Skip to content

Assumptions

This section outlines the key modeling assumptions that differentiate the Google-GO project from the standard PyPSA-Eur model. While the project builds upon the PyPSA-Eur framework, several assumptions have been introduced to better fit the project scope.

The modeling assumptions are organized into four main groups, each addressing a specific aspect of the energy system that requires customization beyond the default PyPSA-Eur configuration:

  1. Electricity Demand: Assumptions related to the demand, profile, and evolution of the electricity across different sectors
  2. Clean Firm Technologies: Assumptions related to the techno-economic characterization of dispatchable clean generation technologies
  3. CO2 Price: Assumptions related to carbon pricing mechanisms
  4. Generation Expansion: Assumptions related to the capacity expansion of clean and conventional generation technologies

1. Electricity Demand

To focus on the electricity sector, demands from low-voltage, industry, agriculture, heating, and hydrogen are aggregated into a single electricity load. Transport demand remains separate due to its dependency on EV batteries.

The electricity demand framework combines default PyPSA-Eur assumptions with TYNDP-2024 projections to compensate for the absence of full sector coupling.

Specifically:

  • PyPSA-Eur: provides demand for low-voltage, industry, agriculture, and transport
  • TYNDP-2024: provides demand for heating and hydrogen

Within the aggregated electricity demand, commercial and industrial (C&I) demand - which participates in the GO market - is further characterized using data from Eurostat and IEA to determine its share and load profile.

1.1 Heating and Hydrogen Electrification

Both demand components are computed following a consistent methodology:

  • Derived from TYNDP 2024 results
  • Calculated as a share of net final electricity demand excluding transport (since transport is modeled separately in PyPSA-Eur)
  • Based on the National Trends scenario (aligned with most recent policies)
  • Applied to the PyPSA load, neglecting profile differences

Table 1 - Heating- and hydrogen-related electricity demand shares.

Demand Component 2025 2030 2035 2040
Heating 16% 16% 16% 16%
Hydrogen 4% 7% 14% 21%

Heating Demand:

  • Includes space heating, cooling, and hot water needs in buildings and households
  • The source provides only 2019-Historical and 2040/2050-Deviation Scenarios. Therefore, 2019 data remain constant across the planning horizon
  • These shares do not account for hydrogen-related demand at the denominator

Hydrogen Demand:

  • Includes electricity demand from all electrolyzers
  • The source (TYNDP 2024-Scenarios Report-Data and Figures) provides 2030 and 2040 National Trends data. Values for 2025 and 2035 were interpolated assuming zero demand in 2020
  • These shares include heat-related demand in the denominator and should only be used when heating demand is included. The code allows separate handling if needed

Note: By default, only heating demand is added. To activate the hydrogen demand assumptions, append this in config/config.go.yaml.

overwrite_years:
  2025:
    electricity:
        hydrogen_demand:
            enable: true
            share: 0.04

  2030:
    electricity:
        hydrogen_demand:
            enable: true
            share: 0.07

  2035:
    electricity:
        hydrogen_demand:
            enable: true
            share: 0.14

  2040:
    electricity:
        hydrogen_demand:
            enable: true
            share: 0.21

For implementation details, see config.go.yaml and strip_network.py (described in Project Config and Strip Network).

1.2 Commercial and Industrial Demand

Demand from commercial and industrial customers is derived from the C&I share over total electricity consumption as follows:

  • Eurostat is the main source
  • For United Kingdom (GB) and Switzerland (CH), which are absent from Eurostat, IEA energy balance statistics are used
  • Latest available data: Eurostat-2023, IEA-2022
  • C&I shares remain constant across the time horizon, excluding potential future changes (e.g., data center growth or deindustrialization). Users can manually modify these values

For implementation details, see config.go.yaml (described in Project Config).

C&I share Figure 1 - 2023 C&I share over the final electricity consumption (Source: Eurostat).

2. Clean Firm Technologies

Two types of clean firm technologies are considered:

  • Green Open Cycle Gas Turbine (Green OCGT): low-CAPEX, high-OPEX technology using green hydrogen as fuel
  • Advanced Clean Firm: capital-intensive technology serving as a proxy for advanced nuclear or advanced geothermal

Data Sources:

The techno-economic parameters are derived from multiple sources:

  • Investment cost, fixed operation and maintenance (FOM), variable operation and maintenance (VOM), and lifetime: Lazard LCOE - June 2025. Where ranges were provided, the mean values were used
  • Fuel cost and efficiency: PyPSA Technology Data
  • Green hydrogen price: EU Clean Hydrogen Observatory. The source provides 6.71 €/kg in 2024, which was assumed to decrease up to 5 €/kg by 2030 and 3 €/kg by 2040.

Currency Conversion:

Lazard costs are reported in $2025, while PyPSA-Eur uses €2020 as reference currency. The conversion was performed in two steps:

  1. Dollar-to-Euro conversion using the 2025 exchange rate
  2. Inflation adjustment to €2020 using Average annual Harmonised Index of Consumer Prices (HICP) - EU27

Cost Processing:

The input data were processed to derive the two input cost parameters used in PyPSA-Eur, i.e., annualized capital and marginal costs. For Green OCGT, fuel costs available for 2030 and 2040 were interpolated to obtain values for 2025 and 2035, covering all planning horizons. For Advanced Clean Firm, all values were kept constant across the planning horizon.

The final cost parameters are computed as follows, where the annuity factor is calculated following the standard PyPSA-Eur approach (see process_cost_data.py), using the lifetime and a discount rate of 7% (\(r\) is the discount rate and \(n\) is the lifetime in years):

Note: 7% is the default in PyPSA-Eur for long-terms rate of returns.

\[ \text{annuity factor} = \frac{r}{1 - \frac{1}{(1 + r)^n}} \]
\[ \text{annuity_factor_fom} = \text{annuity factor} + \frac{\text{FOM}}{100.0} \]
\[ \text{Annualized Capital Cost (€2020/MW-a)} = \text{annuity_factor_fom} \times \text{investment cost} \]
\[ \text{Marginal Cost (€2020/MWh)} = \text{VOM} + \frac{\text{fuel cost}}{\text{efficiency}} \]

Table 2 - Techno-economic characterization of clean firm technologies.

Investment Cost ($2025/kW) FOM ($2025/kW-a) VOM ($2025/MWh) Fuel (€2020/MWh) (€2020/kg) Efficiency Lifetime Annualized Capital Cost (€2020/MW-a) Marginal Cost (€2020/MWh)
Green OCGT 1,150 – 1,450 10.00 – 17.00 3.50 – 5.00 '30: 150.0 (5.0)
'40: 90.0 (3.0)		 0.40 30 72,790 '25: 444.7
'30: 371.1
'35: 297.5
'40: 223.8
Advanced Clean Firm 9,020 – 14,820 136.00 – 158.00 4.40 – 5.15 3.41 0.33 70 585,153 13.6

For implementation details, see config.go.yaml (described in Project Config). For details on the techno-economic characterization of the other power plants and storage technologies, see PyPSA Technology Data.

3. CO2 Price

Four CO2 price levels are identified to study their interaction with GO markets as sensitivities:

For implementation details, see config.go.yaml (described in Project Config).

4. Generation Expansion

The assumptions on extendable generators follow the default PyPSA-Eur configuration (for details, see config.default.yaml and PyPSA-Eur documentation). Extendable technologies include:

  • All solar and wind types
  • Open Cycle Gas Turbine (OCGT) and Combined Cycle Gas Turbine (CCGT)
  • Clean firm technologies described in Section 2

Additionally, renewable generation targets are derived from TYNDP-2024-National Trends scenario results. These targets are designed for use in the Renewable Portfolio Standard (RPS) sensitivity (see Scenarios Config) as minimum generation constraints relative to total electricity generation. For implementation details, see config.go.yaml (described in Project Config) and Solve Network Constraints.

Such shares are defined for each planning horizon and country, including system-level (denoted as EU+). Values are taken from TYNDP-2024 National Trends scenario, which provides data for 2030 and 2040 only. Therefore, 2035 values were interpolated. Kosovo is the only country absent from TYNDP data, for which a 43% target is imposed for all years based on 2030 value from an Ember study. (TYNDP and Ember are well aligned, e.g., system-level targets at 2030 from both sources are essentially the same ~73%).

Table 3 - RPS at country and system (i.e., EU+) level.

Country 2030 2035 2040
System (EU+) 0.74 0.78 0.83
Albania (AL) 0.93 0.96 0.99
Austria (AT) 0.92 0.92 0.92
Bosnia (BA) 1.00 0.92 0.85
Belgium (BE) 0.58 0.66 0.74
Bulgaria (BG) 0.34 0.33 0.32
Switzerland (CH) 0.82 0.90 0.98
Czechia (CZ) 0.33 0.34 0.35
Germany (DE) 0.91 0.92 0.94
Denmark (DK) 0.84 0.89 0.94
Estonia (EE) 0.69 0.80 0.92
Spain (ES) 0.82 0.88 0.94
Finland (FI) 0.77 0.83 0.89
France (FR) 0.36 0.44 0.52
Great Britain (GB) 0.83 0.82 0.82
Greece (GR) 0.71 0.81 0.91
Croatia (HR) 0.92 0.90 0.88
Hungary (HU) 0.29 0.32 0.35
Ireland (IE) 0.89 0.91 0.93
Italy (IT) 0.72 0.78 0.85
Lithuania (LT) 0.94 0.96 0.97
Luxembourg (LU) 0.95 0.96 0.96
Latvia (LV) 0.89 0.89 0.89
Montenegro (ME) 1.00 1.00 1.00
North Macedonia (MK) 0.47 0.57 0.68
Netherlands (NL) 0.81 0.86 0.92
Norway (NO) 0.99 0.99 0.99
Poland (PL) 0.64 0.60 0.56
Portugal (PT) 0.96 0.98 0.99
Romania (RO) 0.54 0.58 0.62
Serbia (RS) 0.65 0.67 0.69
Sweden (SE) 0.77 0.83 0.89
Slovenia (SI) 0.49 0.53 0.58
Slovakia (SK) 0.28 0.32 0.35
Kosovo (XK) 0.43 0.43 0.43