On 29 June, Italy's grid asked for more electricity than on any other day of the year: 57.4 gigawatts, chasing a heatwave that had already made the month the hottest ever recorded in western Europe. Two weeks later it broke its own record again — nearly 58 gigawatts, air conditioners humming in unison from Palermo to Turin. Somewhere between Milan and Brussels, grid operators were watching demand climb with the mercury, renewable output swing with the clouds, and import flows priced against a war that refuses to end.

There was no single needle to watch. Picture instead a scale built with three arms, not two — climate neutrality, energy security, industrial competitiveness — where lowering any one sends the other two swinging out of reach. This is the "trilemma" Europe's energy experts will carry to Brussels on 21–22 October, when they convene for the European Climate and Energy Modelling Platform (ECEMP) — this year titled "Climate Change, Energy Security and Competitiveness: Modelling an energy system under stress". There, they will ask a question with no comfortable answer: can tools built to optimise for one variable be trusted to hold three at once?

When reality escapes the models

"We are in the heart of Europe, close to the actual policymakers," says Erik Delarue, who leads the Energy Systems Integration and Modelling group at KU Leuven's Department of Mechanical Engineering. He frames ECEMP around the energy trilemma: "cost-optimal solutions and cost competitiveness; security of supply and resilience; and sustainability." "Ten years ago, focus was strongly on CO2 emissions," he recalls — Europe had long been a frontrunner there. "But in the past couple of years, these two other corners have again become more prominent," pushed forward by the war on Europe's doorstep and its fallout on industry and supply. Furthermore, the difficulty is that some of the factors now shaping Europe's energy system are precisely those that models struggle to capture. One is human behaviour. "People don't buy a car, or decide to install solar panels, or choose how to charge an EV, purely as rational optimisers," Delarue points out — pushing researchers like him toward among other agent-based modelling that simulates real behaviour instead of assuming it.

The drivers of the system are also evolving. One again showed up this summer, Delarue notes: "Weather has become a very influential factor — during summer evenings, you for instance get a very steep ramp in the residual load — and that shows up directly in very volatile electricity prices." Europe's other dependency runs deeper than any heatwave. "The EU genuinely needs to find a medium-term strategy for natural gas," says Pedro Crespo del Granado, energy economist at NTNU and research coordinator on iDesignRES, one of ECEMP's co-organising projects. But the models planning that shift, he argues, "assume a perfect market, full cooperation, which is not the case." A deeper gap looms further away: "By 2035, maybe 2040, the gas infrastructure that is in place today in Europe will struggle to become economically viable" — yet "the models will not tell you that, because the models will not increase the price of gas to keep the infrastructure running."

The same pattern runs through the social side of the transition. Planning "has the assumption that it's going to be a wider and strong collaboration among all European nations," Crespo del Granado notes — implying rapid grid deployment that shows up everywhere in model results. What doesn't show up is public acceptance of that expansion. "They never account for the effect of social implications of this rapid transmission expansion, which means change in regulations, political parties should agree across governments." Comparing scenarios with and without grid expansion, he argues, is often the only way that hidden cost becomes visible.


Modelling a system at war

Nowhere is the gap between model and reality starker than in Ukraine, where iDesignRES runs a case study on planning an energy system under wartime conditions. "The main thing we have learned is about accelerating methods to analyse war modelling impacts, for example drone waves," Crespo del Granado reports. "The Ukrainian system might know a few hours in advance the first wave or the second wave of drone attacks, and they respond already hours before with reconfiguring blackouts, reconfiguring the grid." What surprised his team more came from stakeholder workshops: asked what to prioritise, Ukrainian counterparts said "accelerate the grid interconnection and get market coupling with Europe as soon as possible." The questions that followed went further still, he adds: "What does it imply for Europe and Ukraine? And how can regulation itself move faster, using the knowledge these models produce?"


The right model for the right question

But acknowledging these blind spots does not necessarily mean asking a single model to optimise climate, security and competitiveness simultaneously. Delarue clarifies: "We should see a model as a tool, to gain insight in certain matter. Different questions might require different modelling approaches." Then, he adds: "What you usually have is a cost metric that the model optimises for, while many of these other considerations are handled as constraints rather than as objectives in themselves." Emissions get capped; security is built in through safety margins and contingency planning. "A multi-objective optimisation — a single function weighing cost, emissions and resilience together — is possible too," Delarue continues. "Next to approaches like modeling-to-generate-alternatives, both can be used to explore the solution space around the strict cost-optimal solution."

The challenge, then, is not simply to make one model do everything, but to make sure the right models are used for the right questions — and can work together when those questions overlap. While Delarue and Crespo del Granado expose the limits of today's models, Niina Helistö of VTT turns to possible solutions. She coordinates Mopo, a European project developing a toolset to plan sustainable and resilient energy systems in a cost-effective manner, while also exploring how different energy models can be better connected. "If a model is used to answer a question it wasn't built for, it might give an answer, but it can be misleading because the modelling assumptions don't fit the purpose," she cautions. Climate change and variable renewables, she notes, have pushed the field toward more comprehensive models and, for the same reasons, toward coupling several of them — with a balance still to be found between the two. That connective approach is what her project has spent its final year testing, and its demonstration cases speak directly to the trilemma. In the Baltic countries, which in February 2025 synchronised their grids with continental Europe after cutting their last ties with the Russian and Belarusian networks, the question is security.

The new set-up still demands careful planning of how the system is designed and operated, Helistö points out, while keeping the whole European system in view — hence the link between pan-European modelling and detailed operational modelling of the region. In the industrial cluster spanning the Netherlands and Belgium, the question is competitiveness and climate at once: industries are looking for cost-effective ways to decarbonise, in line with European energy and emission scenarios. The project therefore tests what the pan-European perspective adds to the analysis of the cluster — a way, she says, to "demonstrate the trade-offs with sectoral and geographical detail and scope." Without shared interoperability standards, Helistö warns, "model couplings are much harder and prone to errors." A year on, the project's tools have moved beyond the research team that built them — into university courses, PhD projects, public agencies and, increasingly, commercial work.

Connecting models, not eliminating trade-offs

Yet connecting models does not eliminate the compromises built into them. Ask Crespo del Granado what will shape the next decade more — better models or better politics — and he names one of the hardest problems without hesitating: prices. "Understanding the impact on prices is difficult to inform and difficult to assess," he acknowledges. "In all the models that we use, they all cut corners somewhere — when you're analysing these prices, that leads to a blurry story for policymakers to act on, for example in market design and incentives."

Back on the grid, none of that uncertainty pauses for the heat — and the WMO now puts the odds of an intensifying El Niño persisting into 2027 at near 100%. Add a geopolitical landscape still unsettled by the war in Ukraine and beyond, gas supply bottlenecks and rising costs, and the trilemma remains unresolved, its scale still with three arms.

What Mopo and other European initiatives are testing isn't a way to finally balance it, but a way to see more clearly how the choices made on one side affect the others — before the next shock puts the system under pressure again.

Photo by Amorie Sam on Pexels


Contacts:

Coordinator
Niina Helistö, VTT
info@tools-for-energy-system-modelling.org

Communication manager
Erika Novellini, ICONS
info@tools-for-energy-system-modelling.org

Project website: https://www.tools-for-energy-system-modelling.org
LinkedIn: @Tools for Energy System Modelling  
Mastodon: @MOPO@mastodon.energy
Bluesky: ‪@mopoproject.bsky.social