What is FerroAI and how do I use the function?

FerroAI was introduced on 25-04-22. 

During the first few weeks, the software will be updated frequently to gradually improve the user experience. Feel free to send feedback through the user app (under Account Settings).

For release notes and updates see the following link:

FerroAI Releaser

What is FerroAI?

FerroAI optimizes the use of your Ferroamp system. By analyzing consumption patterns, production data, weather conditions, and electricity prices — including grid tariffs — FerroAI makes daily, data-driven decisions to maximize the economic benefit of your system.

Activating FerroAI

1. Download or update the Ferroamp user app on your mobile phone.
2. Select the “Control” tab in the bottom icon menu (three lines with circles).
3. Expand the menu in the upper left corner and select FerroAI.
4. Go to Electricity Settings and select “Electricity Area.” Enter the area your installation is located in.
5. Select “Tariff Type.” Indicate whether you pay power-based tariffs to your electricity grid provider. Choose “Reduced tariff at night” if you have lower or no tariffs during the night.

Instructional video for activation

Enter Grid Tariff

For FerroAI to manage your power peaks correctly, you need to specify whether you are charged for peak power (demand charges) by your grid operator and how those charges are structured. This information is usually found on your electricity network invoice or your grid operator’s website. There are three options:

• No demand charges

• Same demand charges all day

• Different demand charges for night and day (you’ll need to specify the time periods for reduced/no charges)

Deactivation

You can deactivate FerroAI by changing the mode under the “Control” tab in the app. When switching back to manual control, the latest settings will be reactivated.

Control

FerroAI combines data on spot prices, grid area, tariffs, weather, and solar radiation with your electricity usage history from EnergyHub to minimize electricity costs.

Every evening, just before midnight, an analysis is performed for the coming day.

The calculation is based on:

• Forecasts of energy consumption based on your historical usage data and time of year

• Expected solar production based on weather forecasts and historical weather data

• Grid costs based on power demand, taxes, and transmission fees

• Spot price data from the electricity market

The goal is for the battery to be used in the way that provides the greatest financial benefit.

Charging the battery

1. Solar power
   FerroAI prioritizes charging the battery with surplus solar production. If solar power is expected during the day, the battery may refrain from charging from the grid even if electricity is cheap. When there is a lot of sun, the battery will prioritize charging during the hours when spot prices are lowest, so that you are paid as much as possible for any export.
2. From the power grid
   If solar production is not expected to cover the need, the battery will charge from the grid when electricity prices are low. If you pay power-based fees, FerroAI spreads charging over several cheap hours rather than maximizing charging when prices are at their lowest, to avoid creating power peaks.
3. Only when economically beneficial
   Charging from the grid only occurs if FerroAI calculates that it will be financially beneficial.

Discharging the battery

To create financial benefit, FerroAI prioritizes discharging in the following order:

1. Reduce power peaks (if you have a power-based fee)
   FerroAI reserves battery capacity to reduce expected power peaks based on historical electricity usage and seasonal variation.

2. Maximize self-consumption of solar power
   It is more profitable to use self-produced electricity than to sell it. In addition to the electricity price itself, you avoid transmission fees, electricity tax, and VAT.

3. Avoid price peaks
   If there is more capacity in the battery than needed to reduce power peaks, the battery will prioritize discharging when electricity is most expensive.

4. Sell to the grid
   Discharging for sale only occurs if the price difference between buying and selling is calculated to exceed transmission fees, electricity tax, VAT, and energy losses. Otherwise, the energy is saved for later use in the property.

Real-life Examples

Example 1 – Demand peaks expected, no solar surplus:
FerroAI charges the battery during cheap night hours and sets it up to shave demand peaks. Any remaining energy is used to reduce grid imports during expensive evening hours.

Example 2 – Demand peaks and solar surplus expected:
The battery is prepared to handle peaks but also discharges to make room for expected solar surplus. Remaining energy is used during the evening to reduce imports and maximize self-consumption.

Example 3 – No peaks, no solar surplus expected:
FerroAI charges the battery during cheap night hours. The energy is used for self-consumption during expensive morning and evening hours.

Example 4 – No peaks, solar surplus expected:
Discharges the battery in the morning to make room for solar surplus. Energy is also used to reduce grid imports during expensive hours.

Forecasting

FerroAI’s consumption forecasts are tailored to each installation. The model is trained on historical data and identifies patterns such as time-based variations (daily, weekly, monthly), average daily consumption, and recurring power peaks. The system also estimates how much peak reduction can be achieved each month and adjusts over the year.

Video explaining the graphs:

FAQs

When can FerroAI be activated?
If you already have a system, FerroAI can be activated and begin controlling the system within a few hours. If you have a brand-new system, you may have to wait up to a week until FerroAI has enough data to make its calculations.

Why didn’t the battery charge at all during the night?
This may be because the spot price differences during the day are so small that FerroAI does not expect it to be profitable to charge the battery at night considering fees, costs, and energy losses.

Why did the battery charge so slowly during the night and not all the way to 100%?
This may be because there is no need for that much battery capacity for the discharge planned during the day.

In other cases, it may be because other loads, such as a heat pump or EV charging, were running at the same time as charging. In that case, there may not have been enough room under the calculated power limit or main fuse rating to charge the battery at the planned power level. If the pattern repeats, FerroAI will adjust charging after a few days according to the new conditions.

Why didn’t FerroAI sell electricity to the grid during the price peak?
This may be because FerroAI calculates that energy will be needed from the battery for upcoming power peaks or for use within the property even after the price peak, when taxes and fees on purchased electricity are expected to exceed the price difference.

Does FerroAI work without a battery?
No. Initially, it only works if a battery is present in the system.

Does it work with retro-fit systems (no SSO, solar connected to another inverter)?
Yes. FerroAI can’t read solar production directly without connected SSOs, but it observes exported surplus and uses that for future forecasts.

Does FerroAI support large systems with multiple batteries?
Yes, it works on large systems with several batteries.

Does it work in PowerShare (multisystem)?
No, FerroAI does not work in PowerShare setups.

What does FerroAI cost?
FerroAI is free in the initial phase and will later become a subscription service. Pricing has not been determined yet, but the value will clearly exceed the cost.

Can I control when the battery discharges for selling electricity?
No, individual variables can’t be controlled when FerroAI is active. It optimizes based on overall economic gain. If there's surplus and the battery is full, electricity is sold — unless the price is negative.

Does FerroAI prioritize peak shaving over cheap electricity?
Yes. Since power peaks often incur higher costs than price fluctuations, FerroAI prioritizes saving enough capacity for peak shaving.

Does it consider time-of-use tariffs (day/night pricing)?
Yes. You can choose between three modes: no tariff, same tariff all day, or different tariffs by time.

Can I set a price threshold regarding when to use the battery?
No, FerroAI makes automatic decisions and avoids using the battery for minor price differences.

Can FerroAI interact with external systems like HomeAssistant or Homey?
To ensure optimal performance of FerroAI, it is recommended that no external systems are used for control. However, systems such as Homey and Home Assistant can still be used for data collection and visualization without affecting FerroAI.

What happens if FerroAI is turned off?
The most recent active setting is restored automatically, ensuring continued operation without AI control.

How does FerroAI handle integrated EV charging?
FerroAI currently does not account for OCPP-based control. If you want to avoid peak charges caused by EV charging, limit the charging power and schedule it in the car.

How does it work with non-integrated EV chargers? Will it recognize charging patterns?
Charging without OCPP can’t be controlled, but regular usage patterns are detected and included in forecasts like any other household load.

How does FerroAI affect connection to ancillary services (grid services)?
This depends on your grid service provider's requirements. Currently, FerroAI cannot be combined with ancillary services from Varberg Energi. If FerroAI is enabled, your system may not be available for their services.

Can FerroAI’s calculations be wrong?
Yes, FerroAI is based on forecasts that can be affected by weather, solar radiation conditions, and major changes in electricity usage patterns.

If the forecasts are wrong, the calculations will also be wrong.

Example:
Solar power surplus is lower than estimated. This means the battery does not charge according to the forecast in the app and therefore does not have enough energy to meet the needs.

Possible explanations:

• An incorrect weather forecast

• Solar radiation is limited by shadows, leaves, or snow on the panels

• Higher consumption than forecast resulted in lower solar surplus than expected

In the case of longer deviations, such as snow on the panels or changed consumption, FerroAI adjusts its calculations within a few days. New systems or changed electricity usage also mean a learning period before optimization becomes fully accurate.

Errors may also occur due to bugs or conditions not included in the calculation models. Ferroamp continuously works to improve FerroAI and minimize these. Although temporary deviations may occur, the most important factor is long-term benefit. The system is designed to provide the best results over time so that total electricity costs are reduced each month.