What an energy management system changes

An energy management system (EMS) is the controller, software and hardware together, that decides when a home battery and an EV charge or discharge. EnergyFlow models two levels of that control. What the extra intelligence is worth depends on your prices.

Rules and price-aware optimisation

Rule-based control is what a plain inverter (the box between your panels and your house) and battery do. Each hour it asks one question: is there more PV than the house needs? If yes, it stores the surplus and exports the rest. If no, it draws on the battery first and the grid second. It knows nothing about prices.

A price-aware controller looks at a whole day. It knows the hourly prices and the expected PV and demand, and it plans the battery for the day that minimises the bill: buy when energy is cheap, hold it, use it when it is dear. Buying low to use or sell high is called arbitrage. Both controllers work within the same battery limits.

In EnergyFlow the second mode is an optimiser: it finds the cheapest plan among all those that respect the limits. It plans one calendar day at a time (see the methodology article). Heat pump, hot water and pool demand are inputs. Only battery and EV flows are planned.

Which prices it works from

A dynamic tariff charges a price that changes every hour with the wholesale electricity market, instead of one fixed price. EnergyFlow builds it from day-ahead prices: the wholesale price for each hour of tomorrow, set in an auction the day before. We hold hourly day-ahead files for each country we cover: France, Belgium, the Netherlands, Austria and Switzerland each have their own price zone, and Luxembourg and Germany share one. The import price adds a network fee and VAT, and the export price is a share of the market price. The optimiser plans each day knowing that day’s 24 prices, as a real controller could once they are published.

When a battery charges from the grid

In rule-based mode, never. The battery charges from PV surplus only.

The optimiser charges from the grid when buying now and using the energy later costs less than the alternative. As a rule of thumb, it buys when the price it will avoid paying (or earn by exporting) later in the same day, multiplied by the round-trip efficiency (the share of energy that comes back out of the battery), is still above the price now. At the default 92 %, the battery returns 92 kWh for every 100 it takes in, so the later price must beat the price now by a clear margin. It can also export PV in a high-price hour instead of saving it for the evening, which raises revenue and lowers self-consumption.

Why a fixed tariff leaves little to optimise

With one import price all day there is no cheap hour to buy in. The only decision left is what to do with a PV surplus: store it or export it. When the export price is below the import price, the usual case, storing it wins whenever the battery has room. That is exactly what the rule-based mode does. So on a fixed tariff, which in EnergyFlow has one import price and one export price, a price-aware controller has little to add and the optimiser does not run.

Comparing the two in EnergyFlow

There is no dispatch mode switch. Dispatch is the hour-by-hour sharing of electricity between PV, battery, EV and grid, and its mode follows your tariff and battery:

The recap on the results page names the mode that ran, the price-aware optimiser or the rules.

Three comparisons work well. Keep the house, weather, schedules and PV the same in each.

• Fixed tariff against dynamic tariff, with the same battery. The difference mixes two things, the price structure and the control.
• Dynamic tariff with and without a battery. The gap shows what the battery and its optimiser add together.
• Dynamic tariff with different battery sizes. This shows where a bigger battery stops adding value for that price series.

EnergyFlow does not run the rule-based mode on a dynamic tariff with a battery, so the optimiser’s contribution cannot be separated from the tariff’s.

Mind the weather year. A real-year run is priced with that calendar year’s market prices, and a typical year at a fixed reference year, so changing the weather year changes the price year too.

The car as a second battery

Vehicle-to-home (V2H) lets a parked electric car power the house through a bidirectional charger. More car models offer it each year, and a car battery is often larger than a home battery, so it is worth testing as extra storage before you buy a bigger home battery.

A house with PV, a home battery, a heat pump and a car that can give power back has many flows to share out every hour. That is the job of the EMS, and the more storage the house has, the more its decisions matter. The car also comes with limits a home battery does not have: it is away during the day, it must keep enough charge for the next trip, and carmakers cap how much energy it may give back. Add V2H to the same house and compare the self-consumption and the annual cost with and without it. The methodology article explains how the model handles the car.

What to look at in the results

The annual cost is split into energy cost net of export and a fixed grid charge. The fixed charge does not depend on the mode, so compare the energy cost. Then check the flows. Grid import to the battery stays close to zero in rule-based mode, and it is where the optimiser’s grid charging shows up. Export may rise, and self-consumption may fall even when the bill improves (see the self-consumption article).

Limits of the model

• The optimiser knows each day’s PV and demand exactly. A real EMS works from forecasts, so a real saving would be smaller than the model suggests.
• It plans each calendar day on its own and puts no value on charge left at midnight, so it does not hold energy back for tomorrow’s prices.
• Only battery and EV flows are planned. Heating, hot water and pool demand do not follow prices.
• Grid export caps, peak-power fees (charges based on your highest power draw), supplier restrictions and the real firmware of a commercial product are not modelled.
• At the moment the default dynamic-tariff formula uses Luxembourg values, so adjust the fee, VAT and export share for your contract.
• The result is an estimate, not a prediction of what a specific product will do in your house.