The increasingly integrated renewable energy sources have a strong intermittency in electricity supply, resulting in increasing mismatches between electricity supply and demand. This has led to an increase in imbalances in the electricity grid, heightening the need for regulating power, such as flexible assets. Currently, the balancing markets are experiencing a high volatility in system imbalance and imbalance prices, resulting in market and system instability.

Previous studies have largely focused on how renewable energy sources lead to an increase in system imbalance and have stressed the need for flexible assets. Furthermore, several studies show that flexible assets can make significant profits in today's balancing markets by optimising their implicit balancing strategy. However, little attention has been devoted to the price incentives provided by the different pricing designs in the EU and how their remuneration structure affects the implicit balancing behaviour of flexible assets. As flexible assets are increasingly deployed on the balancing market, any strategic or gaming behaviour might have a significant negative impact on grid stability and system costs. Therefore, this research aims to answer the following research question:

What implicit balancing behaviour do flexible assets show in different imbalance settlement designs and what balancing market design recommendations can be provided based on this?

This research question is explored using mainly a quantitative approach. First, a literature review on European balancing markets is provided, explaining how the EU Balancing Guideline forms a framework for the balancing markets in EU. Furthermore, the balancing markets in the Netherlands and Belgium, which serve as case studies in the modelling, are explored in detail. Then, two linear optimisation models are developed to simulate the balancing markets of both countries and to show how the price incentives provided by the markets influence the behaviour of a flexible asset that aims to optimise its profit. Additionally, the impact of the asset's implicit balancing, following three different strategies, on imbalance prices and the asset's own profit is examined. A comparison is made between the results under the two different pricing designs, showing how different design choices affect the opportunities for strategic behaviour and potential market exploitation. The main insights obtained can be summarised as follows:

- A flexible asset can, under certain circumstances, exploit the market with gaming behaviour, in which it intentionally amplifies the imbalance price and thereby enlarges its own profit.
- Marginal pricing, provides a larger opportunity window for strategic and gaming behaviour compared to the averaging of marginal prices. Establishing the imbalance price based on the average of all marginal prices during an ISP can mitigate the asset's effect on the imbalance price, but can eventually not prevent extreme imbalance prices either.
- There are advantages and disadvantages to both single and dual pricing. Single pricing provides a more stable pricing system, but is limited in discouraging overreactive implicit balancing. The dual pricing design discourages overreactive balancing stronger, but this can lead to great losses for BRPs, increasing their financial risks, and can decrease market stability.
- The volume of the system imbalance plays an important role; in ISP with lower system imbalances, flexible assets can only increase the imbalance price to a very limited extend, before a market saturation point is reached where all system imbalance is balanced implicitly. With higher system imbalances, flexible assets can push the imbalance price to much higher, or even extreme values.

As balancing markets evolve with increasing volatility in system imbalances and higher volumes of flexible assets, transaction costs are expected to rise due to higher levels of market participation, growing complexity in forecasting, and potential market saturation effects in quarter-hours with lower system imbalances. These expectations for future conditions emphasise the need for pricing designs that ensure grid stability and market efficiency. Therefore, this research provides several recommendations:


- TSOs should consider re-evaluating the overall pricing design for FRR and imbalance. These two pricing designs are highly interconnected, as imbalance prices should reflect the costs made for FRR. Marginal pricing has both advantages and disadvantages, as it can reduce strategic bidding for FRR, but can reward gaming behaviour in implicit balancing, leading to market instability. A different pricing design might provide a more stable market, where strategic implicit balancing is discouraged.
- The choice between single and dual pricing needs to be carefully considered. Based on this research's findings, single pricing leads to less risks for BRPs and contributes to market stability. However, additional research into the decision-making process of flexible assets under the risks present in both pricing designs is needed to be able to make a stronger recommendation.
- Adjusting other market design variables could help reduce volatility, such as loosening FRR qualification requirements to encourage flexible assets to provide FRR instead of implicit balancing and reducing the ISP to five minutes. Yet, further research is needed to assess the full implications of adjusting these design choices.

The findings presented by this research contribute to current efforts to improve balancing market designs, by showing that both pricing designs have vulnerabilities that allow flexible assets to exploit the market under certain circumstances, which is inconsistent with the objectives of the balancing market. The discussed design choices and future research recommendations contribute to improving balancing markets, to regain grid stability and market efficiency.

However, there are some limitations to consider. The model contains significant simplifications, such as the aggregation of multiple flexible assets into a single asset, which does not fully capture the decision-making process of multiple individual actors with a flexible asset in the real world. Additionally, the assumption that the flexible asset has perfect information and foresight does not correspond to reality. Furthermore, the heuristic optimisation method used may not always produce the optimal strategy for the flexible asset. Based on these limitations, several suggestions for future research are made. For example, using a bi-level optimisation method might provide more refined results of the asset's strategic behaviour. Future research could also consider the wider context of the market, such as how imbalance pricing influences the decision-making processes of numerous BRPs acting simultaneously, without perfect or with limited market information. An analysis from a game-theoretic point of view, showing the interactions among various actors, might lead to more refined or different outcomes when the different pricing designs are compared.