EMS packages

Monitoring, control, then AI

Three packages, matching the simulator: Monitoring, Control, then AI. Start with visibility, add dispatch, then forecasts and a rolling schedule. Quoted per estate. There is no public list price. After sign-in, switch Monitoring, Control, and AI in the simulator header to see the same split on a live-style site.

01

Base

Monitoring

Live power, SOC, grid kW, alarms and reports. Dispatch does not change.

A single view of BESS, solar, CHP and grid status, including sites that already have batteries.

  • Live dashboard: site energy overview, power summary, battery SOC, and grid import/export kW
  • Sites and fleet portfolio: solar, load, average SOC, and alarm counts across locations
  • Asset inventory: BESS, solar, CHP, generators, and other on-site plant
  • Energy page: live balance and session trends (load, generation, import, export)
  • Alarms for faults and constraint breaches, not routine dispatch noise
  • Reports: peak and average load, generation, SOC range, and a rule-events log
  • Site configuration overview (read-only)

Simulator · Dashboard, Sites, Assets, Energy, Alarms, Reports, Settings.

02

Includes Monitoring

Control

Write dispatch rules for 30-minute tariff windows, peak shave, and export or import limits. Commission the site in the same EMS.

Sites that charge on 30-minute tariff windows, hold DNO caps, or need a discharge schedule. New plant or a control retrofit on existing BESS.

  • Everything in Monitoring
  • Dispatch rules: 30-minute tariff-window charge and discharge, SOC thresholds, load deficit, time-of-use schedules, peak shave, export limit, import limit, import pressure, grid charge, and composite rules
  • Rule runner: see which conditions pass and which rules fire, condition by condition
  • Scenario playbooks: evening peak, battery depleted, import pressure, sunny surplus, off-peak charge
  • Site editor: assets, topology, coupling, and grid limits
  • Live asset enable/disable and dispatch changes from the dashboard

Simulator · Adds the Rules page, site editor and playbooks. Still kW and kWh, not pounds.

03

Includes Control

AI: predict, optimise, control

Forecasts and a rolling schedule for net site value. A model does not run the plant.

Sites that want a 24 to 48 hour plan that updates when solar, load, tariffs or plant change, plus a P&L of what ArcGrid did.

  • ML forecasts of load, solar, prices, and asset behaviour from site telemetry and external data
  • Optimisation engine: best schedule subject to SOC, grid limits, CHP, production, battery degradation, and reserve
  • Setpoints to the existing control layer. ArcGrid sits above OEM controllers.
  • Rolling re-optimise every 5–15 minutes across the next 24–48 hours as actuals arrive
  • Net site value P&L: baseline vs optimised cost, flex revenue, degradation, and the plan in plain English
  • Everything in Control

Simulator · Simulator AI mode previews forecasts, rolling suggestions, and £ vs a grid-only baseline. Live sites run the optimiser against commissioned telemetry.

How AI works

How the AI package works

Forecasts feed an optimiser. The optimiser sends targets into the Control layer you already run.

  1. 01

    Ingest

    Site telemetry plus tariffs, weather and market prices.

  2. 02

    Predict

    ML forecasts load, solar, prices, and how assets will actually behave.

  3. 03

    Decide

    The optimiser solves the schedule against SOC, grid limits, CHP, production, degradation, and reserve.

  4. 04

    Control

    ArcGrid sends bounded setpoints to the existing OEM control layer.

  5. 05

    Learn

    Actual performance feeds back in. The problem is solved again.

Tomorrow the best plan might be: charge the BESS from excess solar, hold SOC for the evening peak, run CHP only in the highest-value window, shift a flexible load by 45 minutes, and keep 20% battery reserve because forecast uncertainty is high.

Maximise net site value = export and flexibility revenue + avoided import + avoided peak charges − energy purchases − fuel − battery degradation − other operating costs.

Built in stages

Commercially useful before the ML models are sophisticated. The site model and optimiser come first; forecasts get better as telemetry accumulates.

  1. 1

    Deterministic optimiser

    Tariffs, SOC, battery limits, solar, CHP costs, grid limits and schedules. MILP / MPC first.

  2. 2

    ML forecasting

    Site load and solar first, trained on the telemetry you already collect.

  3. 3

    Financial optimisation

    Degradation cost, efficiencies, dynamic pricing, and flexibility revenue in the objective.

  4. 4

    Asset intelligence

    Battery degradation, abnormal rack behaviour, inverter efficiency, predictive maintenance.

  5. 5

    Portfolio optimisation

    Several ArcGrid sites together for flexibility and market opportunities.

Package questions

What are the ArcGrid EMS packages?
Three packages, matching the simulator: Monitoring (live view, alarms, reports), Control (dispatch rules, playbooks, site editor), and AI (forecasts, rolling schedule, setpoints into Control, and a P&L). Control includes Monitoring. AI includes Control. A model does not run the plant.
Does AI control the site?
No. ML predicts load, solar, prices, and asset behaviour. An optimisation engine decides the schedule. ArcGrid then sends bounded setpoints into the existing Control layer and OEM controllers. Safety-critical plant control stays where it is. The plan re-solves every 5–15 minutes across the next 24–48 hours as actuals change.
Can I try the packages before buying?
Yes. The free site simulator uses the same three modes. Switch Monitoring, Control and AI in the header after sign-in. It shows dispatch behaviour. It is not a live control system.
How are the packages priced?
Commercial ArcGrid EMS is quoted per estate (sites, assets, DNO and OEM scope). The three packages describe capability, not a public catalogue price. Email hello@arcgrid.co.uk or use the enquiry form for a proposal.

Enquire about a package

Tell us whether you need Monitoring, Control, or AI. Or email hello@arcgrid.co.uk.

We use this to reply about a demo or proposal. See our privacy policy.