As solar floods the grid at noon and vanishes by evening, prices swing from negative to triple-digit within hours. Stromfee Ai is the intelligence layer that reads those swings and turns them into value — forecasting, dispatch and battery optimisation, proven on a German fleet, now for the Indian Ocean.
The glass battery · Mauritius
Watch the battery made transparent: charging on cheap midday solar, discharging into the evening peak. Vendor-agnostic dispatch on top of the certified safety controller — the heart of the whole program.
Midday over-generation and curtailment. Steep evening peaks on ageing thermal plants. Falling system inertia, frequency and voltage stress. The fix is flexibility — batteries that charge when power is free and discharge when it's scarce — but a battery without intelligence leaves most of that value on the table. That intelligence is what we bring.
Vendor-agnostic, market-aware, built on two decades of running real assets. Each tool works on its own — together they form the operating brain for a renewable island grid.
Day-ahead & intraday price preview — the signal that drives every dispatch decision. With learning units.
live · learnHour-by-hour day-ahead market curve with charge & discharge windows. With learning units.
live50 Hz stability monitoring — fast frequency response & reserve.
coreCharge-low / discharge-high dispatch on any battery make. The value engine.
Orchestrates charging across heterogeneous assets within grid & SLA limits.
Detects PV plants, models expected yield and revenue exposure.
Watches each plant for under-performance, curtailment and faults.
The BESS Optimizer's training modules — how the engine learns each asset.
Blends every flexible asset into one optimal day-ahead arbitrage plan.
A real European day-ahead reference, 14 June 2026. Power was negative at midday (−25 €/MWh) and +112 €/MWh by the evening peak. Every euro of that spread is what a well-run battery captures — and what an idle one misses.
The optimiser buys the trough and sells the peak — automatically, every day, on whatever battery hardware is installed.
Source: ENTSO-E day-ahead reference via Stromfee data pipeline. Mauritius runs a single-buyer model today — the same engine drives SLA-based grid-support dispatch.
A high-renewable island grid has less inertia and is more vulnerable to disturbance. Grid-forming batteries deliver fast frequency response and reserve — our layer decides when, how hard, and for how long.
The BESS Optimizer is vendor-agnostic — it speaks to 173 manufacturers (Modbus / SunSpec / API), reads and writes the control registers, and dispatches for revenue on top of the certified safety controller. The Charge Director coordinates it across a heterogeneous fleet, within every grid and SLA limit.
No black box. Every cell, every set-point, every euro of revenue is visible — the "glass battery". That transparency is how operators, banks and grid operators learn to trust autonomous dispatch.
The optimiser isn't a fixed rule-set — it learns the behaviour of every asset and market it runs. These are its training modules.
Map the battery: protocol, registers, limits, round-trip efficiency, degradation profile.
Learn day-ahead & intraday patterns, weekend over-generation, evening peaks.
Translate forecast + SLA into charge/discharge set-points within grid limits.
FFR, reserve, reactive power — when to hold capacity vs. arbitrage.
Trade cycles against revenue so the asset still meets its 20-year guarantee.
Measure realised vs. expected, correct the model, compound the gains.
Before you can optimise the grid, you have to see it. PV Radar finds and models the solar fleet; PV Watcher keeps every plant honest.
Detects PV installations, models expected generation against irradiance & weather, and flags regulatory / revenue exposure (curtailment, feed-in rules). The map of the asset base before a single decision is made.
Watches each plant continuously for under-performance, soiling, string faults and unexpected curtailment — turning "it should be producing" into "it is, and here's the proof".
The Automixer blends every flexible asset — batteries, CHP, solar — into a single day-ahead arbitrage plan, automatically. It decides the cheapest, highest-value way to serve the grid, hour by hour.
Stromfee Ai is the AI platform of HR Energiemanagement GmbH — a German energy-management company with more than 20 years of operating real generation and storage assets against live power markets. We don't theorise about flexibility; we have run it daily, across ~50 sites and ~72 MW, for years. There is no harder school than the German grid — extreme solar and wind penetration, negative prices, redispatch and frequency stress make it the world's toughest proving ground for AI BESS dispatch and control frameworks. Battle-tested there, ready for the Indian Ocean. Mauritius can be helped — with proven hands, not promises.
Whether you operate a grid, finance a project, or build the batteries — the intelligence layer is ready. Let's talk about the first pilot.