When home batteries create grid peaks, price feedback can change the answer

Cheap electricity can encourage thousands of home batteries to charge together, potentially creating a new peak at a neighborhood transformer. A new mathematical model explains why that risk can look much smaller when simulations account for charging’s effect on wholesale prices.

Home batteries responding to the same electricity tariff can behave like a crowd rushing through one doorway: everyone chooses the cheapest hours. But the doorway is not necessarily fixed. If enough batteries charge together, their demand raises wholesale prices, making adjacent hours more attractive.

In a preprint, Lion Hirth of the Hertie School and Neon Neue Energieökonomik develops an algebraic explanation of how that feedback changes local grid peaks. The paper derives formulas and compares them with earlier German simulations by the same author. It is not a field trial or an independent empirical validation.

Start with the transformer’s headroom

The model considers a low-voltage grid behind one transformer. Its background load is household demand net of local solar generation, before batteries act. Charging adds to that load; discharging subtracts from it.

A new import peak occurs when battery charging pushes demand above the grid’s previous maximum. The crucial comparison is therefore between charging power and the available headroom: the difference between the old peak and background demand during the charging hour.

When wholesale prices are treated as fixed, identical batteries facing the same signal charge at full rated power in the cheapest hours, under the model’s assumptions. The resulting charging-hour load is simply background demand plus the combined charging power of the local fleet.

In the German base case, 50 batteries each rated at 4 kilowatts add 200 kilowatts to a background load of 78 kilowatts. The resulting 278-kilowatt peak matches the simulation and exceeds the no-battery import maximum of 169 kilowatts.

How price feedback spreads charging

The alternative allows wholesale prices to rise with national residual load—electricity demand after accounting for wind and solar, with flexibility then changing the balance.

Charging fills the low-demand trough. In the idealized equilibrium, prices become flat across the charging window, removing the incentive to concentrate charging in its initially cheapest hour. Because the model assumes prices increase strictly with residual load, that price plateau corresponds to a flattened residual-load trough.

The amount of energy charged determines how deeply the trough is filled. This energy balance, rather than the precise price curve, sets the aggregate charging profile in the simplified case.

To translate that national profile into a neighborhood peak, Hirth assumes charging is allocated among batteries in proportion to their energy. This is important: equal prices alone do not uniquely determine which individual batteries charge when.

For a V-shaped trough, the local charging peak is s√(rE), using consistent units. Here, s is the neighborhood’s share of participating flexibility, r describes how steeply the trough’s sides rise, and E is the total energy charged into it. Adding local background demand gives the charging-hour transformer load.

Battery energy and the trough’s shape become decisive; rated charging power drops out only when it is high enough not to constrain that profile. With little participating flexibility, the charging window can remain narrow and full-power synchronization can return.

A close match, with revealing exceptions

For the critical winter night in the German simulation, the model predicts about 91 kilowatts of local charging, compared with 93 kilowatts simulated. Including background demand gives 159 kilowatts versus 163 kilowatts—both below the previous 169-kilowatt import peak.

That agreement depends on counting the appropriate flexibility. Storage technologies can have different costs and charging thresholds. In this case, the paper argues that home batteries displace utility-scale storage from the relevant trough. Including both fleets instead would predict only 136 kilowatts of transformer load, substantially understating the simulation.

The scaling relationship also has limits. Doubling battery energy and power produces a simulated charging increase of 57%, rather than the formula’s 41%, because the critical night changes. Removing the price component that does not respond to residual load lowers the simulated peak by 19%, exposing a synchronization mechanism outside the idealized trough-filling result.

A planning framework, not an all-clear

The model does not reliably identify which night will be critical. Battery foresight, profitable charging opportunities and price movements unrelated to residual load complicate that selection. It also assumes sufficient arbitrage incentives for full overnight charging.

Other risks remain: fixed tariff components can synchronize charging, and low local solar output during a national solar peak can leave a neighborhood exposed to midday imports. Transformer planning must also consider export flows and equipment capacity, not just the previous import maximum.

The preprint’s contribution is therefore a conditional framework: compare local headroom with charging shaped by the wider market, rather than assuming every battery always charges together at full power.