Modelling Methodology

Underlying Philosophy

We have attempted to design this model using Zellner’s KISS philosophy: Keep It Sophisticatedly Simple.

Two pitfalls when attempting to model the energy system of the distant future are that enormous complexity and lack of transparency prevents other from gaining value from the model, and that the modeller falsely believes that it is possible to generate meaningful answers. To quote Huntington, we “model to generate insights, not numbers”.

The ETRC model aims to be readily understandable and to give believable outcomes. Broad simplifications are used throughout, as we are not trying to convince ourselves that an accurate projection of life in 2050 is possible. However, the model must advance beyond previous forecasts in terms of geographical resolution and the number of energy vectors considered.

There is a wealth of existing studies of energy in 2050, so we wish to utilise previous studies wherever possible, rather than creating an altogether new forecast which lacks the pedigree and heritage of the IEA Outlook or Shell Scenarios (to name just two).

A key distinction to keep in mind is that this model is a tool for exploring a pre-existing or user-generated scenario for the future; it is not a means of predicting what will happen in the future. The model is not designed to meet any explicit targets or enforce any explicit constraints. It will not search for an optimal mix of technologies that bring about an 80% reduction in CO₂ emissions, or decide how to distribute renewable generators around the continent: these are tasks for the user, as they have been done many times before. Instead, the model allows you to explore the finer details of an 80% reduction (or other scenario of your choosing), investigating the patterns of electricity supply and demand, and the level of transfer that would be required across the continent of Europe.

Summary of the Model's Approach

The modelling process starts from the energy services that people will demand in future, which together with predicted technology shares and efficiencies drive the demands for different forms of energy. This ‘big picture’ projection of the whole economy is made using annual numbers, from which the model drills down to hourly timescales. Year-long hourly profiles of electricity demand in each country are stochastically built up from individual sectors’ demands, supply from renewables is drawn from historic reanalysis data, and supply from conventional generators is simulated using a least-cost optimisation model. Put together, this paints a picture of electricity supply, demand, prices, and transmission between countries.

The model's operation can be summarised by four input stages, four calculation stages and two result generation stages:

1. The model is initially calibrated with energy demands, service demands and technology parameters for a reference year. Data for 2010 from the IEA, World Bank and Eurostat is provided by default.

2. You begin by creating a scenario for the future demand for energy services (such as thermal comfort, economic output, movement of people and goods). Given the large number of sectors and parameters, scenario can be derived from broad macro-economic parameters (population and economic growth, sectoral shifts, etc.).

3. You then provide assumptions about the mix of technologies that will provide these energy services (e.g. conventional cars, hybrids, pure electric, hydrogen, and so on) and their technical efficiency. Pre-defined and user-defined scenarios can be entered into the model to control these aspects easily.

   1. Finally, you supply the installed capacity of different types of power station in 2050 for each country, including conventional and CCS fossil plants, nuclear and renewables (wind, solar, hydro and tidal), and the capacity of transmission interconnectors between regions.

4. The model uses your service demands and technology parameters to calculate the 2050 final energy demands for ten key energy vectors on an annual basis for the forty countries of Europe and North Africa.

5. The model then stochastically synthesises a one-year profile (8,760 hours) of electricity demand in each country, taking this annual base data and building up hourly profile sector by sector. This relies on stochastic weather simulation and one-day profiles for different types of end-use (electric heating, domestic appliances, electric vehicle charging, etc.) which are pre-defined, and can be altered by the user.

7. The supply of electricity from renewable generators is determined from hourly profiles of the available resource in each country, which have been derived from high temporal- and spatial-resolution reanalysis modelling.

8. The modelling is completed by simulating production of the remaining net demand from conventional generators and cross-border trading. This uses a least-cost optimisation model, which finds the simultaneous short-run equilibrium across all countries, thus giving the optimal operating patterns for the supplied set of plant. This can go on to solve the long-run equilibrium, giving the optimal capacity of each type of plant to install in each country.

9. The initial results produced by this final modelling stage give installed plant capacities, operating patterns, hourly electricity cost and price, and the hourly power flow down each interconnector; which in turn give the CO₂ emissions from the electricity sector per country, or the profits by generator type and transmission asset, for example.

10. The annual fuel consumption from the electricity sector in each country is then then incorporated back into the annual model, allowing the final (end-use) demands to be translated into demand for primary energy sources, whole-country CO₂ emissions, and so on.

The model’s temporal resolution is:

• Annual for fuels that can be readily stored (coal, oil, biomass);

• Monthly for natural gas and electricity, giving the seasonal pattern;

• Hourly profiles for electricity over a year.

We consider two levels of geographic resolution. Data for 40 countries across Europe and Northern Africa is presented at the individual country level. The model also aggregates data into ten regions (pictured right), to give a broad overview of the continent, and to simplify the creation of a transmission network during the first iteration. We follow the ECF Roadmap 2050 study in their grouping of European countries into nine regions, adding a tenth for our supplying neighbours in Northern Africa and Iceland. The electricity supply and demand modules are equally capable of running with the 40 individual countries.