the panel does not pretend otherwise.

Change an input - feedstock mix, delivered price, plant size, upgrading purity - and watch margin, the envelope and the calendar respond. That is the core learning loop: make one change, read the consequence, and find which lever your project is actually sensitive to.

They are conservative planning values per feedstock, drawn from the shipped parameter packs with their provenance attached - which you can read in full on the components page. Override them with your own tested figures where you have them, because real Napier and crop-residue yields vary by a factor of two between sites.

Every parameter traces to an entry in the shipped packs, each carrying a value, a band, a status and a note. The components page joins those to the block they belong to and lets you filter for figures nobody has cited yet - which is a real filter, because a lot of the pack is not yet sourced.