About

MIRACA (Multi-hazard Infrastructure Risk Assessment for Climate Adaptation) is a research project building an evidence-based decision support toolkit that meets real-world demands.

This project has received funding from the European Union’s Horizon Europe research programme under grant agreement No 101004174.

Deliverable 2.3: Mapping lifeline network supply and demands to locations of other CI and societal actors. 

Overview

This program is part of the MIRACA Project – Deliverable 2.3, which focuses on mapping lifeline network supply and demands to locations of other critical infrastructure (CI) and societal actors. The energy modelling component of this framework is essential to integrating production units and consumers into the network model and preserving the robustness, scalability, and consistency of the representation of the energy system.

This work provides a methodology for assessing network resilience through the power flow analysis and is based on the core concepts of PyPSA-EUR (Hörsch et al., 2018). We achieve this by enhancing the technical precision and realism of supply-demand analysis through the use of open-source datasets, innovative mapping techniques, and fallback mechanisms. This framework provides a flexible and scalable model for understanding and improving the resilience of the European energy system through the use of dynamic demand modelling and high-resolution transmission grid data.

1. Importing High-Resolution Transmission Grid Dataset

• Data Source and Version: The network structures created from OpenStreetMap (OSM) are based on the prepared transmission grid (version 0.6), provided by Hörsch et al. (2025), obtained from the Zenodo repository.

• Key Features:

  • Coverage: It includes high-voltage AC/DC lines, substations, transformers, and converters.

  • Integration: The transmission grid is pre-processed in CSV format for easier integration into analytical workflows with less processing load and promoting grid reusability. 

• Application: This dataset is then imported into the Pandapower network structure and is used for power flow calculations and optimisation within the transmission grid model.

2. Load Profile Processing and Integration

• Source: Energy consumption and production profiles (solar and wind energy) are obtained from Open Power System Data, which increases the transparency and reliability of the historical data time series.

• Data Scope:

  • Total consumption and renewable energy production are captured on a per-country level at an hourly resolution.

  • The cleaning and handling of the missing value process is applied to ensure data completeness. Missing country consumption data is estimated using comparable countries and adjustment factors to improve the accuracy and realism of the model.

3. Demand Distribution and Region Definition

• Regional Granularity: Demand is assigned using high-resolution NUTS 3 regions collected from Eurostat GISCO datasets. Regions are then used to map electricity consumption based on economic (GDP) and demographic indicators. The methodology follows the PyPSA-EUR concepts.

• Load Distribution Method: In accordance with prior research (Hörsch et al., 2018), a 60/40 distribution key is used, developed from linear regression analysis of per-country data.

• Coverage Limitations: Some countries lack sufficient NUTS 3 granularity, potentially leading to misrepresentations of local load variations, including missing GDP values and populations of the NUTS 3 regions. In this case, the standard values are used based on country level and applied to countries with linear distribution.

4. Consumer-to-Substation Mapping

• Mapping Technique: A KDTree nearest neighbour search is used to link NUTS 3 regions to substations, ensuring effective assignment of loads.

• Priority and Fallback Mechanisms:

  • Primary Strategy:  Substations with the lowest voltage within the designated regions are given priority.

  • Alternate Mechanism: If no low-voltage option exists within a region, the nearest substation is identified based on broader geographical data.

• Options: The model allows users to choose between equal load distribution to all substations within the NUTS 3 region or using the voltage optimisation approach from the primary strategy.

5. Power Plant Data Integration

• Data Source: Power plant data are integrated from the “Powerplantmatching” tool.

• Mapping Method: A KDTree nearest neighbour search is applied to determine the closest network node, allowing effective spatial integration of generators.

6. Cost Data Integration

• Marginal Cost Data: PyPSA Technology Data provides marginal cost estimations for power plant data.

• Fuel and Emission Pricing:

  • Monthly fuel prices are sourced from the German Federal Statistical Office (Destatis).

  • CO₂ prices are sourced from the emission spot market auction report (EEX).

• Usage: These cost inputs are essential for correct DC optimal power flow calculations with economic dispatch.

7. Power Flow Modelling and Simulation

• DC Optimal Power Flow: Additional settings are implemented for DC optimal power flow (DC-OPF) calculations, ensuring that computational models are optimised for allocating supply and demand based on generation cost and network restrictions.

• Time-Series Simulation: Time-series simulations of power flow calculations are conducted over a user-selected interval. By default, the system computes the first 10 consecutive time steps or a single representative calculation for larger systems to avoid excessive computation.

• Flexibility and Efficiency: These simulation settings act as a safeguard against excessive processing and can be modified within the main program to suit different scales of system analysis.

8. Conclusion and Future Work

• Conclusion: This model demonstrates a reliable, scalable approach to integrating diverse datasets. The inclusion of fallback mechanisms and dynamic demand modelling contributes to enhanced accuracy in the allocation supply and demand.

• Future Improvements: Future work should aim to refine the precision of electricity consumption, generation, and network topology representations. Increasing the resolution and completeness of open-source datasets will further improve the realism and robustness of energy system modelling.

This comprehensive model not only integrates multiple layers of data but also introduces a different method to mapping supply and demand while also showcasing the simulation techniques. The methodological rigor and fallback solutions ensure that the system remains both adaptable and robust, paving the way for progressive improvements in energy system modelling.

Acknowledgments

This project utilizes the following datasets, scripts, and tools:

• Prebuilt electricity network for PyPSA-Eur based on OpenStreetMap data [1] from Zenodo repository (version 0.6): https://doi.org/10.5281/zenodo.14144752
• PyPSA-Eur scripts [2]
• OPSD time series dataset [3]
• Powerplantmatching database [4]
• NUTS datasets and territorial statistical units [5]
• geoBoundaries data for NUTS and non-NUTS regions [6]
• Pandapower electricity modeling tool [7]
• European Energy Exchange (EEX) Auctions dataset [8]
• Costs dataset from PyPSA [9]
• Gross Domestic Product (GDP) data by NUTS 3 region [10]
• Regional population data for GDP calculations [11]
• Data on energy price trends from Destatis [12]
• Extracted UK NUTS from Eurostat 2021 dataset [13]