PV forecast
Expected production by interval, calibrated to the actual array and its recent performance.
Battery planning
The right morning battery target is not always 100%.
A useful target covers expected demand and reserve while leaving enough room for forecast solar. It changes when tomorrow changes.
Short answer
Charge for the forecast deficit, not from habit.
A fixed rule such as "charge fully at 02:00" treats every day as the same. Forecast-based charging asks how much energy the home is likely to need before the next good charging opportunity, then adds reserve and an uncertainty margin.
The result may be a full battery before a cloudy winter day, a partial charge before strong midday solar or no grid charge when stored energy is already sufficient.
Five inputs
A weather forecast alone is not a battery plan.
Load forecast
Expected household consumption, including EV, heating, cooling and unusual routines.
Prices
Actual retail import and export values, not only a wholesale chart.
Battery state
Usable capacity, current charge, power limits, efficiency and permitted operating range.
Protected reserve
Energy kept for outage risk, essential loads and household priorities.
Practical model
A target can be simple enough to explain.
target stored energy ~ reserve + expected net deficit + uncertainty margin Suppose a home has a 10 kWh usable battery. It protects 2 kWh for reserve, expects a 4 kWh net deficit before the next charging opportunity and adds a 1 kWh forecast margin. The planning target is about 7 kWh.
This is a heuristic, not a universal controller formula. A real system also converts between state of charge and usable energy, accounts for losses and checks power, warranty, tariff and export constraints.
Three different days
The same battery should not receive the same instruction.
| Forecast | Likely plan | Why |
|---|---|---|
| Sunny, export-limited day | Keep useful battery space for midday PV. | A full morning battery may force low-value export or solar curtailment. |
| Cloudy day, expensive evening | Consider charging in a permitted cheap window. | Stored low-cost energy may avoid later expensive imports while preserving reserve. |
| Uncertain weather or outage risk | Increase the margin or protected reserve. | The cost-optimal plan should not outrank resilience when forecast confidence falls. |
Replanning
A good plan expires quickly.
Forecasts are wrong in useful ways. Clouds arrive early, household demand changes and an EV plugs in unexpectedly. The controller should re-evaluate the remaining day when material inputs change rather than execute a 24-hour schedule blindly.
The fallback matters just as much. If price data or weather is stale, the home should preserve defined reserve, respect device limits and use a conservative local policy until fresh inputs return.
This is also why forecast accuracy cannot be judged only by a weather metric. The real test is whether updated decisions improve cost, solar use and reserve compliance.
Verification
Measure the decision, not only the forecast.
- Net energy costImport cost minus export value under the actual contract.
- Solar curtailment or spillGeneration that could not be used, stored or exported.
- Battery throughputExtra cycling created to achieve the result.
- Reserve complianceWhether the protected minimum was preserved when required.
- Forecast errorPV and load error by the intervals that changed decisions.
The main conclusion remains: the best target is the lowest sensible charge that covers the expected deficit and reserve while leaving room for likely solar.
FAQ
Forecast charging questions.
What is forecast-based battery charging?
It is a control method that updates battery charging from expected solar production, household load, electricity prices, current state of charge and reserve requirements.
Should a home battery charge to 100% every night?
Not always. A full battery can be useful before a cloudy or high-price day, but it can leave no room for expected midday solar. The target should follow the household objective and forecast.
What happens when the solar forecast is wrong?
A robust system keeps an uncertainty margin, replans as new data arrives and falls back to safe limits when forecasts or prices are stale or unavailable.
Does forecast charging always save money?
No. The result depends on tariff rules, import and export prices, battery efficiency and wear, solar size, household demand, weather and the quality of control.
Can every battery charge from the grid?
No. Capability, warranty, installation settings, metering, tariff and local rules vary. The exact system and contract must allow grid charging.
Sources
Technical references.
- SMA Sunny Home Manager operating manual
Documents forecast-based energy management using expected PV, consumption, tariffs, battery state and export constraints.
- Tesla Powerwall Savings Mode user guide
A manufacturer example of continuously forecasting solar and consumption while balancing time-based rates and backup reserve.
- Australian Government: home batteries
Official consumer guidance on battery losses, export limits, tariff configuration and conditional economics.
- NREL BLAST-Lite battery lifetime model
Shows why battery lifetime modelling includes temperature, state of charge, depth of discharge and power, not only a cycle count.
- Assessment of forecasts in PV-battery systems
Research on how load and PV forecasting methods affect residential and commercial battery control under different constraints.