A battery generates value by charging and discharging energy at the right moments — which means its value is decided entirely by the decisions someone makes with it, hour by hour. Take that decision out of the model and you no longer have an accurate picture of your project over 30 years; you have a rough energy balance.
That is why storage cannot be modelled on averages, and why it's easy to understate what a battery earns while overstating how much of it you need.
01 / WHAT'S NEWA simulated operator inside every configuration
The Artificial Operator (AO) runs inside every one of the thousands of configurations in a sweep. Each hour it looks forward across a rolling window, decides whether to charge, hold or discharge to reach its objective, acts, and re-plans.
Decisions, not reactions
The operator shifts wind and solar output out of the hours where it is worth nothing and into the hours that pay — holding back from a worthless midday to sell into an early-evening peak worth three times as much, or into a delivery window you have committed to.
262,800 decisions per configuration, with degradation, round-trip losses and the grid connection limit respected at every hour.
Forecast quality is an input
An operator who knows exactly what the next 24 hours hold will dispatch perfectly. Nobody does. So the AO takes the degree of forecast uncertainty as a setting, applied to resource, price and demand across a rolling look-ahead window.
Run the same configuration at both ends and the answer stops being a single optimistic number.
02 / THE STRATEGIESThree dispatch strategies
You choose what the operator is optimising for, per scenario.
Maximise revenue
Shifts energy into the highest-priced hours of the forecast price profile. For merchant exposure and day-ahead style arbitrage.
Match demand
Serves a delivery obligation — an industrial offtake, or a data centre that must run 95% of the time — to the reliability target you set.
Reactive baseline — blind
Charge on surplus, discharge on deficit, no view of the market at all. Kept deliberately: the gap to a forecast-led run is what dispatch intelligence is worth on your site.
03 / WHY IT MATTERSForecast quality moves the optimum
Dispatch is not a detail downstream of sizing — it decides which configuration wins. An operator with good visibility holds a firmness target with less battery, or earns more from the same battery. Degrade the forecast and the winning corner of the design space moves, usually toward more storage.
Perfect foresight is a clean upper bound — the most the asset could earn and the least storage it could need. Useful as a benchmark, not as a business case. Run both and the cost of imperfect visibility is priced in, rather than quietly assumed away.
04 / IN THE APPAvailable on every run, from the first pass
The AO runs on every simulation from the first pass, on estimated inputs. As measured resource campaigns, real price curves and signed contract terms land, the same model re-runs and the dispatch schedule — with the revenue that follows from it — sharpens with them, from first sizing through to financial close.
Coming next
On the roadmapSub-hourly time steps, so fast-response and ancillary revenue can be modelled at the resolution those markets actually settle at.