What are Ancillary (Balancing) Services and who can earn Money with them?

Balancing services are used to manage frequency fluctuations in power grids. These fluctuations occur when power supply and demand are not perfectly balanced.

The provision and feed-in of ancillary services are some of the so-called system services that transmission system operators (TSOs) purchase for grid operation. These services can be provided not only by power generators but also by battery energy storage systems and power consumers.

There are several terms related to balancing services. To make matters even more complicated, every country uses slightly different terms to describe very similar products which the respective national TSOs use to balance the grid. Here is a distinction of the three most important terms:

Alternating current (AC) power grids must always maintain a precise balance to ensure the system frequency remains steady at exactly 50 hertz (Hz). This is a delicate task, especially since the grid itself cannot store or buffer electricity. As a result, electricity supply and demand must always match in real time.

Even the slightest deviation from the target frequency triggers automatic countermeasures. If the frequency strays by as little as 0.01 Hz, balancing power is activated, prompting the delivery of balancing energy to stabilize the system.

A stable frequency is vital because many electrical devices depend on it. In electric motors, for example, the operating speed is directly tied to the frequency of the AC power. While a small fluctuation might go unnoticed when whipping cream, in industrial settings, even minor variations can lead to costly production errors.

To keep the frequency of the power grid constant, the amount of power fed into the system must always match demand at any given moment. This requires coordination between power producers and consumers. Of course, no one picks up the phone to call the nearest power plant when they're heating up a frozen pizza or brewing a cup of coffee. These everyday consumption patterns are forecasted using so-called standard load profiles. The reason this works is that individual deviations from typical usage tend to balance each other out - one person might make their pizza a bit earlier, another brews coffee a bit later.

Large-scale consumers—such as chemical parks, mining operations, energy-intensive manufacturing plants, as well as municipal utilities, data centers, EV charging stations, and mid-sized production facilities—alongside all major power producers, submit their consumption or generation forecasts directly to the TSOs.

Each of these players operates within what's known as a balancing group, which is managed by a designated balancing group manager called Balancing Responsible Party (BRP). Each manager is responsible for ensuring that their group’s power balanced:

Balancing groups must be settled every 15 minutes - specifically, 15 minutes before the start of each delivery or consumption period. To discourage imbalances, whether accidental or intentional, the responsible parties must bear the cost of the necessary balancing energy at what’s known as the imbalance price (which is mostly made up of the costs incurred for balancing energy). This is typically significantly more expensive than correcting the imbalance by purchasing power on the spot markets.

Despite significant efforts and incentives aimed at keeping balancing groups stable, most frequency fluctuations in the power grid occur when Balancing Responsible Parties are unable to maintain this balance. These fluctuations can go in either direction: if power supply exceeds demand, the grid frequency rises; if demand surpasses supply, the frequency falls.

For example, if weather forecasts turn out to be inaccurate and less renewable power is available than expected, there may not be enough time to purchase the shortfall from another producer. In such cases, generation lags behind consumption, causing the grid frequency to drop.

Conversely, if an unexpected disruption occurs - such as an industrial plant shutting down - this removes a major consumer from the grid. If the surplus power cannot be redirected quickly enough, the grid frequency rises.

In practice, it is nearly impossible to predict the exact output of renewable energy sources or electricity consumption. A certain degree of imbalance is therefore the norm. However, on a broader scale, positive and negative deviations tend to offset each other to a large extent, preventing serious destabilization of the grid.

Minor frequency fluctuations - within a few percent - do not pose a problem for the power grid itself. However, the grid infrastructure, including power lines, substations, and transformers, is designed to operate at a target frequency.

In the Continental European grid, a frequency range between 49.8 and 50.2 Hz is considered acceptable. Within this range, grid operators respond to deviations by activating standard balancing power. Balancing groups must be fully stabilized no later than one hour after a deviation occurs, allowing the use of balancing energy to be scaled back and returned to reserve power.

Across the Continental European power grid, frequency deviations from the so-called “deadband” (dt.: Totband) - between 49.99 Hz and 50.01 Hz - are an almost constant occurrence. This is the range within which no corrective measures are activated. In the vast majority of cases, however, ancillary services are sufficient to stabilize frequency levels before they drift outside the “balancing range” (dt.: Regelband) of 49.8 Hz to 50.2 Hz.

In rare instances, the available balancing power may not be enough to keep grid frequency stable. This can lead to an “uncontrolled brownout,” also known as a voltage sag. In such cases, transmission system operators initiate a load-shedding procedure. That means parts of the grid - ranging from large industrial facilities to entire neighborhoods or regions —-are temporarily and deliberately disconnected (a controlled voltage sag) in order to prevent a widespread and uncontrolled blackout.

If the grid frequency falls below 47.5 Hz despite all countermeasures, there is a risk of damage to both consumer equipment and grid infrastructure. In such cases, power supply in the affected area of the Continental European grid - occasionally in multiple regions at once - is briefly shut down entirely and then systematically restored.

Balancing services are categorized by the time intervals in which they respond, and by whether they are designed to compensate for a frequency surplus or deficit.

Grid frequency regulation is structured in several stages, activated sequentially if the previous stage fails to restore frequency within the deadband.

Grid frequency can deviate in two directions: upward, for instance when a large consumer disconnects or private solar installations - which are not centrally controlled - feed in more power than expected; or downward, when generators fail or consumption unexpectedly spikes. As a result, energy flows must also be able to act in both directions to correct these imbalances.

Since ancillary services are essential to maintaining the functionality of the electricity grid, the requirements for participating in the balancing power and balancing energy markets are strict and complex. Providers must undergo an extensive application process, submit a detailed concept, and complete a series of test runs - known as prequalification.

In principle, all units must be able to deliver a minimum of one megawatt (MW). However, it is also permissible to pool smaller installations - such as Battery Energy Storage Systems (BESS) - to meet this threshold. Moreover, different types of facilities are better suited to provide specific forms of ancillary services.

Grid operators favor speed-regulated power plants for the provision of primary reserve. Where available, pumped-storage power plants (PSPs) are often used. Their water turbines can be activated within seconds, with many reaching full output in under 30 seconds.

Other power plants equipped with synchronous generators are also suitable. Thermoelectric plants - such as those running on gas or biogas - can provide primary balancing energy as well, although they operate more efficiently when running at a constant output.

Since 2014, BESS have increasingly been deployed for primary reserve. They are capable of storing and releasing power at full charging and discharging capacity almost instantaneously. Importantly, a provider’s ancillary services do not have to come from a single facility or battery system. Today, even micro-installations with less than 10 kW are permitted, provided they are grouped into so-called virtual power plants.

Starting with secondary ancillary services, systems that provide either only positive or only negative balancing power can also participate in the market. These systems are also allowed more time to respond, opening the door to a wider range of service providers beyond PSP, BESS, and virtual power plants.

In addition to PSP and BESS, positive secondary reserve (or: upward aFRR) is frequently provided by gas or biogas-powered gas turbines or steam power plants. While these facilities typically take longer to ramp up, they are often able to deliver higher output than PSPs and battieries. A general requirement for participation in secondary reserve is a minimum output of 5 MW. Smaller units are permitted only under specific conditions.

Secondary reserve can also be provided by pure power consumers. Many large industrial facilities can adjust their consumption with considerable flexibility - particularly where power is used to generate heat or cold. When called upon, these facilities reduce their power consumption for 15 minutes. Thanks to thermal inertia, this can be done without significantly disrupting operations. Large-scale heat pumps used in district heating networks are also potential candidates.

PSP and BESS are also well-suited for negative secondary reserve (or: downward aFRR), as operators seek to recharge them as cost-effectively as possible. Thermoelectric power plants can contribute negative balancing energy by reducing their output. Industrial facilities and heat pumps, in turn, can provide negative balancing power by increasing their power consumption.

The same types of facilities used in secondary reserve can also participate in tertiary reserve, both positive and negative (or: upward and downward mFRR). However, PSPs and batteries do not always have sufficient energy reserves to meet the requirement of delivering their full balancing power over a full hour. For these systems, it can sometimes be more profitable to offer four times the balancing power over a 15-minute interval instead.

In contrast, coal-fired power plants can be used here, as they are often too slow to react for primary or secondary reserve. Many coal units can generate over 100 MW, which allows them to make the necessary adjustments by varying their output by only a few percentage points.

PSPs and BESS can, of course, be charged with renewable power - and discharge it accordingly. Biomass or biogas plants, as well as run-of-river hydroelectric plants, are also suitable for secondary and tertiary reserves and routinely take part in the market.

Since 2020, wind farms have been able to prequalify for all levels of ancillary services, both positive and negative. As of spring 2025, however, they are still primarily used in tertiary ancillary services - largely because the economic incentives remain weak. Voluntarily shutting down wind turbines results in lost feed-in revenues, which the premiums for ancillary services do not always offset. With appropriate monitoring and control technology, solar power plants could also participate in the ancillary markets, though for obvious reasons, only during clearly defined periods.

The various levels - or products - of balancing services are coordinated differently. The earlier a reserve is activated in the sequence, the more automated its operation becomes.

The primary reserve is coordinated by ENTSO-E, the European association of transmission system operators, and operates across national and grid boundaries. However, each national TSO remains responsible for ensuring a sufficient supply of ancillary services within its jurisdiction.

Unlike other reserves, FCR is not manually activated by grid operators. Instead, system service providers autonomously respond to frequency deviations. Their equipment continuously monitors grid frequency and automatically compensates for fluctuations as soon as the frequency moves outside the designated deadband. Once normal conditions are restored, the systems deactivate without external input.

When a need arises, the responsible TSO transmits the demand for secondary reserve in real time to dispatch platforms. These platforms then activate service providers within the grid area and, when available, utilize cross-border transmission paths to meet the requirement. Activation of aFRR follows a merit-order system, starting with the lowest-cost bids and progressing upward until the demand is covered. As the secondary reserve ramps up, the primary reserve naturally winds down in response to the stabilizing frequency.

If it becomes apparent that FCR and aFRR will not be sufficient to offset frequency fluctuations, the TSO manually notifies service providers - through digital platforms and in some cases still by phone - again based on the merit order. These providers must then immediately initiate the delivery of minute reserve (mFRR). This reserve must reach its full balancing effect, whether upward or downward, within 15 minutes of activation.

Ancillary services in general and balancing services in particular are not traded on the power exchange. Instead, TSOs organize separate auctions for each of the three market segments of balancing services. Every day, they tender the 24 hours of the following day in six blocks of four hours each.

Bids are awarded according to the merit order principle - ranked in ascending price until demand is met. This process applies to primary reserve, secondary reserve, and minute reserve alike. However, there are notable differences between the individual market segments.

A key feature of FCR is that providers cannot submit separate bids for positive and negative energy - they must offer both together. As a result, both directions are jointly tendered.

Consequently, only balancing power is remunerated in this market segment. Providers are paid solely for keeping their systems on standby to deliver balancing energy if needed. They do not receive compensation for the actual amount of power injected into or withdrawn from the grid.

Just like day-ahead power auctions, providers submit their bids in a pay-as-clear auction, which is conducted by the respective TSO. Starting with the lowest offer, bids are accepted in ascending order according to the merit order. As with standard power trading, uniform pricing applies - meaning all successful bidders are paid the price of the highest accepted bid.

Prices for balancing power in the secondary reserve and minute reserve segments are also determined through auctions held by the TSOs. However, unlike the primary reserve, the pay-as-bid principle applies here: successful bidders are paid exactly the amount they offered in the auction. Additionally, auctions for positive and negative balancing energy are conducted separately. Balancing power and balancing energy are also tendered independently.

The separation of positive and negative ancillary services offer an economic advantage, as service providers can tailor their bids with greater precision. Two examples illustrate this:

Similarly, the operator of a virtual power plant composed of battery storage systems might submit a highly competitive bid for negative ancillary services between 4 PM and 8 PM on a cloudy day. If the storage systems were not charged earlier due to low solar production, then a call for negative ancillary services would provide a welcome opportunity to charge the batteries. Later, between 8 PM and midnight, the operator could offer a favorable bid for positive ancillary services, since the storage would be full and balancing energy typically commands higher prices than exchange-traded power.

A key distinction from the primary reserve lies in the separate tendering processes for balancing energy. Auctions are first held for the six time slots of balancing power, where service providers commit to holding reserve available. This ensures that the TSOs can rely on sufficient balancing services being accessible when needed.

Following this, the auction for balancing energy takes place. Providers who secured a contract for balancing power are obligated to submit a bid here as well. However, providers who did not win in the initial power auction are also permitted to participate. This mechanism is designed to maintain strong competition during the energy tendering phase.

When balancing energy is actually needed, providers with the lowest bids are called upon first, regardless of whether they were successful in the earlier power auction. This allows a provider who was not awarded a balancing power contract to displace a winning bidder in the merit order for balancing energy.

The cost of ancillary services in general and balancing power in particular is usually covered through grid fees, which power consumers pay via their utility bills. In contrast, balancing energy is billed monthly to balancing group managers (BRPs). TSOs calculate the average price of imbalance energy - essentially the cost of the activated balancing energy - at 15 minutes intervals and multiply it by the volume attributed to each balancing responsible party.

In Germany, the regulation of balancing energy markets falls under the authority of the Federal Network Agency (Bundesnetzagentur), which oversees fair market access and the stability of electricity supply. At the European level, overarching oversight and coordination are the responsibility of the Agency for the Cooperation of Energy Regulators (ACER). Together with national regulatory authorities, ACER ensures that the internal energy market — including balancing services — functions efficiently, transparently, and without discrimination.

A key operational role is played by ENTSO-E, the European Network of Transmission System Operators for Electricity. ENTSO-E coordinates the technical and organizational framework of the electricity market and is therefore instrumental in developing shared platforms for the cross-border exchange of balancing energy as well. Two flagship initiatives in this context are PICASSO (Platform for the International Coordination of Automated Frequency Restoration and Stable System Operation), which handles automatically activated frequency restoration reserves (aFRR), and MARI (Manually Activated Reserves Initiative), which coordinates manually activated frequency restoration reserves (mFRR). These platforms aim to harmonize the balancing energy market across Europe, facilitate the cross-border use of flexibility resources, and enhance system stability in a more cost-effective and efficient way.

PICASSO is the European platform for the exchange of automatically activated frequency restoration reserves (aFRR), developed by ENTSO-E in collaboration with national transmission system operators (TSOs). Technically, PICASSO relies on a central algorithm that evaluates balancing bids from market participants across countries every four seconds and allocates aFRR capacity based on need and cost-efficiency. Participants submit standardized bids into a common market platform, specifying both capacity and price. The activation algorithm prioritizes the most cost-effective bids, regardless of national borders, provided there are no grid constraints. Since the platform's official launch in July 2022, several countries — including Germany, the Netherlands, Austria, Slovenia, and the Czech Republic — have joined. However, not all European TSOs are yet technically integrated, meaning that the platform currently taps into only part of its full European potential.

MARI (Manually Activated Reserves Initiative), by contrast, is the central European platform for the exchange of manually activated frequency restoration reserves (mFRR), established under the EU’s Electricity Balancing Guideline. Unlike PICASSO, which operates automatically, MARI is based on market-based activation by TSOs in response to concrete system needs. As in PICASSO, market participants submit standardized bids into a shared IT system, indicating price, available capacity, and activation time. A central algorithm then determines, on a minute-by-minute basis, the most cost-efficient combination of available bids and allocates them accordingly — including across borders, as long as the grid can support it. Since its launch at the end of 2022, MARI has gradually become operational, with more than ten countries now either actively participating or in the final stages of implementation.

A major technical challenge for both platforms lies in the complexity of real-time communication and IT interfaces between national systems and the central platform. In addition, grid constraints and differing regulatory frameworks across EU member states continue to slow full market integration. The harmonization of lead times, product definitions, and pricing mechanisms is also an ongoing process. Nevertheless, both PICASSO and MARI are seen as milestones on the path to a fully integrated European balancing energy market — one that enhances efficiency, strengthens security of supply, and better integrates renewable energy sources.