Beyond the Flexibility Needs Assessment Keeping Belgium’s energy system in balance Understanding future flexibility needs and the actions required to meet them
Elia, October 2026
ELIA I Beyond the Flexibility Needs Assessment
Purpose of this report Belgium can maximise the value of its growing renewable generation by accelerating market flexibility while ensuring that system operators retain proportionate last-resort tools to keep the electricity system balanced under exceptional conditions. Drawing on the Flexibility Needs Assessment (FNA), this report sets out the implications for system balancing, distinguishes the measures needed to develop market-based flexibility from those required to safeguard system security, and identifies the actions that must begin now. The messages in this report are supported by DSOs within Synergrid.
Content 01
The FNA in brief
02
Scenarios and key assumptions
03
Key results of the FNA and our interpretation
Delivering the PV Response identified: secure and 04 extend the backstop while building future capability
05
More Flexibility, Through Collective Actions
The FNA (drafted by Elia together with DSOs within Synergrid) was submitted to public consultation from 24 July to 4 September 2026 and to CREG for approval on 3 October and approved on 8 October 2026. The present report is published alongside the approved assessment which can be found on Elia’s website.
ELIA I Beyond the Flexibility Needs Assessment I
Our three key takeaways from the FNA Belgium’s energy transition is entering a new phase. For two decades the central question was whether there would be enough electricity at moments of peak demand. That question has not disappeared, but a second one now sits alongside it: what happens when domestic generation exceeds what the country can consume, store or export. This is usually referred to as ‘incompressibility’. On sunny spring and summer days with low electricity demand, these situations are expected to become increasingly frequent over the coming decade. It reflects the rapid pace at which the electricity system is decarbonising. To make the most of the growing volume of clean electricity, the system will also need more flexibility, storage and electrified uses that can adapt consumption to periods when clean electricity is abundant. The Flexibility Needs Assessment (FNA) quantifies how much flexibility the electricity system will need in the years ahead, both to support growing shares of renewable
1
FLEXIBILITY IS GROWING, BUT NOT FAST ENOUGH TO MEET RISING SYSTEM NEEDS
As Belgium is electrifying and adding more renewable energy, the need for flexibility is increasing rapidly. The electricity system will need greater flexibility in both directions: to increase generation or reduce demand when electricity is scarce, and to absorb additional generation or reduce output when electricity supply is abundant. Significant growth is already expected in batteries, demand response and other flexible resources, with available flexibility increasing two- to fourfold by 2030 and 2035. Yet even under these assumptions, there will still be periods when excess low-carbon electricity cannot be fully absorbed by the system. Closing that gap will require an even faster deployment and activation of flexibility, both on the demand side and among decentralised energy resources including renewable ones.
2
energy and to maintain a secure and balanced system under all conditions. It also identifies the remaining gap between the flexibility that is expected to be available and the flexibility that will be needed. Flexibility needs related to network congestions are limited to a quantitative analysis on RES curtailment on distribution level following constraints on the TSO transformer at the TSO/DSO connection point. In contrast, the flexibility required to manage network congestion related to electricity consumption (load congestion) is not assessed within the scope of this document and is therefore excluded from the reported flexibility needs. This report highlights the main findings of the FNA according to Elia when it comes to balancing the grid, explains their implications for Belgium’s energy system and explores the actions that can help close that gap. In particular, PV can remain a cornerstone of Belgium’s energy transition—provided it also contributes to system balance, first through market response and, when needed, through a targeted backstop.
UNLOCKING THE FLEXIBILITY POTENTIAL OF PV — A CRUCIAL CONTRIBUTOR TO SYSTEM BALANCING
During sunny summer days, solar generation often becomes the largest source of electricity production. As a result, reducing solar output is often the main lever available to maintain the balance between electricity supply and demand. Making PV more flexible will therefore be essential to further develop renewable energy, and maintain a secure and efficient electricity system. Given the uncertainties on effective battery deployment, the development of end-consumer flexibility and nuclear energy, the results of the FNA show a range between 1 to 5 GW of uncovered needs by 2030, to be met by 2 to 8 GW of PV installed capacity. The needs rise to 2 to 7 GW by 2035, to be covered by 4 to 11 GW of PV installed capacity. This corresponds to 10–45% of the installed PV capacity expected in 2030 and 15–50% in 2035, compared with 3-5% that responded to market prices in 2025. Given the significant uncertainty, the evolution of this range will be closely monitored and periodically updated as new evidence emerges.
3
MARKET RESPONSE FIRST, WITH A TARGETED BACKSTOP AS INSURANCE
The best way to make the most of low carbon energy is to give consumers, businesses and energy producers the right incentives to adapt their consumption or production when the system needs it. In most situations, market signals will help keep the electricity system in balance. At the same time, a reliable electricity system also needs a solution for the exceptional and rare situations where the market does not provide enough flexibility. The objective is therefore twofold: to encourage the development of market-based flexibility, and at the same time to ensure a backstop mechanism - a last-resort technical safety net operated by system operators - is in place should the market not unlock all the flexibility required. The ability of the backstop should be anchored, not the amount. The amount, frequency and duration should be monitored and evaluated in time. Because such measures take years to design and implement, the reflections and implementation need to start today.
1
2 I ELIA I Beyond the Flexibility Needs Assessment
Our call to Action: Securing now a robust and flexible framework for decentralised PV response Even under ambitious assumptions on battery deployment and end-user flexibility, the FNA identifies periods in which the system cannot deliver enough downward flexibility to be operated safely. On sunny days with low demand, generation can exceed what Belgium is able to consume, store or export. Normally the market resolves this on its own: prices fall, consumption increases where it can be shifted, and generators reduce output. That mechanism only works, however, if demand is able to respond, if generation is efficiently exposed to the price signal, and if the steerability to act on it is available. This is where decentralised PV becomes decisive. When the system is long, solar is close to the only generation still running — and today only a few hundreds of MW of PV are reacting in the market. The FNA shows that in most scenarios, including those assuming ambitious battery deployment and end-user flexibility, there will be moments when more electricity is injected into the system than can be consumed, stored or exported, and the system cannot be brought back into balance. These are incompressibility risks. They are not an economic inconvenience; they can lead to system security issues. They are a call for PV to evolve and provide market response. As installed PV capacity continues to grow, the share that responds to price signals should increase at least as quickly, if not faster.
Market first — the most efficient lever A market-driven response is cheaper, better targeted and more predictable than any technical measure, and it should remain the primary route. It is also the lever with the largest untapped potential. The Belgian PV market is unusual by international standards: the large majority of capacity sits at residential and smallscale level, and a long history of support schemes has shielded those installations from any incentive to react to market prices. Fewer than 1% of households are on a dynamic contract today, although newer contract formats are emerging and gaining ground faster than dynamic contracts did. Beyond that, a series of barriers still holds back distributed flexibility — tariff structures, smart meter roll-out, data exchange, market access rules for aggregators, customer engagement and the
interoperability of devices and platforms. Removing these barriers is a no-regret measure: it allows more low-carbon electricity to be integrated, lowers system cost, and shrinks the volume that any safeguard would ever need to cover. The enablers are set out in section 5 and are, in Elia’s view, the first priority. Each of these barriers is actionable, and removing each barrier directly reduces what has to be secured by other means. Barrier removal deserves to be recognised as a policy pathway in its own right, alongside developing additional flexibility and installing additional renewable capacity.
But the market alone will not close the gap Even if the market delivers the flexibility already assumed in the scenarios—including increased demand-side response, additional storage—a residual system-security need remains. Across the FNA scenarios and sensitivities, and based on the onehour criterion, the total PV response required is estimated at 1–5 GW in 2030 and 2–7 GW in 2035. These ranges represent the total need, not the volume that must necessarily be directly steerable by system operators. Neither the expected pace of market development nor today’s technical requirements can be relied upon to cover this risk, including rare but severe events. Under the current framework (with response on all units above 1 MW installed after 2019), only around 1.1 GW of derated PV capacity would be steerable by system operators as a last resort in 2030, at the very bottom of the identified range. Without a broader safety net, system operators would have to fall back on far more intrusive measures, such as the automatic curtailment of generators or the disconnection of entire distribution areas. The safety net does not replace the market. Market response remains the first and primary lever; the backstop would be activated only as a last resort, when that response falls short.
A robust and flexible backstop mechanism as last resort A backstop is a technical last resort fail-safe allowing system operators to reduce generation on a subset of the PV capacity and other distributed generation,
ELIA I Beyond the Flexibility Needs Assessment I
activated only when the market response falls short and under clearly defined and regulated conditions. Belgium is not alone in exploring a robust and effective backstop. Elsewhere in Europe, national legislation already gives system operators the right to remotely curtail small generators as a last-resort measure. While these countries do not have the same share of small PV units as Belgium; Germany, Poland, Czechia, Northern Ireland, Italy and Switzerland already apply this to small commercial units and for some residential, and Austria will follow from 2028. The trend points towards greater controllability: upcoming European grid codes are expected to introduce additional requirements for small generation, and ENTSO-E is calling for lower national capacity thresholds. This should not be seen as a constraint, but as recognition that small generators now play a significant role in the generation mix and its operation. It is worth being explicit about what the backstop replaces. Without it, the only tools left to Elia during a severe event would be more intrusive and less proportionate measures: the automatic curtailment of generators, which applies indiscriminately to PV systems of all sizes under a European-wide obligation, or the disconnection of injecting distribution transformers — switching off entire neighbourhoods, consumers included, to resolve a surplus of electricity. Compared with that, reducing the output of a targeted subset of the Belgian solar fleet, in narrowly defined circumstances and for a very limited amount of time, is a far more proportionate instrument. For the large majority of PV owners it would pass unnoticed: the proposed volumes are sized against a criterion of, on average, one hour per year, and the ability to consume electricity from the grid is not affected.
How much volume would need to be secured ? Following the identified range for the PV response needs, Elia proposes to formalise and extend the existing operational backstop currently agreed annually with the DSOs. This measure allows system operators, as a last resort, to modulate individual larger PV and wind installations (> 1MW) when market-based flexibility has been exhausted and an incompressibility event threatens system security. Elia believes that this measure should be embedded in a robust regulatory and legal framework through the System Defense Plan. This framework should clearly define its exceptional purpose, activation conditions, governance and safeguards. In addition the regional legal frameworks and technical grid codes should provide the corresponding implementation basis. They should define the appropriate
technical set-up, including secure communication and control capabilities, so that installations of 250 kVA (installed after 2019) and above can be steered when the measure is activated. Elia believes that extending the backstop to installations of 250 kVA and above is a no-regret first step. It would raise the PV capacity steerable as a last resort from around 1.1 GW (derated) under the current framework to up to 2.2 GW (derated) by 2030, covering the lower part of the 1–5 GW range identified by the FNA.
The backstop must be regularly recalibrated as evolving needs and available levers affect its required size Whether further steps are needed — including whether PV installations below the proposed 250 kVA threshold should eventually be included in the backstop — cannot be decided today. It should depend on how the system evolves. On the need side, the key uncertainties are the pace of renewable deployment, particularly PV, the growth and profile of electricity consumption, and the development of inflexible generation such as nuclear. On the delivery side, they are the uptake of batteries and end-user flexibility, the share of PV responding to market signals, the speed at which new installations become steerable, and whether existing installations can be made controllable in a feasible and proportionate way. If these levers develop as assumed, the volume proposed above may prove sufficient; if they fall short, the scope may need to be extended. Rather than fixing the final volume today, regular monitoring should determine which path applies.
The reliability objective must be clearly established The proposed backstop mechanism aims to serve a specific purpose: preventing situations that would otherwise require more intrusive and disproportionate measures. Therefore, its sizing should not be derived solely from the standards defined for scarcity events, but from a broader assessment of the desired level of security of supply and risk tolerance. A less demanding criterion lowers the volume to be secured and limits how far down the installation scale the measur e has to reach, but leaves a larger residual risk of falling back on automatic curtailment of generators or disconnection of injecting distribution transformers. A more demanding criterion reduces that risk, at the cost of extending controllability requirements to smaller installations.
3
4 I ELIA I Beyond the Flexibility Needs Assessment
The backstop is an exceptional system security measure and does not replace the market
What is required, and by when for further developing the backstop
Alongside the backstop, removing the barriers to unlock more flexibility and particularly more PV response in the market is a must. Those are further outlined in the FNA and in this report. Additionally, Elia will assess whether and how downward flexibility products could efficiently unlock part of the PV response as well as additional demand side flexibility.
As regulatory reforms, technical standards and asset roll-out take several years to design, legislate and implement, work on this framework must start now to ensure that the required capabilities are available before the system needs materialise.
Our Call to Action: Securing now a robust and flexible framework for decentralised PV response
00 observes the market
04
T HE OVE R AR C HING PR INC IPL E
MARKET FIRST Flexibility (from demand, generation and storage) needs to be unlocked through the market. MO N I TOR C HANG E S | M ANAG E U NC E RTAINTY
adjusts the backstop
Continuous monitoring of needs and levers Track flexibility needs and available levers over time, and adjust the measures as the balance shifts
FUTURE-PROOF Ready for change
03
SECURE AND EXTEND THE BACKSTOP for exceptional conditions, when the market alone will not close the gap
01 Market first
Anchor the backstop in the System Defense Plan The last-resort backstop and other technical measures become part of the System Defense Plan Monitored continuously
02
Secured by a backstop
Extend the current volume as a no-regret step Lower the backstop threshold to 250 kVA, harmonised across Belgium This excludes any residential installation Future-proof by design
Steerability capability by design Remote steerability becomes a connection requirement (also for PV), enabling flexibility across markets.
ELIA I Beyond the Flexibility Needs Assessment I
1
ANCHORING THE BACKSTOP IN THE REGULATION
Elia currently believes the framework should be anchored at federal level in the System Defense Plan, the instrument recognised under European emergency and restoration legislation for last-resort measures safeguarding system security. A policy decision is needed on whether to integrate the production shedding backstop into it and, if so, the System Defense Plan should define the activation conditions, the governance and the safeguards that keep it proportionate and exceptional. At regional level, the legal and technical framework must give distribution system operators the right to act on individual decentralised assets when required to implement the measures foreseen in the System Defense Plan. Regional regulations and technical grid codes must set the requirements for secure communication and steerability, as a first step, for units above to or equal to 250 kVA. As regulatory reforms, technical standards and asset roll-out take several years to design, legislate and implement, work on this framework must start now to ensure that the required capabilities are available before the system needs materialise.
2
UNLOCKING ADDITIONAL PV RESPONSE WITHIN THE BACKSTOP AS NO-REGRET STEP
Market-based response should remain the preferred route. To ensure sufficient steerability if the market falls short, Elia proposes bringing installations down to 250 kVA within the scope of the backstop across all regions. This would increase the derated PV capacity available as a last resort in 2030 from around 1.1 GW under the current framework up to 2.2 GW (derated) depending on whether some existing installations fall within scope. Given the range of outcomes and uncertainties identified in the FNA, unlocking this volume is a proportionate, noregret measure that strengthens system security while preserving the primacy of market-based flexibility.
3
STEERABILITY CAPABILITY BY DESIGN
In parallel, connection requirements such as grid codes and technical regulation should be adapted so that all new installations (including smaller systems) – whether they consume, generate or store electricity – can communicate and be steered through standardised, interoperable interfaces. This should also be the case for PV installations. This requires coordinated action between system operators and the competent federal and regional authorities. This controllability is first and foremost a market enabler (alongside the other barriers brought forward in Chapter 5). It allows flexible assets – such as heat pumps, electric vehicles, batteries and PV installations – to respond to market signals through market parties and contracts, or to optimise locally via energy management systems. Market-based response is the primary route. The requirements should be developed together with stakeholders to ensure they are fit for purpose.
4
MONITORING THE MARKET AND RECALIBRATING THE BACKSTOP ON A REGULAR BASIS
The required backstop volume should be reviewed regularly as both the drivers of the need and the available flexibility levers evolve. Monitoring should cover, in particular, RES deployment, electricity demand and inflexible generation, alongside the uptake of demand-side flexibility, storage and market-responsive renewable capacity. If these developments materially change the residual system-security gap, the scope and sizing of the backstop should be adjusted accordingly.
5
6 I ELIA I Beyond the Flexibility Needs Assessment
The FNA in brief Building on more than a decade of flexibility assessments conducted by Elia, the Flexibility Needs Assessment (FNA) mandated by the European regulation broadens the scope of previous AdeqFlex studies and provides a basis for defining Belgium’s future non-fossil flexibility needs. Earlier publications focused mainly on system security — how much flexibility is required to keep the system balanced in real time. The FNA retains that objective and adds two dimensions: renewable integration, assessing how flexibility can absorb high renewable output and limit structural excess, and grid constraints, assessing how congestion can restrict flexible resources and lead to renewable curtailment. Together these three dimensions inform indicative national objectives for nonfossil flexibility and help determine whether additional support mechanisms may be required. The assessment is a European obligation. EU Regulation 2024/1747 requires Member States to produce a flexibility needs assessment every two years, following a methodology developed by ENTSO-E and the DSO Entity and approved by ACER. The Belgian implementation plan, setting out the methodology and scenarios to be used, was approved by the CREG. The FNA is drafted by Elia as designated entity, in cooperation with Fluvius, Sibelga, ORES, RESA and Arewal, and is subject to approval by the CREG — it is not an Elia opinion piece but a regulated analysis grounded in an EU methodology. The FNA quantifies flexibility needs. It does not decide how they should be met. Its output is a trade-off: either additional non-fossil flexibility is developed, or greater reliance is placed on modulating renewable output, or some combination of the two. That choice belongs to policymakers. In Elia’s view, that modulation should be market-based in the large majority of cases; only for system security events should it take place outside the market. ‘Curtailment’ is often read as a failure of the system to integrate renewables. At high shares of renewable generation the opposite is true: modulating renewable active power, used proportionately and alongside other integration tools, is one of the most effective means of integrating more renewable capacity. The same market signals that trigger this modulation should also encourage market parties to increase consumption at such moments, or to invest in flexibility, such as storage, that shifts surplus energy to later hours, where this is economically justified. Relying on it, however, is only viable if enough renewable capacity can actually be modulated safely when required — which is where this report focuses.
ELIA I Beyond the Flexibility Needs Assessment I
NEEDS are assessed and combined
Renewable integration
RE S Integration needs
System security
Short-term and ramping needs
Grid constraints
Additional flexibility needs After accounting for flexibility expected to be delivered in the Market/CRM
Network needs
SOLUTIONS are quantified
1
Additional non-fossil fuel flexibility needs for full RES integration (no curtailment)
2
RES curtailment without additional non-fossil fuel flexibility
Additional RES response needs After accounting for expected response in the market
Figure 1: The three dimensions of the FNA, bundled to define indicative national objectives for non-fossil flexibility
The FNA answers what the power system needs are. This report goes one step further and also addresses the consequences: which risks matter most, what they mean in practice, and which measures should follow. Except for the section on barriers and enablers, the figures presented in the call to action are not part of the FNA itself, although they remain factual and are directly derived from it.
Official documents FNA Flexibility Needs Assessment Public consultation Consultation document & responses CREG approval Final decision / approval
7
8 I ELIA I Beyond the Flexibility Needs Assessment
02
Scenarios and key assumptions As agreed with the DSOs and the CREG, the analysis uses the Current Commitment (CC) and Constrained Transition (CT) scenarios for target years 2030 and 2035 outlined in the Adequacy & Flexibility study published in June 2025 (AdeqFlex’25). CC reflects current climate policies; CT assumes lower renewable development and lower electrification across Europe. The scenarios deliberately leave the adequacy gap unfilled: they assume no new thermal capacity nor large-scale batteries beyond those already contracted in CRM auctions before 2025, corresponding to 1.5 GW of large-scale storage in Belgium. That figure does not include the additional 0.7 GW contracted through the CRM auction of 2025, so it should be read as a minimum deployment assumption rather than a forecast. In other words, the capacity that will eventually be built to close that adequacy gap — and which would also provide flexibility — is not counted, so the resulting flexibility needs are, on this point, conservative. Given recent market developments, additional variants (BAT+) were built assuming the large-scale battery potential identified in AdeqFlex’25 materialises — 4.5 GW in 2030 and 5.7 GW in 2035. Large-scale batteries here cover all non-residential battery assets connected to either the transmission or the distribution network; residential batteries are modelled separately. These BAT+ variants are distinct from the sensitivity of the same name used in AdeqFlex’25. It is however important to note that the CC and CT scenarios do already include additional residential batteries and increased flexibility shares from consumers (both residential and industrial loads). Sensitivities were also added on nuclear capacity given ongoing discussions on potential extensions of closed reactors and on slower household flexibility penetration, mainly regarding electric vehicles.
ELIA I Beyond the Flexibility Needs Assessment I
4 scenarios (based on CC and CT from AdeqFlex’25) Constrained transition (CT)
Current commitment (CC)
CC-BAT+
CT-AF25
1.5 GW
4.5 GW in 30’ 5.7 GW in 35’
CT-BAT+
Focus on expected excess and impact on renewable targets
1.5 GW
ECONOMIC DISPATCH SIMULATIONS
4.5 GW in 30’
2030
5.7 GW in 35’
2035 SYSTEM SECURITY NEEDS Focus on unexpected excess and impact on realtime balancing
2 types of
sensitivities
3 types of analysis RES INTEGRATION NEEDS
Low wind Low load / flex
Current policies (NECP)
CC-AF25
2 target years
3 – 4 GW
Unlocking rate / 2
3 – 4 GW
APPROVED DISTRIBUTION AND TRANSMISSION NETWORK DEVELOPMENT PLANS Figure 2: Scenarios and target years used in the FNA
GRID CONSTRAINTS
9
10 I ELIA I Beyond the Flexibility Needs Assessment
Renewable capacity expected to nearly double by 2035, but price responsiveness remains limited today Installed renewable capacity is expected to increase sharply by 2030 and 2035 (based on regional and federal Energy and Climate Plans), roughly doubling total solar and wind capacity compared with today. Onshore wind is assumed to grow more slowly in CT as outlined in AdeqFlex’25. Offshore wind is assumed to reach 4.4 GW by 2035.
The lower bound expresses structural reactions measured while the higher bound is measured against the most extreme conditions only — the 1% deepest negative prices — the volumes reacting are materially higher, and first observations for 2026 show them increasing further, particularly for PV. Elia will put both the method and the updated values to consultation in the framework of AdeqFlex’27. This is the core of the incompressibility challenge: renewable capacity is expanding rapidly, but only a small share of installations regularly adjusts output during periods of excess generation.
Based on observations over 2025, the current ability of renewable assets to respond to market signals remains uneven. The figures below measure structural reactions, meaning assets that reduced output during at least a given share of negative day-ahead price periods:
The FNA therefore also focuses on how much renewable response is required to maintain system security, and includes a sensitivity in which all wind is assumed controllable, in order to isolate the remaining need for additional PV response.
― Around 3-5% of PV capacity, roughly 300-650 MWp, reacted structurally to market prices in 2025;
― Around 30-60% of onshore wind capacity, roughly 1,000-2,100 MW, reacted structurally to market prices in 2025;
― Offshore wind is assumed to fully respond to market prices.
Installed Solar PV (GWp) x
Installed Onshore Wind (GW)
Installed Offshore Wind (GW)
2 20.9
16.5 x
2025
2030
2035
3-5% of PV reacted to market signals in 2025
No market reaction observed
1-2
0.3-0.6
3.6 2025
2
6.0
CC
7.6 scenario 4.2
2030
4.5 2035
30-60% of wind onshore reacted to market signals in 2025
x
CT scenario
2.3
12.0
2
2.3
2.3
2025
2030
4.4 2035
100% of wind offshore reacts to market signals
Current observed market reaction (based on structural modulation of generation during low prices)
Figure 3: Solar and wind roughly double compared with today, however their reaction to the market prices is still limited for decentralised capacities
ELIA I Beyond the Flexibility Needs Assessment I
A DISTINCTIVE FEATURE OF THE BELGIAN PV LANDSCAPE: RESIDENTIAL DOMINANCE
Centralised flexibility expected to grow by a factor of two to four
Belgium stands out compared with many other countries due to the high share of residential photovoltaic (PV) installations in its solar portfolio. Although future growth is expected to include a greater number of medium-scale installations, residential systems are projected to remain the dominant segment. This characteristic presents a challenge from a flexibility perspective, as most small-scale installations currently have limited capability or incentives to respond to market signals and contribute to system flexibility.
Centralised downward flexibility — the ability to absorb surplus electricity by increasing consumption or charging large-scale batteries — comes from nonresidential storage and demand response. It grows by a factor of two to four by 2035, depending on the scenario (AF25 or BAT+), driven mainly by additional large-scale batteries and, to a lesser extent, by demand-side management from newly electrified processes.
The share of residential installations in newly installed capacity is nevertheless declining, and the segment mix is one of the factors that determines how responsive a national fleet can become.
Current share of residential PV installations across selected countries [%]
These volumes express power, not energy. Figure 5 shows the capacity available to absorb surplus electricity, not how long it can be sustained. Batteries and pumped storage are limited by their energy content, typically a few hours, and demand response by process constraints. Since excess-generation events can last beyond those constraints, the actual contribution during long events will be lower than the installed power suggests.
X4
GW
7.9 GW
8
6.7 GW
7
2.3
BAT Potential set in AF25*
4
0.7
Large BAT (CRM 25)
3
1.0 0.6
6
~65
3.5 Assumed in CC/CT BAT+
5 43 33
BE
NL
HU
28
UK
22 ES
22 DE
20 FR
2 5 GR
Sources: SolarPower Europe Market Report, EnergyVille DER Report and national data sources. Definitions of residential/small-scale PV differ by country: Belgium (<10 kVA), Netherlands (<15 kW), Spain (non-ground-mounted PV only), United Kingdom (<10 kW), Germany (<10 kW), Hungary (residential rooftop PV only) and France (<36 kVA).
Figure 4: Share of residential PV installations across selected countries [%]
1 0
2.1 GW 0.5 0.3
BAT DSM
1.3
PHS
2025
2030
Large BAT (CRM <25) Additional DSM**
Assumed in CC/CT AF25
2035
* Based on existing projects being realised, CRM auctions and other drivers ** DSM based on the CC scenario *** BAT includes all non-residential batteries
Figure 5: Assumptions regarding the evolution of downward flexibility in large scale installations (PHS, BAT and DSM)
11
12 I ELIA I Beyond the Flexibility Needs Assessment
Decentralised flexibility is assumed to triple Decentralised downward flexibility could almost triple under the Current Commitment scenario (Figure 6), driven mainly by electric vehicles and residential batteries. This depends on a major increase in the share of enduser devices that actually respond. The right-hand part of Figure 6 shows how many devices are assumed to become responsive by 2030 and 2035, reflecting continued electrification, digitalisation and wider access to control signals. These assumptions are ambitious compared with what is observed today (see Figure 7 for the assumed shares of responsive residential devices). Four points matter when reading these figures:
― More devices do not mean proportionally more capacity. What can
― Availability is lowest when it is needed most. Heat pump flexibility is at its
lowest in spring and summer, and many EVs are not plugged in around midday, which is precisely when solar-driven excess occurs. ― The GW values are estimated averages over the simulated periods. The volume actually available at a given moment can be higher or lower depending on the hour, the type of day (weekday or weekend), preceding and following conditions, and whether devices are connected. ― The volumes express power, not energy. As with centralised flexibility, Figure 6 does not show how long this power can be sustained. Since some excess events last several hours, the actual contribution during these events will be lower than the power shown suggests.
realistically be mobilised depends on user behaviour, technical limits and comfort.
GW
million unlocked devices
8
4
Contributions limited during spring and summer
3.3 6
4
2
0
HP
3
X3
1 GW
2
HP EV
1.6 0.7
0.9
2025
2030
1.5
2.7 GW
2035
1
BAT 0
0.4 2025
2.2
EV
Flex to unlock
0.7 BAT
2030
2035
Figure 6: Assumptions regarding the evolution of downward flexibility in residential installations (BAT, EV, HP) based on Current Commitments scenario
ELIA I Beyond the Flexibility Needs Assessment I
Responsiveness at residential level is the weak link Observations show that electric vehicles respond today to around 30%; heat pumps follow at around 10%; residential batteries are effectively fully responsive, since optimising when they charge and discharge is their basic purpose. Explicit exposure to the market, however, remains marginal: fewer than 1% of households are on a dynamic contract. This explains the reason for the very limited share of ‘market based optimisation’ observed today and the large share of ‘local optimisation’. Indeed only 2% of EVs and 20% of all installed home batteries are found to be optimised on market signals.
consumption or avoiding peak tariffs, rather than by market or system signals. Although these objectives are legitimate, they do not necessarily coincide with the needs of the electricity system at a given moment, such as maintaining system balance or relieving transmission grid congestion. In light of these observations, the shares of responsiveness assumed in 2030 and 2035 should not be considered as a given. Many barriers are to be relieved to speed up the pace of participation of end-users.
Many of these devices are driven by local signals, such as maximising self-
Unlocked shares of residential devices - % of total fleet (market & local) + scenario* Electric Vehicles
%
Heat Pumps
100
15 Today: 10% (local) 1% (market)
Today: 28% (local) 2% (market)
Today: 1% contracts are dynamic
Today: 80% (local) 20% (market)
80 85
55 45
20
28 0
Dynamic Contracts
Home batteries
23’
24’
2
10
15
25’
30’
35’
23’
24’
10
10
25’
30’
100
20 20
20 35’
23’
24’
25’
30’
35’
0.2 0.4 0.9
5
23’
30’
24’
25’
Local optimisation Market-based optimisation * Based on assumed shares in the Current Commitments scenarios of AdeqFlex’25 used as basis for the FNA.
Figure 7: Unlocked shares of residential devices as % of the total fleet and evolution assumed in the CC scenario
20 35’
13
14 I ELIA I Beyond the Flexibility Needs Assessment
03
Key results of the FNA and our interpretation This section aims to highlight and explain the key findings of the FNA. It is not intended to provide an exhaustive assessment, and certain interpretations and selections of key figures reflect Elia’s perspective. Readers are encouraged to consult the official FNA report for a comprehensive analysis and full context.
A. Renewable integration: curtailment (in yearly energy terms) stays limited Rising shares of renewable generation will lead to more frequent periods in which domestic supply exceeds demand. Prices may turn negative and renewable plants may need to curtail production — a trend already visible today. Towards 2030 and 2035, annual renewable curtailment in Belgium due to excess domestic generation is expected to reach around 1 to 2% of total yearly solar and wind generation.
The results of the FNA show that without additional non-fossil flexibility beyond what the scenarios already assume, annual RES curtailment — including the curtailment needed for system security — remains limited:
― 1.0 – 1.3% of total wind and solar generation in 2030, falling to 0.4 – 0.5% in the BAT+ scenarios;
― 1.9 – 2.5% in 2035, falling to 1.1% in both BAT+ scenarios.
ELIA I Beyond the Flexibility Needs Assessment I
Beyond the annual totals, three characteristics matter:
― Curtailment peaks of 2.7 – 4.4 GW in 2030 and
5.4 – 7.5 GW in 2035 are possible; ― Events typically occur around midday in spring and summer, and are therefore mainly PV-driven; ― Some events can last up to 10 hours which goes well beyond what can be easily stored in electrical storage installations. Unlocking end-user flexibility is a no-regret measure, because curtailment levels remain highly sensitive to how flexibility develops. They fall if additional flexible capacity is unlocked and rise if end-user participation falls short of the already ambitious assumptions embedded in the CC and CT scenarios. Weather matters too: sunny weekends and holidays materially increase the risk.
Total curtailed energy** compared to the generated wind and solar [TWh]
2030
60 RES Generation [TWh]
These volumes are concentrated in a limited number of hours: around 200 hours in 2030 and 400 hours in 2035, roughly halving under BAT+. They do not translate one-to-one into negative price periods observed in the wholesale markets, since negative prices can also result from a lack of flexibility from non-renewable generation or from curtailment abroad propagating to Belgium.
2035
40
20
0
0.5%
1.0%
CC
1.3%
0.4%
1.1%
1.9%
1.1%
CC
CT
2.5% *
CT
Total curtailed energy** [TWh]
0.16 0.34
0.13
0.41
0.60
1.02
1.18 0.50
= BAT+ scenario with more large-scale batteries * Percentage curtailed compared to total generated wind and solar ** Total curtailment (after finetuning) combining the needs for excess RES generation, ramping needs, short term and network needs.
Figure 8: Amount of curtailed energy compared to the wind and solar generation for each scenario and target year
15
16 I ELIA I Beyond the Flexibility Needs Assessment
Two ways to offset curtailed energy and one seems more reasonable Federal and regional Energy and Climate Plans do not account or define targets for curtailed generation 1. If the same renewable energy volume must be delivered, there are two possible routes (as also outlined in the FNA):
― develop additional non-fossil flexibility to achieve full integration, or ― develop additional renewable capacity to compensate for what is curtailed. Intermediate combinations exist. Other measures can also help to a certain extent such as more east–west PV orientation or accelerating electrification in transport and heating beyond the assumptions taken in the CC & CT scenarios so that additional demand coincides with daytime renewable production. The arithmetic is stark. Eliminating almost all curtailment would require several gigawatts of additional flexible capacity — up to 9 GW in 2030 and 18 GW in 2035, even under BAT+ — largely to cover rare peak events lasting one
to a few hours. Offsetting the same energy with additional renewables would require around +0.5 GW of onshore wind or +1.7 GW of solar PV by 2035, even accounting for the extra curtailment that additional capacity itself triggers. That is broadly equivalent to bringing forward renewable deployment by about one year at the current pace. The final trade-off is a policy choice for the National and Regional Energy and Climate Plans. The results indicate that building large amounts of additional nonfossil flexibility solely to avoid a limited volume of curtailed energy is unlikely to be economically efficient, and that the business case for such flexibility would not stand on its own. The two routes are not directly comparable in economic terms, however: they represent different system outcomes, and a full comparison would also have to account for the value of the renewable energy not produced, the effect on captured revenues and financing conditions, and the cost of the flexibility itself. Such an assessment falls outside the scope of the FNA and of this report.
Additional flexibility [GW] needed to avoid curtailment when allowing for 0.1% of RES to be curtailed
Additional RES capacity [GW] needed to compensate for the curtailed energy OR
9
CC
18 17
CT
1h
2h
4h
8h
2030
CT
0.5
2035
9
23
CT
0.2 0.2
CC
2030
CC
21
CT
0.4
CC
CC
2035
2030 2035
CC
2030
11
2035
11
CT
CC
CC
0.7
0.3
CT
0.6
0.2
CT
0.5 0.5
1.3 1.7
CT
0.1 0.1
0.2 0.2
BAT+ scenario with more large-scale batteries
Figure 9: Additional non-fossil flexibility (left) or additional renewable capacity (right) required to compensate for RES curtailment 1
To Elia’s knowledge, there are no indications of curtailed energy assumed in those plans.
ELIA I Beyond the Flexibility Needs Assessment I
B. System security: the real constraint is power, not energy The analysis confirms the AdeqFlex’25 trend: short-term flexibility needs increase in both directions with higher renewable penetration. It also confirms a growing need for downward flexibility to manage excess generation. In periods of incompressibility, when solar is close to the only generation still running, that downward flexibility has to come mainly from PV. Based on the FNA results, the total PV response required is estimated at 1–5 GW of derated capacity in 2030 and 2–7 GW in 2035. Applying the 65% derating factor (reflecting that the PV fleet does not generate at its nominal capacity during relevant system conditions), these ranges correspond to approximately 2–8 GWp of nominal PV capacity in 2030 and 4–11 GWp in 2035. This means that 10–45% of solar capacity should be able to react by 2030, and 15–50% by 2035, depending on other market developments. This is the point where the two halves of the story separate. The energy volume of curtailment is expected to be small in Belgium (as found in the renewable integration analysis). However, the power requirements at the moments
Upward flexibility needs* [MW] - shortage
+35%
when the system is long are not small. A limited annual energy problem and a significant instantaneous capacity problem are two different questions, and they call for different requirements.
Short-term flexibility needs increase in both directions: roughly +40% over the next decade Short-term needs, derived from forecast errors and forced outages between intra-day and real time, rise steadily. Total downward needs increase from 2.4 GW in 2026 to 2.8 GW in 2030 and 3.4 GW in 2035. Total upward needs increase from 3.4 GW in 2026 to 3.8 GW in 2030 and 4.7 GW in 2035. These are gross needs, before accounting for available flexibility. After comparing needs with available margins, remaining uncovered upward needs are expected to continue being managed through existing crossborder balancing markets and reserve procurement. Remaining uncovered downward needs are substantial and may require additional flexible resources.
Downward flexibility needs* [MW] - excess By 2035, the Belgian system will require
4,650 CC CT
+40%
3,800 3,400
300 2026
2,000
1,700
400 2030
2026
2035
Fast
450
350
300
Very fast Very fast flexibility required to react in 5 minutes
2,400
2,400
2,100
350
CT
2,800
2,700 1,900
3,400 CC
Fast flexibility required to react in the last quarter hours
2030
Total
2035
3 - 5 GW of total flexibility in the last hours before real time, of which 2 - 3 GW fast needs to be able to react in the last quarter hours, and up to 0.4 GW needs to react within 5 minutes.
Total flexibility required from day ahead to real-time
* P0.1/P99.9 flexibility needs assessed across the whole year
Figure 10: Evolution of upward (left) and downward (right) short-term needs towards 2030 and 2035
17
18 I ELIA I Beyond the Flexibility Needs Assessment
PV response is the missing link in managing excess generation During periods of excess generation, especially sunny low-demand hours, PV is close to the only sizeable generation source still injecting into the system. The issue is therefore not whether curtailment is acceptable in principle, but whether enough PV capacity is technically able, contractually enabled and economically incentivised to turn down when required — and, failing that, steerable as a last resort. Without that capability, excess generation remains on the system and becomes a system security issue rather than a market optimisation question. Incompressibility situations are illustrated in the figure below, which distinguishes between two types of events. In the first case (type A), the system
is already expected to face an excess of generation under perfect foresight: the surplus is visible from the day-ahead stage and progressively carries through to real time. In the second case (type B), the system is already tight but the expected surplus can be resolved in day-ahead. However a forecast error (for instance higher renewable generation than expected) creates an excess in (close to) real-time that must then be resolved operationally. Both situations require corrective action and can lead to incompressibility issues. Recent events observed in 2026 can be mapped to these two categories as shown on the figure.
(B) Uncovered short-term needs (no flexibility left)
MW
Easter Monday April 6, 2026
Forecast errors
(A) Perfect foresight needs (no flexibility left) Sunny Sunday April 26, 2026 Labour day May 1, 2026
Load (with flexibility)
(A) Perfect foresight needs / structural excess Ability of the market to manage well ‘predicted’ situations of high renewable generation. It is related to the ability of market parties to maintain a balanced portfolio during high renewable energy conditions (B) Uncovered short-term needs / deviations to perfect foresight Ability of the system to maintain sufficient shortterm flexibility to manage unexpected outages or forecasting errors. It is related to available downward flexibility in the system
PV Wind Thermal must run Nuclear
0h
12h
24h
Figure 11: Illustrative example of a sunny weekend day and types of incompressibility events
ELIA I Beyond the Flexibility Needs Assessment I
What is the remaining PV response required ? After accounting for all (existing and additional) flexibility assumed in the four scenarios and for existing observed response on large and decentralised assets, the analysis determines the additional renewable response required. The results show that additional onshore wind response alone cannot cover the need: additional PV response is required. In addition, when assuming all wind is steerable and 5% of the existing PV fleet responds; the need amounts to a derated PV capacity of (in case of a 1-hour criterion):
― in case of 2 GW nuclear in the system: 0.8 – 3.6 GW in 2030 and 1.7 – 5.2 GW in 2035.
― in case of 3 GW nuclear, the needs are roughly increased by 1 GW.
These remaining needs assume that all wind already responds structurally to negative prices (not yet the case today), that new storage is commissioned, and that new end-user flexibility is unlocked through markets, tariffs or other incentives — ambitious relative to what is observed today. They also come on top of the PV response already observed in the market of around 5% of today’s fleet (the upper end of the 2025 observations). The assumed response of 5% extrapolated in the future corresponds to around 0.4 GW (derated) in 2030 and 0.6 GW (derated) in 2035, as the responsive share is assumed to grow with the installed fleet. The total PV response required therefore equals the remaining need shown in Figure 12 plus these volumes — around 1.2–4.0 GW (derated) in 2030 and 2.3–5.8 GW (derated) in 2035, with 2 GW of nuclear.
PV response needed when already accounting for full wind response and additional central and decentral flexibility as outlined in the scenarios: case with 2 GW nuclear in the system Additional PV response needs in 2030
Additional PV response needs in 2035
After accounting for 5% PV response (≈0.4 GWd) 100 % onshore and offshore wind responding
6
After accounting for 5% PV response (≈0.6 GWd) 100 % onshore and offshore wind responding
6
Least optimistic scenario
1 hour 5
5
4
1 hour
Least optimistic scenario
3
3.6 GW CT-AF25
2.2 GW CT-BAT+ 1 hour
1
Most optimistic scenario
3.5 GW CC-AF25 3.5 GW CT-BAT+
3
2.7 GW CC-AF25
2
4
2
0.8 GW CC-BAT+
5.2 GW CT-AF25
1 hour
Most optimistic scenario
1.7 GW CC-BAT+
1
0
0
1
3
10
20
30
40
50
Hours at risk [h]
60
70
80
90
100
1
3
10
20
30
40
50
60
70
80
Hours at risk [h]
Figure 12: PV response needs under the CC/CT and AF25/BAT+ scenarios, assuming full participation of decentralised wind generation: in case of 2 GW nuclear in the system
90
100
19
20 I ELIA I Beyond the Flexibility Needs Assessment
These volumes are expressed as derated PV capacity because the Belgian fleet does not generate at its nominal capacity at national level, a derating factor of 65% applies (see the explanatory box). Securing a given derated volume therefore requires roughly one and a half times as solar panels installed capacity to be made controllable. These volumes are set against a reference criterion of one hour per year: they are sufficient to secure the system while still allowing, on average, up to one hour of uncovered excess conditions per year. This is a statistical result across multiple climate years, consistent with AdeqFlex’25 — not a prediction that the volume would be triggered for one hour in every year; some years would require more, others none. It is a reference point rather than a regulatory requirement as no reliability standard exists for it today, in either the European methodology or Belgian legislation. The choice is material: applying a less demanding criterion of three hours per year instead of one would reduce the volumes of PV response needed by roughly 1 GW. Setting that level is a policy decision, not one for the system operator. What Elia can state is that, as system operator, it needs operational solutions available at least up to the one-hour level; this does not mean that every instrument has to be dimensioned on that basis. Because those instruments exist to prevent situations that would otherwise require emergency or last-resort measures, the criterion should follow from a deliberate assessment of the desired level of security of supply and risk tolerance, rather than be inherited from other frameworks such as the adequacy one.
Many uncertainties remain on the effective amount required The future PV response need varies significantly across scenarios, depending on how much other flexibility actually materialises. Additional large-scale storage and demand-side capacity reduce these needs; slower market development increases them. Both bounds are contested, and in opposite directions: the lower bound assumes a battery deployment that several market parties consider unlikely without further investment signals, while the upper bound rests on observed renewable response levels that the renewable sector considers too conservative. All in all, the calculated PV response needs incorporate the uncertainty associated with future developments and the differing perspectives of various market participants. The sensitivities on nuclear lifetime extension and on lower end-user flexibility both point upwards. Figure 13 sets out these outcomes for each scenario and target year for the 1-hour criterion for the total PV that would need to react in the system (not making a distinction on whether those volumes are unlocked in the market). The resulting range is wide by construction: it reflects genuine uncertainty on how the market will evolve, not imprecision in the assessment. Across the FNA scenarios and sensitivities, on the total PV response required to meet the one-hour criterion ranges looking into the different uncertainties/ scenarios, a range of 1 to 5 GW can be derived for 2030 and from 2 to 7 GW for 2035. These ranges should be treated as planning ranges rather than targets: they reflect uncertainty about how the system and available flexibility will evolve and do not prejudge the instruments through which the response should be delivered. As explained above, the ranges are expressed as derated PV capacity. A 65% derating factor is applied to convert installed nominal PV capacity into the effective output expected to be available during relevant system conditions. Consequently, securing a given derated response requires a larger volume of nominal PV capacity to be controllable. This leads to a nominal capacity of 2 to 8 GWp required to respond (or 10 to 45% of the fleet) in 2030 and 4 to 11 GWp (or 15 to 50% of the fleet) in 2035. The resulting volumes are summarised in the figure beside.
ELIA I Beyond the Flexibility Needs Assessment I
Total decentralised PV (100% available) response needs for a 1-hour reliability criterion and system needs to be prepared for (based on FNA results)
2035
2030 2 GW
2 GW
4 GW
3 GW
[Derated GW]
3 GW
EV
7.9
Unlocking rate / 2
6.2 5.2 4.2
2.7
3.2
2.2
6.2 6.1
5.2 4.7
3.7
6.9
5.9
5.2
4.2
3.2
4 GW
4.2
System needs to be prepared for a total PV response of
5.1
4.5
4.3
4.1
~ 1 to 5 GWd or 10 to 45%
2.3
of the fleet
2.8
3.3
1.2 0 Current Commitments Constrained Transition
Range
AF25
0
BAT+
The numbers represent the total PV response required (including the ones already responding today). To convert the numbers in nominal capacity (MWpeak), the numbers should be multiplied approximately by 1.5.
Figure 13: Total PV response required for each scenario based on 1-hour criteria
System needs to be prepared for a total PV response of
~ 2 to 7 GWd or 15 to 50% of the fleet
21
22 I ELIA I Beyond the Flexibility Needs Assessment
NOMINAL PV CAPACITY x 65 % = DERATED PV CAPACITY The volumes of needed PV controllability in this report are expressed as derated PV capacity — the effective injection actually available at the moment the system is long — not as installed nominal capacity (GWp), the maximal power the PV panels can produce. The two differ materially, and confusing them overstates what a given fleet can deliver. A similar parallel can be made with adequacy where derating factors represent the contribution of a certain technology to adequacy.
This ratio is rather stable. Across all incompressibility-risk quarter-hours — those with negative day-ahead prices — in the last 3 years, peak derating factors remained consistently within the 60–70% band, with the same behaviour observed in the summers of 2024, 2025 and 2026. A derating factor of 65% is therefore used throughout this report when converting nominal to derated volumes. As an illustration, 20 June 2026 — the day before the solstice and one of the sunniest days of the year — saw Belgian PV output peak at 8.12 GW at 13:30, against 12.55 GWp installed at that time which is around 65%. The practical consequence is that the nominal capacity that must be made controllable is roughly one and a half times the derated volumes quoted in this report.
14
Installed capacity ≈ 12.5 GWp
12 National PV production (GW)
Belgian PV output peaks at around 60–70% of installed capacity, even on the sunniest days. This is structural rather than accidental: inverters are commonly undersized relative to the panels installed behind them, and at national scale the fleet aggregates a wide range of orientations, tilts and local shading conditions, so installations do not peak simultaneously.
Belgian solar PV output - 20 June 2026 (sunny summer day)
10
Peak: 8.12 GW at 13:30 ≈ 65% of installed capacity
8 6 4 2 0
00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00
Time
Figure 14: Illustration of the PV output on a sunny summer day compared to the nominal installed PV capacity
ELIA I Beyond the Flexibility Needs Assessment I
C. Grid constraints: limited overall, with local exceptions Based on a high-level assessment within this FNA, and within the perimeter assessed, grid constraints during periods of high renewable generation remain limited overall. Curtailment on distribution level following constraints on the TSO transformer at TSO/DSO coupling points is expected to occur only under high-penetration or specific local conditions, while batteries connected to the transmission grid are expected to face minimal offtake constraints during peak renewable infeed. Such analysis will be further expanded in upcoming versions of the FNA. On the transmission grid, offtake constraints on flexible assets — particularly large-scale batteries — are expected to remain very limited even during periods of high renewable output above 60% of RES capacity, and to improve as reinforcements are delivered in 2030 and 2035. This assessment is based on detailed grid connection studies and provides a first view of the link
Transmission system scope
Unavailability of flexible resources
Quantitative assessment on unavailability of TSO-connected batteries during high RES conditions
During high RES conditions, no significant offtake unavailability on large-scale batteries is expected
between congestion and flexibility. It is a first iteration: a fuller assessment of transmission-level constraints during high renewable infeed is foreseen in the next FNA, and its conclusions may refine the picture presented here. It is a system-wide result; outcomes for individual assets can differ significantly depending on local conditions. On the distribution grid, RES curtailment resulting from constraints at TSO/ DSO connection points is limited to particular high-RES conditions and specific locations, remaining well below 1% of total renewable generation in each region in both 2030 and 2035. For the first time, Elia and the DSOs assessed these needs jointly — a basis for improved coordination on the steerability of renewable generation. This first assessment covers constraints at TSO/DSO coupling points; it does not yet quantify internal distribution congestion or voltage-related inverter disconnections, and the limited result should therefore not be read that network constraints are immaterial.
Distribution system scope
Network needs on TSO/DSO Coupling Point
Quantitative analysis on extra flexibility needs on distribution level following constraints on the TSO transformer at the TSO/DSO connection point
Figure 15: Summary of the results regarding grid constraints
During high RES conditions, distribution level curtailment is limited to specific conditions and locations Annual impact remains well below 1% of annual renewable generation levels
23
24 I ELIA I Beyond the Flexibility Needs Assessment
04
Delivering the PV Response identified: secure and extend the backstop while building future capability The FNA shows that in most scenarios — including those with ambitious assumptions on market-based flexibility — there will be periods when the system cannot deliver enough flexibility to operate safely. These are excess generation periods, identified as incompressibility risks. They arise because the pace at which decentralised PV becomes responsive to price signals does not follow the pace at which it is installed. Three layers can address this. The first is market-based response, which requires the barriers identified in section 5 to be removed. The second is downward flexibility products, which could unlock part of the PV response where justified. The third is a technical backstop. This section focuses on the third, for two reasons. First, the market layer depends on a long chain of actors — regulators, suppliers, aggregators, device manufacturers and ultimately consumers themselves — and on adoption by a large number of small asset owners; moving that mass takes many years, and the needs identified here materialise by 2030. The FNA elaborated a long list of market barriers to be relieved first to unlock this flexibility. Second, neither of the first two layers is designed for exceptional conditions. A capability activated on the order of an hour a year has no viable business case on activation revenue alone, and procuring additional downward capacity may reserve flexibility already available in day-ahead and intraday rather than add to it. Neither delivers the firm availability that a system-security measure requires. Nevertheless, Elia continues to work actively with market parties and the responsible authorities to ensure that the first and second layers further develop and unlock the full potential.
ELIA I Beyond the Flexibility Needs Assessment I
Options to unlock additional PV response
Required next steps
1
Market first: price-driven response Use to the fullest extent possible
Remove barriers and enable PV to respond to price signals
2
Targeted market mechanism If needed Under investigation
Elia will assess whether and how (new) downward flexibility products can efficiently unlock part of the PV response
3
Technical backstop (fail-safe) Last resort, when market response is insufficient
...the time to build the toolbox is now !
Anchor the backstop and extend the current volume as a no-regret step Federal: Establish regulatory framework. Integrate the backstop into the System Defense Plan Regional: Set legal, technical and technological controllability requirements
Figure 16: Summary of the proposed three levers
What is a technical backstop ? A technical fail-safe that would reduce the output of selected PV installations, activated only when market response is insufficient, under clearly defined and regulated conditions, so the system can be operated safely at all times. Its design should preserve, not displace, the incentive to develop market-based flexibility.
25
26 I ELIA I Beyond the Flexibility Needs Assessment
What is at stake if nothing changes It is worth being explicit about the alternative, because it is frequently misunderstood. Without a targeted ability to reduce PV output in exceptional conditions, the only tools left to Elia during a severe event would be to rely on radical measures which include disconnecting injecting distribution transformers — switching off entire neighbourhoods, consumers included, to resolve a surplus of electricity, or rely on the European obligation for all generators (including residential PV systems) that requires to automatically reduce their output. These options, while effective from a system security perspective, would lead to uncontrolled and undesirable situations for all stakeholders. A backstop that reduces the output of a targeted subset of PV installations in narrowly defined circumstances is a far more proportionate instrument: cheaper, better targeted, and without the societal impact of disconnecting communities. Similar reflections and developments are under way in other European countries, driven by a range of system needs. Germany, Poland, Czechia, Northern Ireland, Italy and Switzerland already allow remote curtailment of small commercial units and, in some cases, residential installations. Austria will follow from 2028, lowering the threshold to 3.68 kW. Other countries are also assessing how their backstop measures need to evolve, and ENTSO-E, the European association of transmission system operators, is calling for similar developments across Europe. It is worth noting that, while these countries have high shares of renewables, none has a residential PV penetration comparable to Belgium, which is the European leader in this respect. For the large majority of PV owners the backstop activation would pass unnoticed: the volumes are sized against a criterion of, on average, one hour per year, and the ability to consume electricity from the grid is not affected. Currently, Belgian system operators can rely on such a fail-safe mechanism but only on a minor part of the Belgian RES capacity and only made of larger systems. These systems are furthermore expected to react to market price reducing the fail-safe capabilities to contribute to exceptional conditions.
A Market-First Approach, Supported by Four Targeted Actions to cover Exceptional Conditions
00
TH E OVERARCH IN G PRIN CIPL E
MARKET FIRST
Flexibility (from demand, generation and storage) needs to be unlocked through the market. A market-first approach remains the cornerstone for unlocking flexibility from demand, generation and storage. By allowing market participants to react to price signals, flexibility can be developed and activated efficiently. For market-based response to materialise, this requires, among other things, addressing the barriers identified in Chapter 5 to further unlock flexibility from consumption, generation and storage at both transmission and distribution level. Nevertheless, the analysis shows that exceptional conditions may lead to residual flexibility needs that cannot be fully covered through market reactions alone. To address these situations, four complementary actions are proposed while maintaining market mechanisms as the primary source of flexibility.
ELIA I Beyond the Flexibility Needs Assessment I
01
02
Anchor the backstop in the System Defense Plan
Extend the current volume as a no-regret step
The last-resort backstop and other technical measures become part of the System Defense Plan
Lower the backstop threshold to 250 kVA, harmonised across Belgium. This excludes any residential installation
To make the backstop an effective measure, the necessary regulatory framework — including the right for the system operator to modulate, supported by appropriate technical standards and technological capabilities — must be established without delay. Concretely, this means two things.
Today, the PV volume covered by the existing technical measures2 — including installations above 25 MW — and therefore steerable as a last resort amounts to around 0.6 GWd. Based on the current findings in the FNA, it is no regret to further unlock capacity as part of the backstop.
Elia currently believes that the required regulatory framework should be established through a clear and robust federal framework anchored in the System Defense Plan, which is the legally recognised framework under European emergency and restoration legislation defining last-resort measures available to safeguard system security. A policy decision is needed on whether the ‘backstop’ should be integrated into this framework. If such a decision is taken, the System Defense Plan should clearly define the conditions under which the measure can be activated, the governance surrounding its use, and the safeguards ensuring that it remains a proportional and exceptional measure of last resort. Such a robust, transparent, and predictable framework is essential to provide legal certainty, clarify responsibilities, and ensure that system operators can act effectively when system stability is at risk.
Based on the current framework of the backstop and expected evolution of the PV fleet, the PV capacity in the backstop in 2030 would amount to around 1.1 GW (derated), compared to 0.6 GW (derated) today. The increase compared to today would be driven by new capacities entering under the current rules of the framework.
At regional level, the necessary legal and technical framework must be developed to operationalise the measures defined in the System Defense Plan. The regulatory framework should ensure that distribution system operators have the right to intervene on individual decentralised assets when required to implement the measures foreseen in the System Defense Plan.
To ensure sufficient steerability if the market falls short, Elia proposes bringing installations down to 250 kVA within the scope of the backstop across all regions. This would increase the derated PV capacity available as a last resort in 2030 up to 2.2 GWd depending on whether some existing installation will fall within the scope of the measure. Given the range of outcomes and uncertainties identified in the FNA, unlocking this volume is a proportionate, no-regret measure that strengthens system security while preserving the primacy of market-based flexibility.
In parallel, the necessary technical and technological requirements should be defined through the relevant regional regulations and technical grid codes, to ensure that decentralised assets are equipped with the capabilities needed for secured communication and steerability.
2
Other technologies are also part of the backstop but will slightly contribute to the challenge during the identified moments
27
28 I ELIA I Beyond the Flexibility Needs Assessment
Evolution of PV response needs compared to the installed derated PV capacity 20.9 GW
Expected amount of installed PV nominal (GWpeak)
16.5 GW 13.6 GWd
Expected amount of installed PV (derated)
12 GW 10.7 GWd
7.9 GWd
Total PV response 2 to needed (base 7 GWd 1-hour criterion) 1 to 5 GWd
0.6
2025
2030
2035
Current PV volume in the backstop Assuming a 65% derating to peak installed power (GWp)
Figure 17: Evolution of the incompressibility gap compared to the installed derated PV capacity
03
Steerability capability by design
Remote steerability becomes a connection requirement, enabling market reaction. Given the rapidly growing need for flexibility and the time required to update technical standards and deploy control solutions at scale, action should be taken now to ensure that future flexible assets are ready to contribute. All new installations able to modulate their generation, consumption or storage – including PV systems, batteries, heat pumps and EV chargers, and smaller units – should therefore be equipped from the outset with the technical capability to adjust their injection or offtake remotely. This capability should become a standard connection requirement in the relevant federal and regional grid codes and technical regulations, supported by secure, interoperable and cost-efficient control solutions. Embedding it at commissioning will be simpler and less costly than retrofitting a rapidly expanding fleet of devices later. Requiring this capability does not mean that every installation would be remotely controlled. Its primary purpose is to enable voluntary, market-based response, either directly or through suppliers and aggregators. Should additional steerable capacity nevertheless be required in the future, the functionality would already be in place and could be further unlocked.
04
Continuous monitoring of needs and levers
Track flexibility needs and available levers over time, and adjust the measures as the balance shifts. The need for flexibility—and the mix of resources available to provide it—will evolve over time. Key drivers such as renewable deployment, electricity demand and inflexible generation should therefore be monitored alongside the uptake of batteries, demand-side flexibility and marketresponsive PV. Because these developments can move independently, the residual need for PV response may be higher or lower than currently estimated. The measures should consequently remain adaptable rather than being fixed once and for all. Monitoring against transparent indicators at regular intervals should determine whether the backstop’s scope and size remain proportionate, need to be expanded, or can be reduced as marketbased flexibility develops. This approach would maintain adequate system-security assurance while avoiding unnecessary intervention or investment.
ELIA I Beyond the Flexibility Needs Assessment I
Flexibility requirements during incompressibility are linked to many uncertainties: Continuous monitoring is required Drivers of the need
Levers to cover the need (More) Demand-side flexibility
(More) RES penetration (mainly PV)
(More) Storage (small, large)
(More) Inflexible generation (e.g. nuclear)
EVs, heat pumps, industry – uptake uncertain
Deployment pace uncertain
Home batteries, large scale batteries amount V2G – adoption pace uncertain
Must-run – phase-out / extension timing uncertain
(More) Flexible share of RES (mainly PV)
Curtailable / steerable RES – limited reaction today, uptake uncertain
(Less) Electricity consumption Less electrification means more need
? Every lever moves independently – the balance must be continuously monitored and steered
Key uncertainties: pace of uptake, market barriers, costs, regulation, consumer behaviour, incentives...
CONTINUOUS MONITORING
29
30 I ELIA I Beyond the Flexibility Needs Assessment
05
More Flexibility, Through Collective Actions Unlocking market-based flexibility at scale is a first-order ambition. The scenarios already assume an ambitious development of demand flexibility and still leave a residual need, so every additional increment matters: the more the system can adjust through markets, the smaller the volume that has to be secured through other means. Unlocking flexibility at scale requires a coordinated effort across the entire energy ecosystem, since PV and other decentralised resources — batteries, electric vehicles, heat pumps and other distributed assets — still face a range of obstacles that keep them from participating fully in these markets. Several enablers have progressed since Elia’s earlier publications, but accelerating outcomes depends on stronger collaboration among all stakeholders in the flexibility value chain. The following five key challenges emerged from discussions with the different energy stakeholders.
ELIA I Beyond the Flexibility Needs Assessment I
1
MAKING IT EASY FOR CONSUMERS AND GRID USERS
Flexibility requires attractive and easy-tounderstand offers, clear communication and continued smart meter roll-out. Particular attention is needed to unlock the potential of electric vehicles and heat pumps. A first, and in many ways decisive, challenge concerns the end-user. Current consumer policy tends to prioritise bill predictability and simplicity, which inadvertently favours traditional tariff structures over the smart, flexible contracts that would let households monetise their own flexibility. As a result, flexibility remains unknown, too complex, or too risky for most people to engage with. Regional differences in the regulatory framework, compound this: suppliers must tailor their offers region by region, which fragments the market and prevents new products from reaching scale. However, a growing number of suppliers start to offer simple flexibility products alongside dynamic contracts, but uptake is still largely confined to early adopters, and even engaged consumers often see only limited value from participating, which weakens the case for changing behaviour. Moving beyond this early-adopter stage will require flexibility offers that are both commercially attractive and genuinely simple, adequate consumer protection to build trust, and sustained efforts – backed by regulators and public authorities – to raise awareness of what flexibility means and what it can concretely deliver for consumers. Finally, the electrification of heating and cooling, transport and industrial processes is a critical enabler of new flexibility. The right market and policy conditions must be created to foster the development of these new uses, and to ensure that the assets concerned are smart and can be flexibly controlled from the moment they enter the market.
2
GIVING THE RIGHT PRICE INCENTIVES
Price signals and tariff structures should reward consumers for shifting consumption to the right moment and support the growth of flexibility, while avoiding conflicting incentives between commodity and grid tariff components. A number of existing regulatory and tariff mechanisms currently discourage rather than encourage flexibility. Production support schemes can reward output regardless of whether the system actually needs it at that moment; arrangements that offset a household’s consumption against its own solar production blur the incentive to react to price signals; and flat or time-invariant grid tariffs give consumers little reason to shift consumption to the moments when the system needs it most. Suppliers and other market intermediaries will only invest in developing flexibility if the price signals consumers receive are strong, predictable and genuinely reflective of system needs. Addressing this calls for future support mechanisms designed to reinforce rather than undermine flexibility or grid tariffs that evolve toward structures rewarding welltimed consumption and production.
3
DELIVERING INFRASTRUCTURE AND INTEROPERABILITY
Common standards, interoperability between devices and platforms, and efficient processes for connecting flexible assets are key to enabling flexibility at scale. A third obstacle lies in the physical and digital backbone needed to make flexibility visible and usable at scale. The roll-out of smart meters remains incomplete in some regions, limiting the ability to measure and reward flexible behaviour, while the landscape of connected devices and communication protocols remains highly fragmented. This fragmentation drives up integration costs, erodes the business case for flexibility, and creates a lock-in risk that leaves consumers tied to a single provider or technology, unable to freely switch supplier or equipment. The way forward is to complete smart meter deployment where the flexibility potential is greatest, to make flexible devices interoperable by design rather than through costly case-by-case integration, and to establish open communication standards across the value chain.
31
32 I ELIA I Beyond the Flexibility Needs Assessment
4
ENABLING MARKET ACCESS
Open and proportionate rules are needed for aggregators and flexibility providers, including improvements to the Transfer of Energy model for day-ahead and intraday participation, simpler onboarding and prequalification, and scalable communication solutions. A fourth barrier concerns the quality of settlement data. Balance responsible parties are settled on a quarter-hourly basis and receive allocation data at the same time resolution. However, these volumes are not always based on the actual measured consumption of the individual connection points in their portfolio. This weakens portfolio management and forecasting, and gives balance responsible parties little reason to unlock flexibility within their portfolios. Basing allocation on actual measured quarter-hourly values would give them the right signals and make decentralised flexibility easier to account for.
5
ENABLING OPEN FLEXIBILITY MARKETS
Flexibility should be accessible and monetisable independently of the supply relationship, with open standards instead of exclusive partnerships and simpler, scalable participation rules for small businesses and households. A fifth challenge concerns flexibility services themselves which are often built on exclusive partnerships with specific equipment or software providers. It creates the very lock-in that open standards are meant to avoid. Meanwhile, neutral platforms capable of working with many types of devices still lack a straightforward way to bring that flexibility to market on their own. And although independent participation in flexibility markets exists in principle, the mechanisms that would make it work in practice remain too complex especially for small businesses and households. Progress requires maturing the frameworks that let flexibility be accessed and monetised independently of the energy supply relationship, and making participation simpler and more scalable so that independent flexibility trading becomes workable well beyond large industrial players.