EV chargers, heat pumps, residential batteries, rooftop solar, smart hot-water cylinders — the load behind every meter is now a controllable asset. CEDeris is the platform layer that forecasts, aggregates, and dispatches them at the scale of a portfolio, on the same 27-zone European dispatch model that powers CEGridSight and the same battery optimiser that powers CE BESS Arbitrage.
Per-asset hourly forecasts for EV charging sessions, heat-pump COP-weighted demand, residential battery state-of-charge, and behind-the-meter PV — driven by weather, household type, tariff signal, and observed historical behaviour. Aggregates honestly: tail risk doesn't disappear when 50,000 households are summed.
Co-optimised participation in wholesale, balancing, and distribution-flexibility markets across a heterogeneous fleet — with per-asset opt-in, comfort constraints, and customer-fairness rules built into the objective rather than tacked on afterwards.
DSO constraints — feeder thermal limits, voltage envelopes, transformer ratings — enter the dispatch as hard constraints, not advisory caps. Every aggregator instruction is feasible against the local network, not just the wholesale price.
Plain-English first; this section for anyone vetting the maths. Skip if you're not running the aggregator yourself.
EV charging arrival/departure and energy-per-session modelled as marked point processes per household. Heat-pump load conditioned on outdoor temperature and a building thermal model with two RC time constants. Residential PV uses the same Perez transposition + Monin–Obukhov stack as CompoundVision. All trained on metered data and re-fit weekly.
Aggregate-level MILP sets price-quantity bids into wholesale and balancing markets; per-asset LP allocates the cleared volume back across the fleet under each asset's comfort, state-of-charge, and customer-opt-in constraints. Solved by CEMeridian (in build) once it ships; HiGHS today.
Feeder topology imported from DSO licence area data; thermal and voltage limits enforced as linear constraints in the per-asset LP. Where feeder data is unavailable, we fall back to a transformer-bus aggregation with conservative defaults rather than ignoring the network.
Heat-pump dispatch respects an indoor-temperature comfort band per household. EV dispatch respects a per-session energy and ready-by time. Residential battery dispatch respects a per-customer minimum reserve. Comfort violations are forbidden, not penalised — the optimiser cannot trade them away for revenue.
A platform layer underneath your customer app — forecasting, optimisation, and bid generation for assets you've already onboarded. You keep the customer relationship; we handle the dispatch.
Time-of-use and dynamic-tariff design that knows what your customers actually do — a heat-pump customer's tariff is co-designed against the dispatch model that will run their pump.
Procure flexibility you can trust to deliver. Every bid is backed by an asset-level dispatch plan that respects your feeder limits, not just the wholesale signal.
CEDeris is on the year-three slot of the platform roadmap. We're prototyping the per-asset forecast layer against an EV-charging dataset now; production launch follows once CE BESS Arbitrage and CEMeridian are stable. Customers on early access shape the spec — particularly on tariff structures, opt-in semantics, and DSO data formats.