Battery storage in a greenhouse: peaks or reserves?

Both, if the capacity allows: battery storage in a greenhouse shaves the demand charge peak and can take part in the reserve markets at the same time, as long as discharge power and energy are deliberately divided between the revenue streams. Sizing decides which stream comes first and how much room is left for the other.

Which three revenue streams are open to the battery?

A greenhouse suits a storage system unusually well, because lighting, heating and ventilation create large and predictable power peaks. The same battery can earn in three ways.

1. A lower demand charge. The network operator measures power in 15 minute periods, and in demand tariffs the charge is typically set by the highest 15 minute average power of the month. The calculation basis varies between network operators, so we check it in the operator’s own price list. The metering period and the calculation basis above apply to Finnish network operators; for a site in Sweden we check the network operator’s demand tariff and metering period separately. The battery discharges across exactly the period where lighting and the other loads coincide. We wrote about finding the peaks in more detail in demand charge optimisation.
2. The reserve markets. Fingrid pays for the readiness to adjust power, and a battery is the fastest resource for that. The products differ in activation speed and in the duration required.
3. Shifting own or cheap power. Solar generation lands in the middle of the day, the need for lighting in the morning and the evening. The battery moves the day’s generation, or power bought during a cheap hour, to the hours when the lights are on.

All three use the same discharge power and the same energy. That is why the order has to be settled before the size of the battery is fixed.

The advantage a greenhouse has over other sites is predictability. Lighting runs on the clock and on the light sum, so the timing of the peak is known in advance to the day and often to the quarter hour. That turns both peak shaving and reserve bid planning into calculation. In the example greenhouse it looks like this: lighting starts before the heating night setback has recovered, the overlap lasts about two hours, and the battery’s 350 kW discharge is timed to exactly that window.

How is battery storage in a greenhouse sized for peak shaving?

Sizing starts from metering data, not from a supplier’s product range. The example greenhouse has its own input data: the highest 15 minute average power of the month is 1,450 kW, the target level is set at 1,100 kW, and the peak forms in the morning when lighting starts before the heating night setback has recovered. The overlap lasts about two hours.

The calculation runs step by step.

1. Power to be shaved. 1,450 kW less 1,100 kW is 350 kW.
2. Energy to be discharged. 350 kW times two hours is 700 kWh.
3. Nominal capacity. Assuming 80 per cent depth of discharge, 700 kWh divided by 0.8 is 875 kWh.
4. Discharge power. The battery has to deliver 350 kW for the full two hours, not just for a moment.

The target level is not a free choice but is derived from data. If the month has only one period at 1,450 kW and the rest stay below 1,100 kW, a target level of 1,100 kW is realistic. If instead there are dozens of periods at different heights, the same target level requires the battery to discharge several times a day, and then the time charging takes becomes the binding constraint. The battery has to be charged before the peak, so charging is timed to night or midday hours.

If the same peak lasted three hours instead of two, the energy discharged would rise to 1,050 kWh at the same 350 kW of discharge power. Duration decides the capacity and the height of the peak decides the power. That difference shows up only in real data at 15 minute resolution, because sizing based on an estimate tends to land wrong precisely on duration.

What do the reserve markets require from the battery?

The reserve products differ in how fast and for how long the power has to be available. The fastest activate automatically on frequency, the slowest at Fingrid’s request.

Activation times and durations: Fingrid’s product pages FFR and FCR. The figures in the table are Fingrid’s and apply in Finland. For a site in Sweden we check the transmission system operator’s product requirements case by case.

In a greenhouse battery the requirements translate straight into kilowatt hours. At the example’s 350 kW, FFR ties up about 2.9 kWh over a 30 second duration, while the FCR-D requirement of at least 20 minutes ties up about 117 kWh during the hours the battery is on the market. The fast products therefore buy power, the slow ones energy. The products also have a minimum bid size and a prequalification process, so a greenhouse takes part in practice through an aggregator that combines several sites into one bid. Down regulation reverses the requirement: the battery needs room to take in power, so the state of charge cannot be kept full. Technical terms and compensation levels change, so the product is chosen on the characteristics of the load and the battery; we wrote about the products more broadly in demand response and reserve markets.

Why do solar power and cheap hours compete for the same battery?

Lighting is weighted towards the dark season, when there is no own solar power. The battery shifts midday generation to the lighting hours, and in winter the same move is made with power from a cheap hour. In the example greenhouse that means charging when generation or price is favourable and discharging when the lights are on.

This is where the conflict with the other two revenue streams appears. A fully charged battery cannot offer down regulation, because there is no room. A battery emptied for morning lighting no longer shaves the peak if a second peak arrives the same morning. And if the example’s 117 kWh of dischargeable energy is reserved for FCR-D, the energy available for shifting shrinks by exactly that amount (at 80 per cent depth of discharge, about 146 kWh of nominal capacity).

The answer is not to pick one stream and drop the others, but to agree their order hour by hour. During the morning peak period the demand charge comes first, at midday the battery can sit in reserve, and evening lighting uses the energy that is not tied up elsewhere.

How are the revenue streams combined in one battery?

Combining them is settled in the control logic and in the contracts. The order we at Iiva work through:

1. Reading the data. A year of metering data at 15 minute resolution, the lighting control programme and the generation profile of any solar system.
2. Sizing the peak. Target level, power to be shaved and duration of the peak as described above.
3. Size of the reserve allocation. How many kilowatts and kilowatt hours are left to the reserve and during which hours.
4. Control hierarchy. Which signal wins when the demand charge limit and a reserve call arrive in the same moment.
5. Contracts. The aggregator’s terms, battery suppliers’ offers and the network operator’s basis for calculating the demand charge side by side.

The fourth point is the one most often left unagreed. If the aggregator may control the battery freely, a reserve call can empty the charge just before the morning peak, and the month’s demand charge is set by that period. So the condition goes into the contract: protecting the demand charge takes priority over the reserve bid during the hours the data marks as peak risk.

We at Iiva do not sell batteries and do not act as an aggregator, so we calculate the options on the same basis and also say when a storage system is not justified. Demand charges, demand response and battery storage are calculated from the same data, because they compete for the same capacity.

The first step is light. You send a year of metering data at 15 minute resolution and the lighting control programme, and you get back the sizing of power and capacity plus an assessment of which revenue streams fit in the same battery.

---

We size the battery with the greenhouse’s actual metering data and compare the offers on the same calculation basis. Read about battery storage or about demand response and reserve markets.