4.1 Overview
This chapter describes the methodology applied to develop the demand inputs used in the TYNDP 2026 scenarios, including the Central Scenario (NT+) and LEV / HEV. As the market model simulations are performed on an hourly basis, the demand inputs must also be expressed as hourly time series.
Developing demand scenarios is a key part of the TYNDP 2026 Scenario Building process and follows the integrated toolchain described in Chapter 3. The approach is based on a bottom-up logic, starting from the collection of annual sectoral energy demand data coordinated between electricity and gas TSOs and subsequently translating these annual values into hourly demand profiles suitable for market simulations.
The information relevant for the simulation model covers demand for hydrogen, electricity, electric vehicles, hybrid heat pump (HHP) heat and synthetic fuels.
Figure 3 shows the overall demand-related part of the toolchain, from annual demand values established through TSO data collection to the hourly demand timeseries used as inputs in the market model. For illustration purposes, a single hourly profile per energy carrier is shown. However, the demand profiling tools generate differentiated hourly profiles by sector and use, which are subsequently aggregated to a single nodal demand profile per carrier in PLEXOS®. Further details on the modelling approach are provided in Chapter 8.
Figure 3: Demand data flow from annual values to hourly values
According to the Joint Data Collection Guidelines, demand inputs are to be provided by national TSOs and must be in line with the most recent NECPs. Where relevant, these inputs are complemented by the latest national policy measures and planning assumptions. This ensures consistency between the demand data used in the TYNDP 2026 Scenarios and the official national policy frameworks, while maintaining coherence with parallel European assessments, such as the European Resource Adequacy Assessment (ERAA).
In most cases, annual demand information is obtained from the ETM as described in Section 3.2. Where ETM data are not available or do not require completion, alternative data sources or dedicated data collection exercises are applied. Through this coordinated process, annual demand by energy carrier and sector is determined for each country.
Section 4.2 provides further details on the annual demand data, while Annex VI summarises country-specific data sources and modelling approaches.
Once annual demand values are defined, they are translated into hourly demand profiles following the procedures outlined in Sections 3.3, 3.4 and 3.5 and detailed further in Section 4.3.
Chapter 10 presents the methodology applied to derive the annual demand values for the economic variant scenarios, ensuring alignment and consistency with the Central Scenario.
4.2 Annual Demand Data
ETM countries and scope
The demand input data are jointly collected by gas and electricity TSOs through the ETM.
The ETM1 is a comprehensive, open-access energy system model to interactively build and explore energy system scenarios on country-base, covering all relevant sectors and energy carriers.
ETM covers all EU countries along with the United Kingdom, Norway, Switzerland, and Serbia. In this cycle most but not all these countries make use of the ETM as a demand data collection tool. For some countries a fallback solution is applied in the ETM. For a few other countries ETM was not used at all. They were treated like the countries not included in the ETM which submitted their demand input data via the PEMMDB demand file for 2030, 2035, 2040 and 2050 time horizons. For details, please consult Annex VI.
Starting point of any scenario created in the ETM is today’s energy balance per country and a database which reflects the technical and economic characteristics of the technologies applied in the energy system (e. g. a heat pump). In combination with a user-defined set of scenario assumptions for a specific target year (e. g. 2040), the ETM then projects the future use of energy. Within the TYNDP 2026 scenario building process, only the demand side modelling features of the ETM have been used and integrated into the overall scenario toolchain. The supply part and other sections of the ETM were not used in this scenario-building cycle. The ETM demand section contains the relevant end user (sub)sectors, wherein the user can define various scenario parameters. As an example: For the subsector space heating in the demand sector-built environment, the user can define which share of the heat will be provided by which technology (heat pumps, gas boiler etc.) which in turn determines how much energy of which carrier is being consumed.
On the output side, the model provides a set of data tables and charts to visualise and explore the outcomes of a scenario. All results can be exported for further use in other applications. The ETM can be accessed both online via a Graphical User Interface (GUI) or via an Automated Programming Interface (API), allowing more advanced users to interact with the ETM by scripts to upload data to or export data from the model. The model itself and a comprehensive documentation are available online via this webpage. The underlying energy balances for the reference year can be fully explored via the dataset manager.
For all the countries in scope of the TYNDP 2026 Scenarios and included in the ETM, the demand scenarios were defined by aligning on a set of parameters per each of the following sectors:
- Households
- Buildings
- Industry (energetic)
- Industry (non-energetic)
- Datacentres & Information and Communication Technologies (ICT)
- Transport (national)
- Transport (international)
- Agriculture
- Other
The defined parameters range from general assumptions (e. g. population growth) to more detailed aspects such as technology mix and specific technical characteristics.
The energy demand outputs were retrieved and subsequently processed to support validation, scenario comparison, target achievement assessments and data conversion required for integration with other scenario tools.
The main result of the data collection in ETM is the Final Energy Demand per country. ETM allows a detailed sectoral breakdown enabling the downstream Time Series Generation tools to make use of sector specific temporal characteristics of different subsectors.
1 ETM model is developed by Quintel Intelligence
ETM Fallback Solution
Some of the EU27 countries did not submit ETM demand data themselves but made use of the ETM fallback solution that was offered. The list of countries is given in Annex VI. For those 6 countries, the 2026 ETM values are set to the mean of the 2024 DE and GA scenarios for all four NT+ scenario target years.
Note that in 2024 cycle, the central NT / NT+ scenario is not available in ETM. Since the ETM model and especially its demand input parameters have evolved from the 2024 scenarios to the 2026 scenarios, a value mapping and interpolation process was required to ensure data compatibility and consistency.
Non-ETM Countries
For a third group, which consists of 2 EU countries and 11 non-EU countries (i. e. 13 countries in total), the ETM is not used to define demand data, either because ETM is not available for them or other constraints are faced. Instead, the ERAA 2025 electricity demand dataset is used as input for the e-demand.
Hydrogen demand was assumed to be zero. Thus, these countries are modelled with a reduced functionality and nodal topology. For the United Kingdom, Bulgaria and Cyprus hydrogen modelling is implemented even though a complete ETM data set is not available. For details, again refer to Annex VI.
Additional Data Collections
Following a comprehensive analysis of the scenario development process within the new regulatory framework, several data gaps were identified. To collect the remaining datasets necessary for scenario construction, multiple surveys were distributed to electricity and gas TSOs.
The additional data collections (DC2) covered the following topics:
— District heating, including industrial steam network and agricultural heating:
While ETM tool explicitly captures district heating demand and reports it under the “heat demand” category, additional information was required on the shares of primary energy sources supplying district heating. These shares, combined with efficiency factors, were necessary to calculate primary energy demand for district heating across all energy carriers.
— Demand and supply for synthetic and biofuels:
One further piece of information required to complete the data collection dataset was the total demand for synthetic and biofuels, together with its allocation between domestic production and other supply channels, such as exchanges or imports. The synthetic fuel production values identified for each country determined the corresponding additional hydrogen as well as the need for CO2 as a feedstock.
— Domestic ammonia production for shipping:
The ETM tool includes ammonia demand for both fertiliser production and shipping fuel. While the ETM model specifies whether ammonia for fertilisers is produced domestically, this information is not available for ammonia used as shipping fuel. Therefore, additional data were collected to determine the share of shipping-related ammonia produced domestically. The total ammonia demand was then converted into hydrogen demand, which serves as an input to the modelling tool, together with the additional electricity required for the conversion process.
— Electricity demand for CCS facilities that are not connected to a power plant (CCSnpp):
While CCS attached to power plants is inherently reflected in the plant efficiencies within PEMMDB data collection, an additional survey was required to collect national assumptions for CCSnpp and use them as an input to the scenario model. Furthermore, the assumed conversion efficiency of CCSnpp facilities was collected to determine the captured CO2 volumes relative to electricity consumption.
— Total CCS:
In addition to the CCS not connected to power plants, the data collection also covered the total CCS capacities on a country level as well as the share of CCS that stems from biogenic sources (BECCS).
— EVs survey:
This additional survey was carried out to collect parameters required for modelling the flexibility potential of passenger electric vehicles (EVs).
EVs can contribute to system flexibility by responding to market signals, primarily through smart or flexible charging strategies that optimise charging times. In this survey, TSOs were asked to indicate the share of EVs with fixed versus flexible charging profiles by selecting one of four predefined options for each target year.
If no values were provided, the default assumption applied was the “Balanced” option for all target years.
Additionally, TSOs were requested to define the capability of EVs with flexible charging profiles to provide Vehicle-to-Grid (V2G) services. For this purpose, the EV fleet was categorised into two groups:
— Home chargers: EVs primarily charged at home and acting as prosumers, connected to the prosumer node in the market model
— Street chargers: EVs primarily charged at public charging points, connected to the electricity market node
The predefined options also included trajectories for V2G participation across target years, differentiated between home and street charging configurations (see Table 5).
| Fixed Charging (%) | Optimised Charging (%) | |
|---|---|---|
| (DFT) | (PLEXOS®) | |
| Market Driven | 30 | 70 |
| Balanced | 50 | 50 |
| Users Oriented | 70 | 30 |
| Business As Usual | 85 | 15 |
| V2G (%) | 2030 | 2035 | 2040 | 2050 | |
|---|---|---|---|---|---|
| Home | Low flexibility | 0 | 5 | 10 | 50 |
| Medium flexibility | 15 | 20 | 25 | 35 | |
| High flexibility | 30 | 35 | 40 | 50 | |
| Street | Low flexibility | 0 | 1.5 | 3 | 5 |
| Medium flexibility | 0 | 3.5 | 7 | 15 | |
| High flexibility | 0 | 5 | 10 | 20 |
Table 5: Predefined charging behaviour assumptions and v2g trajectories used in the ev flexibility survey
If no values were provided, the medium trajectory was applied for both home and street charging EVs.
— Survey on Grid Losses
TSOs were asked to submit grid losses for any target year. The fallback solution was to use 3 % grid losses for all target years. For details see Annex VIII.
— Electricity Market Area Split
For countries with multiple market areas (nodes), the national-level demand figures were disaggregated across the respective nodes to ensure accurate representation within the market model. This distribution was based on detailed input provided by TSOs through dedicated questionnaires, which specified the relative allocation of the various demand components among nodes for each target year. As a fallback solution for e-demand, the split reported in the 2024 scenarios was used.
4.3 Hourly Demand Data
Overview
The following sections describe how annual demand values are transformed into hourly demand time series for hydrogen, electricity, thermal energy for hybrid heating and synthetic fuels. As the methane system is not modelled explicitly, methane demand time series are not generated.
Instead, ETM provides annual demand values for non-energy uses, while methane consumption in the energy system emerges endogenously from the model (on an hourly and yearly scale).
Hydrogen demand profiles
Hydrogen demand time series are computed for each hydrogen node and target year. The resulting hourly profiles represent exogenous hydrogen demand and reflect differences between countries, weather years (see Chapter 6) and target years. The differences arise from country-specific temperature assumptions and the composition of hydrogen demand across use cases, in particular the share of temperature-dependent demand such as demand for space and water heating.
The profiled hydrogen demand combined hydrogen demand from ETM and the Additional Data Collections: (1) Hydrogen demand (including sectoral breakdown) collected in ETM, with hydrogen demand for thermal energy removed, (2) hydrogen demand for use in hydrogen boilers in district heating networks from DC2 and (3) hydrogen demand for the production of ammonia for shipping from DC2. Hydrogen demand is not profiled for hydrogen-fired CCGT or OCGT plants, hydrogen use for thermal energy production in residential and tertiary hybrid heat pumps (see below), or hydrogen demand for synthetic fuel production (also see below), as these demands are determined endogenously within PLEXOS®.
Hydrogen demand profiles are constructed by sector and subsequently aggregated to form the total hydrogen demand time series for each node and year.
Residential and tertiary hydrogen demand is dominated by space heating and water heating, and therefore temperature dependent. The profile shape is derived by relating hydrogen demand to Heating Degree Days (HDDs) in a linear regression. This is done separately for space and water heating as the relationship to HDD is different.
In general, this is determined for a reference year and then varied according to the temperature differences between the reference year and the modelled weather year. Hydrogen demand for hydrogen boilers in district heating follows the same temperature-dependent profile assumptions.
Transport-related hydrogen demand consists mainly of hydrogen demand for heavy goods vehicles (HGVs) and for aviation. Hydrogen demand for HGVs is assumed to follow a flat hourly profile. This is because even though some countries impose driving restrictions on HGVs (such as Sunday bans), these effects are not modelled. Hydrogen demand for aviation is derived from historical aviation kerosene consumption profiles which are used to capture seasonal variation in flight activity. No assumptions are made regarding changes in demand patterns in future years.
Industrial demand is assumed to follow a flat hourly profile without seasonal variation. This assumption is consistent with standard gas system modelling practices, where industrial demand is typically represented at daily resolution, as short-term variability can be balanced through the use of line pack as short-term storage medium. This industrial hydrogen demand, as set out in ETM data collection, includes energetic uses, such as process heat, and non-energetic uses, such as direct reduction in steel production or feedstock use in the chemical sector. Hydrogen demand for the domestic production of ammonia for shipping is also assumed to be flat.
Electricity demand profiles
1) Grid Losses
Electricity grid losses are calculated based on the yearly electricity demand ETM yields and the shares of the losses the TSOs submitted (see Annex VIII). (ETM loss estimation is not used). The total yearly grid losses are then considered as part of the inputs to DFT for creating the electricity market nodes hourly demand profile.
Note that the efficiencies of power plants are modelled in PLEXOS® endogenously accounting for the self-consumption and the CCS if in place at the site.
2) Electricity Market demand time series are generated using DFT (Demand Forecasting Tool) for most of the countries.
DFT uses, for each market node, a data-driven modelling approach based on historical load timeseries combined with new technologies load profiles. Load timeseries are weather dependent, the weather conditions have been identified in Chapter 6 and Annex III and do not only affect electricity load but also renewable productions as well as other energy carriers. The weather years used are future climate projections and have hourly resolutions in order to be used. The timeseries automatically consider all the characteristics related to the consumption sectors (industry, services, residential and transport) of each specific country based on historical data. The technological evolution is as well considered through assumptions on expected penetration of Datacenters, District Heating and Transports (Part of Passenger EVs, Trucks, Vans, Buses, Trains).
Regarding Passenger Electric Vehicles, a share of the fleet is assumed to charge in public or street locations. This segment is split into:
- a user-oriented (non-flexible) charging component, represented by fixed daily demand profiles derived from Energy Transition Model (ETM) assumptions and profiles directly in the DFT. In the model it is considered as additional electricity demand time series for the e-market node.
- a market-driven (flexible) charging component, corresponding to EVs, that are able to optimise charging in response to electricity prices and charging infrastructure availability. This flexible charging demand is not included in the DFT profiles and is instead endogenously determined by the market model optimisation. See next chapter and Annex I.
The electricity demand associated with Heavy Duty Transport EVs (trucks, buses and vans) is fully embedded in the DFT electricity demand profiles and treated as exogenous load.
3) Prosumer node
Prosumer demand time series comprises only the consumption of fully electric heat pumps. Profiling is conducted according to climatic conditions and includes three distinct categories: heating, cooling, and domestic hot water usage. The profiling is carried out according to DFT. The prosumer node also models the consumption of hybrid heat pumps. These technologies, in addition to adapting to weather conditions, must meet heating demand by utilising either gas (hydrogen or methane) or electricity, depending on which option is economically preferable.
For Passenger EV home charging, two components are considered:
- a user-oriented (non-flexible) charging component, represented by fixed daily demand profiles derived from ETM assumptions and profiled directly in the DFT as additional electricity demand time series for the prosumer node.
- a flexible charging component, whose charging behaviour is optimised endogenously in the market model, following the same modelling approach adopted for other flexible EV charging segments.
Consistently with the overall EV modelling framework, only the flexible share of Passenger EV charging is represented explicitly in the market model, while fixed charging demand is treated as exogenous electricity load.
4) Behind-the-meter
Few market nodes have chosen to simulate behind-the-meter production separately. Residential rooftop PV systems combined with batteries are optimised independently over a 24-hour horizon. Batteries may only be charged from associated Photovoltaic (PV) generation, and each of these prosumer nodes only considers its own demand, without accounting for the rest of the system when optimising the battery behaviour. The sum of the PV generation and battery charging and discharging time series is attached to the corresponding prosumer node as a fixed load. It is a negative demand time series corresponding to a supply. For Germany only the battery time series is used as the PV capacities are modelled endogenously. The capacities used for this behind-the-meter approach are given in Annex VII.
Electric Vehicles Transport Demand
In the TYNDP 2026 scenarios, EVs are modelled differently depending on whether they belong to Passenger Cars or Heavy Transport categories (i. e. buses, trucks, vans).
In principle, both categories could be represented explicitly in the market model to capture the interaction between charging behaviour and electricity market prices. However, for the sake of modelling consistency and tractability, only EV Passenger Cars are modelled explicitly within the market model framework. The electricity demand associated with Heavy Transport EVs is therefore embedded in the exogenous electricity demand profiles and modelled in the DFT as additional electricity load.
To represent EV Passenger Cars charging in the market model, the annual transport demand (in kilometres) provided by the ETM must be converted into an hourly time series, as the market model operates at hourly temporal resolution.
This conversion is performed by applying driving profiles taken from the REM 2030 Driving Profiles Database2.
As shown in Figure 4, the driving profiles represent the share of weekly transport demand occurring at each hour of the day. To ensure consistency over a full week, the profiles are constructed such that total demand corresponds to five weekday profiles and two weekend day profiles, summing to 100 % of weekly transport demand.
For example, if a given hour (e. g. 08:00) represents 1.4 % of daily driving in the weekday profile, this value is counted five times (once for each weekday). Similarly, weekend hourly values are counted twice. Summing all hourly contributions across these five weekday profiles and two weekend profiles results in 100 % of weekly demand.
2 Fraunhofer ISI (2015). REM2030 Driving Profiles Database V2015. Fraunhofer Institute of Systems and Innovation Research ISI, Karlsruhe, Germany.
Figure 4: EV driving profiles – Share of weekly driving demand per hour [%]
For TYNDP 2026, the same driving profiles are applied:
- across all scenario horizons (2030–2050),
- across all countries,
- with a single distinction between weekdays and weekend days.
The annual EV Passenger Cars demand (in km) from the ETM is multiplied by these driving profiles to obtain hourly driving demand time series consistent with market model requirements.
As written above in the TYNDP 2026 scenario, EV Passenger Cars are divided into two main charging categories:
- EVs with imposed (non‑flexible) charging profiles, following predefined charging patterns;
- Flexible charging EVs, whose charging behaviour can respond endogenously to electricity market prices.
Only flexible EV Passenger Cars are modelled explicitly in the market model. Consequently, the annual transport demand converted into hourly profiles using the driving profiles refers exclusively to this flexible EV segment. The electricity demand associated with EVs under imposed charging profiles is treated exogenously and is either calculated within the DFT or directly provided by TSOs.
Further details on the definition and modelling of flexible versus imposed charging are provided in Section 8.3.
Thermal energy profiles for hybrid heating
Thermal energy demand profiles define the hourly distribution of thermal demand to be met by HHPs in each target and weather scenario. Thermal energy may be provided by an electric heat pump or a gas boiler. Two sets of profiles were constructed: one for methane HHPs, one for hydrogen HHPs. It is a model outcome whether this thermal demand is satisfied using electricity or the respective gas.
Similar to the approach previously described for hydrogen demand, the profile shape is derived by relating thermal energy demand temperature in a linear regression. This relationship is determined for a reference year and then varied according to the temperature differences between the reference year and the modelled weather scenario.
Synthetic fuels
As described in Section 8.4, demand profiles for the EU27-aggregated e-liquids and SNG nodes are assumed to be flat over the year. This simplification is justified by the relatively small share of synthetic fuels in the overall fuel system, which does not warrant a more granular temporal representation, and by the typical operating pattern of synthetic fuel production plants. Synthetic fuel synthesis is a continuous chemical process, for which an economically optimal operation is generally based on near constant production (and corresponding consumption) 24 hours a day, 7 days a week.
In practical terms, the annual synthetic fuel demand figures from the additional data collection are divided by 8,760 hours to obtain a constant hourly demand. This yields static demand profiles that are independent of weather variability for all synthetic fuel types.
