This README provides an overview of the GeoMEC multi-carrier geolocation planning model developed within the iDesignRES horizon project.
The model is developed, further enhanced and handled by E3-Modelling, part of Ricardo (E3M), under Work Package 2, Deliverable 2.3.
As part of the 3-layer modelling strategy proposed in the iDesignRES project, the geolocation multi-carrier model (GeoMEC) is positioned within Layer 2, corresponding to the NUTS-2 regional level. This middle layer plays a pivotal role in linking the strategic, long-term decarbonisation pathways developed at European scale with the detailed, spatially explicit reality of regional energy systems.
In the iDesignRES architecture, the Pan-European optimisation model in Layer 1 produces long-term decarbonisation scenarios that establish the boundary conditions for all subsequent layers. These boundary conditions include projected energy service demands, macro-level technology deployment trends, fuel-price trajectories, CO₂ budget pathways, and system-wide policy constraints. GeoMEC then translates these high-level outcomes into regionalised, spatially detailed energy system configurations. Specifically, the model determines the optimal allocation of production, conversion, storage and network infrastructure capacities across each NUTS-2 region.
GeoMEC simultaneously covers several energy carriers, electricity, heat, hydrogen, gas, CO₂ flows and e-liquids and identifies the most cost-effective spatial deployment of technologies needed to satisfy projected demand for these carriers. Its allocation logic accounts for the techno-economic characteristics and the distinguishing features of each region, including:
• renewable resource potentials (solar, wind, geothermal, biomass) and land availability and regulatory constraints,
• spatial distribution of demand,
• location of industrial clusters and process heat requirements and
• existing infrastructures (electricity networks, gas grids, DH systems).
This regional tailoring is essential, as the composition of demand, industrial activity, and available resources varies widely across Europe. The novelty of GeoMEC lies in its ability to quantify the effects of sector coupling and the diverse technological elements constituting interconnected multi-carrier energy networks, while retaining a level of spatial and technological resolution that remains computationally tractable. Unlike traditional models that focus on either a single energy vector or on representing the system at a coarse administrative scale, GeoMEC integrates:
• multi-vector conversion chains (power-to-heat, power-to-hydrogen, hydrogen-to-power etc.),
• cross-carrier flexibility options (batteries, pumped storage, hydrogen storage),
• geospatially explicit deployment constraints,
• infrastructure optimisation, including network expansion
The core model of GeoMEC is written in GAMS (General Algebraic Modelling System), using its environment to define sets, parameters, decision variables, equations, and the objective function that formulate the linear or mixed-integer linear problem. These are then solved in the same framework by the optimisation software packages CPLEX and Xpress, although alternatives such as HighS may also be used. Python scripts, including toolkits and libraries that streamline input and output connections to the GAMS model, such as GAMSPy and gamsapi, are used for input data routines and for appropriate handling of results for analysis.
Inputs to GeoMEC
The input data routine has been developed to streamline all available data sources, convert data into a suitable form for the main model, and translate data to the IAMC common definition protocol, where applicable. This routine consists of dedicated Python scripts which collect, parse and disaggregate to NUTS2 level if necessary, the different data files used as inputs for the model. Where data from upstream models is currently unavailable, but expected to be in the next stages of development, suitable placeholders are in place.
Outputs of GeoMEC
GeoMEC provides results at a NUTS2 level and its outputs include:
• Regional distributions of generation, storage and hydrogen production capacities and operation
• Spatial investment in electricity, hydrogen and CO2 transmission corridors, along with the cross-border flows
• Identification of regions with potential infrastructure pressures, based on utilisation levels or renewable integration constraints
• High-level infrastructure adequacy signals, useful for informing dispatch feasibility in more detailed operational models
GeoMEC is formulated as a linear (LP), or mixed integer linear (MILP), optimisation problem solved under perfect foresight, meaning that the model simultaneously determines optimal long-term capacity investments and short-term operational dispatch decisions for the entire modelling horizon, which covers the period from 2025 to 2060, using a 5-year time step. The model minimises total discounted system costs for expansion and operation, while ensuring that energy service demands across carriers and regions are met under all relevant technical, operational and policy constraints.
The choice between an LP and MILP formulation is left to the user, depending on the level of operational detail required and the available computational resources. When operating in LP mode, a linear relaxation of unit commitment constraints is applied. In this case, discrete on/off decisions are approximated through continuous variables and constraints related to minimum unit size, start-up and shut-down behaviour are relaxed. This formulation focuses on capturing the main operational limits and cost drivers, while reducing computational complexity and is suitable for large-scale, multi-region capacity expansion analysis and long-term scenario exploration. When operating in MILP mode, the model enforces a more detailed representation of operational behaviour, including stricter unit commitment constraints and discrete investment decisions reflecting specific technology sizes. This allows for a more explicit treatment of operational feasibility but comes at a higher computational cost. The two formulations therefore offer a trade-off between model detail and tractability.
In terms of the investment decision, the model follows a perfect foresight formulation, in line with the approach adopted in large-scale energy system models. Investment and operation decisions are optimised considering the full-time horizon, including future demand development, technology evolution and policy constraints. To keep the computational burden at manageable levels, the user is able to select that the perfect foresight formulation is implemented using a rolling-horizon structure, whereby the optimisation is solved over successive time blocks while maintaining consistency of long-term investment decisions across periods. This allows the model to represent intertemporal trade-offs and long-term system dynamics without solving the entire horizon in a single optimisation run. The rolling horizon feature is under testing and will be incorporated soon.
GeoMEC can be found in the following repository https://github.com/e3modelling/GeoMEC