i
Supply gets variable.
Wind and solar are weather-coupled, not dispatchable. The system has to be operated against probabilities, not setpoints.
Adaptive power consumption control
Adaptive control for the decarbonised grid.
As electricity supply becomes variable, the action moves to the demand side. eForecast develops the adaptive control layer that lets consumption respond to a transitioning grid, intelligently and automatically, across any load that can adapt.
How it works02 · The shift
A decarbonised grid is a different control problem.
Adding renewables to the supply mix doesn't just lower emissions. It changes the nature of the system. Supply becomes weather-coupled, forecasts replace dispatch certainty, inertia falls, and stability has to be re-engineered from the demand side. The research moves with it.
ii
Forecasts replace certainty.
Operating decisions now ride on forecast outputs. Control has to be robust to the forecast being wrong, sometimes badly wrong.
iii
Inertia disappears.
Synchronous generation is displaced. Frequency and stability margins shrink. The fast response increasingly has to come from loads, storage and inverter-based resources.
iv
The action moves to the demand side.
When supply can't be ordered up, demand has to meet supply. That rewires what a load is: from passive draw to active participant.
03 · Approach
Forecasts in. Control out.
Forecasting electricity carbon intensity has become broadly available infrastructure. eForecast consumes those forecasts and turns them into decisions: when each load should draw, how much, and how to stay correct when the forecast turns out to be wrong.
In · forecast feeds
- EI Grid CO₂ intensity (gCO₂/kWh)
- WX Renewable generation forecast
- DM System demand forecast
- PX Market / price signals
- σ Forecast uncertainty bands
Adaptive control
A control policy that balances objectives (carbon, cost, stability) against constraints (deadlines, comfort bands, ratings) under forecast uncertainty.
Out · control signals
- k(t) Time-varying setpoint
- σ̂ Robustness margin
- P̃ Dispatch schedule
- ↺ Re-plan on update
- ∑ Portfolio aggregate
04 · Problem space
Where the research is concentrated.
Adaptive control is a deep, broad surface area. eForecast's active work clusters around four problems that surface again and again as electricity systems decarbonise.
Grid stability under high renewables.
As inertia falls and inverter-based generation rises, frequency and voltage stability have to be supported from elsewhere, increasingly from controllable loads and distributed storage.
Control under forecast uncertainty.
Forecasts are never exact. The interesting work is in policies that remain correct as the forecast updates, that hedge against tail outcomes, and that recover gracefully when the forecast is wrong.
Intelligent battery operation.
When and how to charge, when to dispatch, how to age the asset. The decision space for storage is rich and high-dimensional. Adaptive control across multiple objectives is well-suited to it.
Control across many sites.
A single load is one problem. A heterogeneous portfolio (different assets, constraints, geographies, owners) is a different one. The coordination is where most of the value, and most of the difficulty, sits.
The list isn't exhaustive. Adaptive control applies wherever a load can shift, defer or modulate, across electric mobility, heating, industry, data centres, distributed energy resources and beyond.
05 · Intellectual property
Active across the energy ecosystem.
eForecast holds multiple patent-pending positions across the broader energy ecosystem: in adaptive control, in the use of forecasts under uncertainty, and in the orchestration of distributed loads and storage.
06 · Company
Based in Dublin, Ireland.
eForecast was founded in 2022, originally to forecast the carbon intensity of electricity. The work has since matured into adaptive control: the systems that turn forecasts of a variable grid into operating decisions for individual loads and portfolios of them.
More about us
Latest
Notes and updates from the work appear occasionally under News.