Over the past few years, the European Union (EU) has been making concerted efforts to shift towards cleaner energy by setting ambitious targets and formulating supporting policies. With its goal of becoming carbon neutral by 2050 (under the European Green Deal) and reducing its greenhouse gas emissions by at least 55 per cent by 2030 (with support of the Fit-for-55 package), it is preparing the continent to make the clean energy transition. To achieve these goals, it will rely on integrating large-scale variable renewable energy (VRE) into the grid. Under the Fit-for-55 package, the EU-level target for renewable energy is proposed to be increased to at least 40 per cent by 2030 (from 32 per cent previously) along with introduction of sector-specific measures for increasing its contribution to the final energy demand. This indicates the need for an accelerated renewable energy deployment. For instance, the EU will need 451 GW of wind capacity by 2030 (up from about 180 GW in 2021) and will have to double its planned annual capacity addition in the coming years.

This shift towards greater renewable integration comes with its own set of challenges for the grid operators along with the issue of grid resilience. Balancing supply and demand at all times is critical for a power system’s reliable operation given that even a small variation can disturb the power system frequency and may also affect system operations and reliability. The large-scale renewable energy penetration brings with it large swings in power supply and demand, thus increasingly making a case for flexible power systems.

In this backdrop, the European Network of Transmission System Operators for Electricity (ENTSO-E) recently released a position paper on the ‘Assessment of Future Flexibility Needs’ in Europe. The paper provides insights on identifying the flexibility needs arising from increasing variability in the balance of generation, demand and storage. 

Need for flexibility

Fundamentally, power system flexibility needs arise from the variations in the power system due to uncertain and fast-changing generation, demand and grid capacity.

In terms of demand, the electrification of heating, transport and industries has led to greater uncertainty and variation. For instance, lack of smart charging in electric vehicles (EVs) as well as larger electric loads subject to temperature variations, difficult-to-forecast customer preferences and uncertain price responsiveness can lead to huge fluctuations in demand, making it uncertain.

On the generation front, increased use of VRE and less dispatchable generation (which can be dispatched on demand at request of grid operators as per market needs) causes these imbalances in supply.

In grids, variation and uncertainty is caused by VRE, distributed energy resources (DER), and inverters [devices that convert direct current (DC) electricity to alternating current (AC)]. With large-scale deployment of inverter-based generators, the system inertia is expected to decrease and lead to increasing rates of change of frequency (RoCoFs), which can present a challenge in terms of frequency stability of the system.

Existing assessment approaches

EU’s extant power system planning methods already assess flexibility needs and the availability of solutions and products to cover flexibility gaps in many cases. These needs are addressed in the European Resource Adequacy Assessment (ERAA), the system operation, capacity allocation and congestions management and electricity balancing guidelines, and the Electricity Regulation. For instance, in recent years, ENTSO-E’s Ten-Year Network Develop Plan (TYNDP) has forecast decreasing inertia levels due to converter-interfaced renewables such as solar photovoltaic (PV), wind and battery resources becoming dominant. Coupled with increasing RoCoF due to the rising size in MW of sudden disturbances, this can lead to a new need for fast frequency response capacities. Indicators such as inertia, RoCoF, area control error or frequency restoration control error quality need to be investigated.

Ireland’s transmission system operator (TSO) EirGrid has illustrated the speed at which frequency response challenges can be met, while the adequacy and flexibility assessment study by Belgium’s TSO Elia provides examples of flexibility metrics that address unexpected generation and demand variations after the day-ahead time frame. Moreover, there are research and innovation (R&I) projects like MIGRATE (Massive InteGRATion of power Electronic devices), which successfully investigated systems with high penetration of power electronics. Another ongoing R&I project funded by the European Commission’s (EC) Horizon 2020 programme is the European OneNet flexibility market project, started in October 2020. Under this, various flexibility market solutions and the complementary tools for electricity network development and operations will be tested over a three-year period. In addition, the project will carry out regional studies on flexibility needs and studies on methodological development needs.

With all these developments, TSOs now need to look for solutions as to which additional flexibilities may be needed at what time in the future and in which European regions and also as to the possible negative effects of a flexibility gap on the power system.

Flexibility challenges

ENTSO-E’s position paper makes an assessment of future flexibility needs related to adequacy in the day-ahead time frame, that is, to flexibility needs arising from increasing variability in the balance of generation, demand and storage. It states that the increasing variability in the forecasted demand and generation in the day-ahead time frame can possibly lead to two new flexibility challenges.

Firstly, the periods when load is increasing and VRE generation is decreasing could require the decreasing amount of weather-independent generation to ramp up or down at a faster rate and over wider overall MW ranges than in the past to compensate for decreases in supply from renewable energy. For instance, the increase in load could become steeper due to increasing penetration of heat pumps and EVs, while the MW scale of VRE generation decreases (for example during sunset) would grow with higher VRE penetration.

Secondly, the decreasing amount of weather-independent generation may become insufficient to cover the demand during extended scarcity periods with very low VRE generation like windless winter weeks.

ENTSO-E’s approach

ENTSO-E conducted a survey of assessment approaches being presently used by 22 TSOs for future flexibility needs and on international examples of flexibility needs assessments, metrics and products. Based on the responses, ENTSO-E zeroed in on two flexibility needs—ramping and scarcity periods, along with the proposed metrics.

To illustrate how flexibility needs for ramping and scarcity periods could evolve by 2025 and 2030, ENTSO-E considered the example of Germany, which is both one of Europe’s largest countries and has one of the highest VRE penetration rates, based on ENTSO-E’s Mid-term Adequacy Forecast (MAF) 2020 data. Though the German system displays sufficient system adequacy in both 2025 and 2030 MAF analyses, its high amount of installed wind and solar capacities relative to both peak loads and installed dispatchable capacities have started posing challenges that are quite visible on ramping and scarcity periods.

Notably, the paper acknowledges that a simple analysis of residual loads cannot replace the complete and realistic chronological simulation of system adequacy, which includes forced outages, different climate years, as well as imports and exports. Therefore, ENTSO-E has proposed metrics that combine the strengths of detailed hourly results from chronological probabilistic simulations with insights gained from the simple residual load analysis with a focus on the two relevant and prevalent variability challenges as explained below.

Ramping flexibility needs: The ramping flexibility needs approach is partly based on experiences from California Independent System Operator (CAISO) and EirGrid. These metrics measure large daily residual load gradients, for example, at sunset in regions with large PV generation capacities. Residual load is the load left after subtracting VRE generation (such as wind, PV and run-of-river hydro) from the total demand, and is a useful indicator to show the flexibility and ramping needs for adequate power system operation. Explicit and implicit demand flexibility was considered as part of the dispatchable capacity, and not in the residual load calculation. As per ENTSO-E, the treatment of these capacities in the methodology could be further improved.

Figure 1 shows the 2025 and 2030 maximum ramps in the residual load over 1-, 3- and 8-hour time steps, which reach a substantial fraction of the total dispatchable capacity and even exceed it at times. The figure that relates the residual load to dispatchable capacities indicates a serious ramping-flexibility challenge, especially for the 2030 data, as dispatchable capacity or other flexibilities would need to be imported from neighbouring countries, or RES would need to be curtailed in a well-coordinated and anticipated manner to cover such ramps.

However, a key point to be noted is that residual loads do not account for imports and exports, which strongly contribute to overall system adequacy, especially for strongly interconnected countries such as Germany. Further, dispatchable capacities, which in any case only provide a very rough reference for the interpretation of residual loads, are adjusted or derated to account for forced outage rates of coal, gas, pumped-storage hydropower generation and other capacities, and include the MAF data for demand side response (DSR).

Broadly, the highest annual residual load MW ramps (calculated as the difference between residual loads 1,3 and 8 hours apart) can be easily compared between market zones and years after they are noramalised to the zone’s dispatchable capacity, accounting for demand response and forced outage derations.

The proposed metrics for ramping flexibility needs are percentage of loss of load expectation (LOLE), expected energy not served (EENS), and curtailed surplus energy during the 5 per cent highest ramp periods. These indicate how the ramping issue can pose an adequacy and economic problem. They will be assessed separately for positive and negative residual load ramps and for 1-, 3- and 8-hour ramps as well as the corresponding prior hours for potential pre-emptive curtailment. Presently, hourly values for LOLE, EENS and curtailed energy are among the outputs of chronological probabilistic market simulations used by adequacy and TYNDP studies. With the necessary fine-tuning of this indicator, the 5 per cent threshold can be addressed and assess how ramping capabilities of all resources are modelled in market simulations, especially demand response and VRE curtailment.

Figure 1: Germany’s maximum residual load ramps

Note: CY 2009 – climate year 2009; DE – Germany
Source: ENTSO-E

Scarcity period flexibility needs: These are metrics focused on contiguous-day EENS issues during scarcity periods, when VRE resources are not available for extended and continuous periods such as windless winter weeks in northern Europe.

Figure 2 shows a windless winter week where residual loads are almost as high as the loads themselves as a result of minimal VRE contributions. It is seen that the annual maximum of the 120-hour (or 5-day average) residual load, plus necessary frequency containment and restoration reserves (FCR+FRR), amounts to 96 per cent of maximum dispatchable capacities for 2025 and 102 per cent for 2030. In the example 5-day period in January, many of the 120 hours far exceed dispatchable capacities, which include demand response capacities, further indicating dependence on support from neighbouring countries.

In case the maximum annual value of 120-hour residual load rolling averages, including FCR and FRR requirements and normalised to the market zone’s derated dispatchable capacity, including demand response, is near 100 per cent, short-term flexibility resources such as batteries or DSR are not likely to cover power needs. That said, this metric can indicate small sets of hours in a given year when flexibility challenges are particularly strong. Meanwhile, market simulations can show quantified reliability risks from detailed simulations of dispatchable capacity, demand response, battery use, and mutual support between countries, as well as weather and outage probabilities.

As in the case of ramping, LOLE and EENS percentages over the maximum 120-hour average residual load periods indicate the fraction of overall adequacy concerns that arise due to seasonal scarcities involving extended periods of high residual load and low VRE generation. For further interpretation of scarcity periods, it may be useful to examine the climate years with high LOLE and EENS contributions during the identified 120-hour scarcity periods in market simulations, as well as the average generation as a percentage of the installed capacities of all VRE resources during these periods. This will help understand which climatic conditions can lead to scarcity periods.

The necessary fine-tuning of this indicator will not only address the focus on the single worst 5-day period, but also involve examining how the availabilities of flexibility resources during scarcity periods are modelled in market simulations, especially implicit demand response and sector coupling resources such as vehicle-to-grid (V2G), or seasonal thermal or hydrogen storage.

Figure 2: Sustained high net loads in Germany


Note: FCR – frequency containment reserves; FRR – frequency restoration reserves; max. disp. – maximum dispatch; CY 1985– climate year 1985
Source: ENTSO-E

Sector coupling and flexibility

Sector coupling can be an important source of flexibility in the energy system, ranging from energy storage technologies to demand-response solutions. Sector coupling refers to the integration of the energy consuming sectors—buildings (heating and cooling), transport and industry—with the power generation sector. It provides options to absorb excess electricity supply from renewable energy sources, to store energy and to provide back-up supply in times of high demand and prices.  

Sector coupling solutions such as power-to-heat with thermal storage and electrolysers using clean electricity combined with gas storage seem relevant for mitigating scarcity period flexibilities, especially for countries with high VRE shares. On the other hand, power-to-gas, smart electrolysers, V2G or smart EV charging are solutions that offer fast response flexibility and ramping flexibility—both before and during the steep evening ramp of the residual load.

Hydrogen-based sector coupling is also gaining traction and there is a growing interest in using hydrogen as a long-duration energy storage resource in a future electric grid dominated by VRE generation. Recently, Denmark’s TSO Energinet and pump producer Danfoss entered into an innovation collaboration to examine whether a plant that converts electricity to hydrogen can be used to help balance the power grid. The collaboration is part of Energinet’s Open Door Lab, which examines how to enable more flexibility in the power grid.

Sector coupling as well as other flexibilities such as batteries and DSR involve load that is connected at the distribution level, implying that their usage for the overall system requires close cooperation between TSOs and distribution system operators (DSOs). ENTSO-E suggests that to promote this cooperation between TSOs and DSOs, a joint assessment of flexibility needs for different use cases at the transmission and distribution level should be developed.

Challenges and way forward

Introduction of flexibility in the power system at a wide scale is undoubtedly a complex and significant challenge and hence the position paper is a step in the right direction. ENTSO-E is further planning to develop several additional flexibility need assessment methods related to stable frequency, congestion management, voltage stability and uncertain variations or forecast errors after the day-ahead frames along with associated metrics in the coming years. Once developed, ENTSO-E may directly apply or recommend to its member TSOs a fine-tuning and application of these methods, metrics and indicators on a national, regional and/or pan-European basis.

With proper planning and execution, Europe’s power system can certainly provide safe and quality power even in the face of an unprecedented energy transition.