Danish imbalance prices: a more volatile future

Denmark has long been one of the leaders in the energy transition in Europe, with more than 7 GW of installed wind power and 3.6 GW of solar. This means that more than 50% of its energy consumption can be met by renewables during certain parts of the year.

At the same time, there are hours without significant wind or solar production. When this happens, the country (as all countries do) relies on conventional generation and imports through interconnectors to meet power demand. The secure operation of the power system, at all times, is one of the prime responsibilities of the Transmission System Operators (TSOs).

To facilitate this, the supply and demand of electricity needs to be balanced, leading to what is known as nominal frequency in the system. This means that supply and demand are matched. The planning and dispatch of supply and demand is coordinated by the TSO, with the help of various electricity markets, that are set up before the delivery of electricity.

The focus of this article is primarily on the balancing markets which are cleared very close to the real-time delivery. There have been some major market changes in the past few months and some more coming up quite soon.

In Europe, the electricity markets are designed with zonal pricing, which mean that all the market actors in a given zone are subjected to the same price. Denmark has two bidding zones, DK1 and DK2. It is a bit special in this sense that DK1 belongs to the continental synchronous area and DK2 is a part of the Nordic synchronous area.

As a result, DK2 participates in some common frequency services markets with the Nordic bidding zones, including Frequency Containment Reserve-Normal (FCR-N), Frequency Containment Reserve-Disturbance (FCR-D), automatic Frequency Restoration Reserve (aFRR) capacity market. Conversely, DK1 operates in a continental European FCR market. DK1 also launched a local aFRR capacity market in October 2024, just before joining PICASSO (Platform for the International Coordination of Automated Frequency Restoration and Stable System Operation).

Denmark’s integration to the PICASSO balancing initiative:

PICASSO is a collaborative European initiative aimed at creating a unified market platform for the activation of automatic Frequency Restoration Reserves (aFRR). Denmark joined Germany, Austria and Czechia in this common platform on 2 October 2024. On 17 October 2024, the Dutch TSO, TenneT also joined the project. Thereafter, Belgium, Bulgaria and Slovakia joined PICASSO.

The concept behind PICASSO is to allow TSOs to share balancing reserves. This is becoming increasingly important as more intermittent renewable generation connects to power networks. Efficient allocation of aFRR reserves is enabled through a common optimisation model which is cleared every four seconds. The market itself is cleared based on marginal clearing mechanism. Having a common PICASSO platform helps to increase transparency and competition among balancing service providers.

Since Denmark joined, we have been observing the price formations and volumes exchanged in this TSO-TSO market. Both bidding zones in Denmark have been subjected to high PICASSO cross-border marginal prices from time-to-time.

A good example comes from 2nd February 2025, when a 45 MW upward activation of aFRR ocurred in in DK2. This resulted in a price of 2,100 EUR/MWh. Looking at the fuel mix in Fig. 1, you can see how a deficit in renewable generation led to this scenario. A lack of wind generation in particular made the local DK2 PICASSO merit-order more extreme. This is also made worse as DK2 is a relatively small bidding zone, including a limited number of assets available to provide upward aFRR reserves.

As DK2 is also a part of the common Nordic aFRR capacity market, it has become lucrative for assets in this area to offer their capacities in the common market and get compensated in the PICASSO market with the pay-as-cleared price.

Even though DK1 is a bigger bidding zone than DK2, it is not shielded from such spikes. On 12 February 2025, we observed a situation in PICASSO where both DK1 and DK2 were coupled and Cross Border Marginal Pricing (CBMP) jumped to 4,618 EUR/MWh in the morning for a few minutes (as shown in Fig. 2). The total aFRR up activation needed for DK1 and DK2 combined was around 140 MW. The entire DK1 upward resources of 95 MW (which were available for around 2,000 EUR/MWh) were exhausted, so the remaining deficit had to be solved by DK2 - causing the price to spike to 4,618 EUR/MWh for both the bidding zones.

Another case from the evening on 27 February shows the PICASSO price for DK1 soaring beyond EUR 4,000/MWh, while DK2 was priced the lowest amongst all the PICASSO bidding zones at EUR 78/MWh. In this situation, the wind generation in DK1 was significantly low. Moreover, there was a lack of liquidity in the aFRR up market. This put the already short market in an even more extreme position, not to mention that the area was isolated from others in real-time when the market was cleared.

These examples of high prices Danish bidding zones show that PICASSO prices can get quite extreme based on both market liquidity and the availability of interconnector capacities in real-time.

PICASSO activation volumes and prices in Denmark:

The scatter plots of activated aFRR volumes and prices in DK1 and DK2 are shown in Fig. 4. By comparing the plots for DK1 and DK2, we can observe that there are larger activations on both sides in DK1. We can also see that the upward activation prices for the same activation volumes could be much higher in DK2 than in DK1. The same is true for the downward activation prices in DK2, which are higher than in DK1. For smaller volumes, even less than 10 MW, the aFRR down prices have reached peaks of around 2,000 EUR/MWh on several occasions. 

Impact assessment of aFRR on imbalance prices 

The imbalance price calculations in Denmark so far do not include the PICASSO prices. Therefore, imbalance price spikes are not yet coherent with PICASSO spikes.

To understand what this means for the future, we can compare the mFRR regulation prices that set the imbalance prices as of today, with the aFRR energy prices which will be decisive in the imbalance price calculation from 18th March 2025. Remember, the aFRR energy market in both the bidding zones has limited volumes and the activation prices can get very extreme, very quickly.

Fig. 5 illustrates the aFRR and mFRR up and down regulation prices in DK1. Fig. 6 also shows this for DK2. The aFRR down regulation prices have been both volatile and extreme just after the introduction of this market. However, the aFRR up regulation prices for both areas have stayed high in the past few months. This is partly due to the seasonal effect, where there has been an extended period of low wind production, but also because this is still a new market with limited liquidity.

Impacts of the change to Danish balancing calculations:

In this section, we analyse the simulated new imbalance prices that would have occurred if the new imbalance pricing method was implemented when Denmark joined PICASSO. We also dive into how the aFRR prices from PICASSO influence the simulated imbalance prices.

At the moment, the imbalance prices are solely based on the mFRR prices. Danish TSO, Energinet, is compensating for the mismatch between the imbalance costs and Balancing Responsible Party (BRP) payments due to PICASSO prices not being reflected in the imbalance prices. Unfortunately, this removes the polluter pays principle - meaning those who cause the imbalance are not responsible for the costs associated with it.

To help solve this, the new imbalance pricing mechanism will be implemented from 18th March 2025. In this new method, the dominating direction is decided based on mFRR activations. Whether the aFRR activation prices influence the imbalance prices depends on the direction of aFRR activations. If these two directions overlap, then the aFRR prices are considered in the calculations. For upward aFRR activations, the volume weighted average price of aFRR for the 15 minute interval is compared with the mFRR upward activation price. The higher of those two figures is then assigned as the imbalance price. The works the opposite way for downward activations with the lowest prices taken into account.

Fig. 7 and Fig. 8 show both the aFRR up regulation prices and the associated imbalance prices under the new imbalance methodology for DK1 and DK2 respectively. It also marks those occasions when aFRR up prices are equal to imbalance prices - indicating that aFRR activations were aligned with the dominating direction determined by mFRR. For DK1, in the upward direction, this happened 13% of the time. In DK2, it happened 11% of the time.

Fig. 9 and Fig. 10 illustrates the aFRR down activation prices and imbalance prices with the new method in DK1 and DK2 respectively.

Down regulation is usually activated where there is an excess of energy in the power system. Since the period in this study mainly covers the winter months (and periods of reduced wind generation) we have seen fewer downward activations. aFRR downward activations overlapped with the dominating direction 24% of the time in both DK1 and DK2. It is also important to note that there can be scenarios where the system is in upward mFRR regulation due to deficit of supply, this becomes overcompensated, and then needs slight downward regulation from aFRR for the final balancing adjustments. This is due to the fact that aFRR is used to fine tune balancing requirements.

Volatility measure of the new imbalance prices:

The imbalance prices in both DK1 and DK2 can be more volatile compared to other Nordic bidding zones, even with the old imbalance price calculation method (without PICASSO prices). As the Danish bidding zones are smaller relative to the other Nordic zones, and have less diversity in sources of generation, the real-time availability of interconnector capacities can be influential in determining the imbalance prices.

Box plots are presented in Fig. 11 for the differences between day-ahead prices and both the old and new imbalance prices. This illustration helps to measure volatility of the imbalance prices. These figures show that the volatility with the new imbalance prices would have been higher than the old imbalance prices for both the areas. Again, it is important to highlight that these observations are based on four months of data without a sufficient change in market signals for the actors.   

Automation in the mFRR activations:

The manual frequency restoration reserve (mFRR) market will shift to automated activation on 4th March 2025 across all Nordic countries. Balancing will take place at bidding zone level. The imbalance prices will be determined based on the request signal of a bidding zone, instead of the activations in that bidding zone. Due to this new algorithmic activation, we can also expect to see more volatility in the imbalance prices as per the Nordic TSOs.

Conclusion

While the change to Danish imbalance calculations will help with correctly indicating the cost of balancing the grid to the market actors, it will also subject them to more volatile prices.

With the wider Nordic shift to automated mFRR activation from 4th March 2025 and new imbalance pricing method being applied, it will be interesting to observe the superimposed effect of the two new market changes. 

The resolution of the intraday markets will also change to 15 minutes, making it more granular. As the expectation of imbalance prices plays an important role in determining intraday market strategies, it will increase the need for advanced market analysis and forecasting to ensure successful trading strategies.