Why industrial sites can't decarbonise without a digital twin of their energy system

The pressure on industrial manufacturers to decarbonise is no longer a distant strategic priority. It is arriving in the form of carbon border adjustment mechanisms, tightening EU ETS obligations, Scope 1 and 2 reporting requirements, and procurement conditions from customers who have made their own net–zero commitments.

For energy–intensive industries – chemical manufacturing, food and beverage processing, paper and pulp, pharmaceuticals, building materials – this creates a genuinely difficult problem. Decarbonisation isn't just a matter of switching to green electricity. It means fundamentally rethinking how steam is generated, how heat is recovered, how cooling and refrigeration systems are powered, and how all of these interact with a production schedule that was never designed with energy optimisation in mind.

The firms making progress on this aren't doing it by intuition or spreadsheet. They're building a rigorous, model–based understanding of how their energy and production systems interact – and using scenario–based feasibility analysis to identify where intervention delivers the most value.

In this blog post, we'll explore why energy modelling is becoming an essential capability for industrial decarbonisation, what a credible modelling approach looks like for manufacturing sites, and where the biggest opportunities typically lie.

The core problem: Energy and production are planned in silos

In most industrial facilities, energy management and production planning operate in separate worlds. The production team optimises shift patterns and line utilisation for throughput. The utilities team manages steam, electricity, compressed air, and cooling to keep the plant running. Finance monitors energy-spend but rarely has the granular data to challenge operational assumptions.

The result is a facility that functions, but not efficiently. Steam is generated at times when waste heat could meet the same demand. Compressors and motors run during peak tariff periods because no one has modelled what load–shifting would cost in production terms. Heat that could be recovered from exhaust gases or condensate is vented because no one has quantified the business case for recovery equipment.

This isn't a failure of expertise. It's a structural problem: without a model that represents energy and production together, in the same analytical framework, you cannot see the interactions that drive inefficiency, and you cannot quantify the value of changing them.

What could a digital twin tell you that you don't know today?

Have you ever wondered how much of your electricity bill could be covered by solar panels on your roof? Whether switching from gas to electricity for your thermal processes would save money – or cost more – under today's prices, and under the prices expected in five years? How much value sits in the waste heat your site is currently venting, both for your own processes and potentially for a nearby district heating network?

Or perhaps: what would it mean financially to buy electricity when spot prices are low, store it in batteries, and run your most energy–intensive processes off that stored power during peak tariff hours? And what happens to that business case if gas prices spike, or the carbon price doubles?

These aren't abstract engineering questions. They are investment decisions – each with a quantifiable return, a quantifiable risk, and a dependency on how all the other systems on your site behave at the same time.

A digital twin of your energy system is what allows you to answer them with numbers rather than assumptions. It models your site as an integrated whole – solar generation, CHP, heat pumps, waste heat recovery, storage, production schedules, grid interaction – and runs scenario and robustness analyses across all of it. The output isn't a qualitative recommendation. It's a euro–for–euro breakdown of what each investment saves, what it costs, and how robust that return is when the assumptions change.

From static efficiency audits to dynamic operational optimisation

Most industrial sites have conducted energy audits. Many have invested in monitoring and metering infrastructure. The typical output is a list of efficiency measures ranked by payback period – insulation upgrades, motor replacements, variable speed drives, LED lighting.

These measures have value. But they address static efficiency: reducing the energy required to perform a given task. They don't address dynamic optimisation: choosing when to perform tasks, and in what sequence, to minimise total energy cost and emissions given time–varying electricity prices, fuel costs, and grid carbon intensity.

Dynamic optimisation is where the larger opportunities often lie for energy–intensive manufacturers – and it requires a fundamentally different analytical approach.

Consider a chemical plant with significant process heat demand, currently met by a gas–fired boiler. A static audit will tell you whether the boiler is operating efficiently. A scenario–based feasibility model will tell you something far more useful: what would full or partial electrification of that heat demand actually cost and save? Is a high–temperature heat pump viable at current electricity prices – and how does that change if gas prices rise? What happens to the business case if you add thermal storage, reducing the peak electrical load the heat pump needs to cover? And how robust is that investment across the range of energy price scenarios that are plausible over the next ten years?

These are the questions that drive real investment decisions – and they require scenario–based modelling, not a one–time audit.

Quantifying the decarbonisation levers

For industrial sites working toward specific emissions targets, simulation modelling provides something that audits and carbon accounting cannot: a quantified, scenario–tested roadmap.

The key levers for industrial decarbonisation are well understood:

Electrification of thermal processes: Replacing gas–fired boilers and direct–fired heaters with heat pumps, electrode boilers, or electric process heaters. The economics depend heavily on the ratio of electricity to gas prices and on grid carbon intensity, both of which vary over time. Modelling these scenarios across a realistic range of future price conditions reveals where electrification makes sense now, where it makes sense in five years, and where alternative approaches are more robust.

Heat recovery and integration – capturing waste heat from high–temperature processes and using it to meet lower–temperature demands, or upgrading it with heat pumps to useful temperature levels. Scenario modelling quantifies the recovery potential, identifies the equipment required, and calculates the displacement of fossil fuel consumption.

On–site renewable generation – solar PV, and in some cases wind, combined with battery storage, can materially reduce both Scope 2 emissions and grid electricity costs. But the real question most site managers are sitting with is: how much of our own electricity consumption could we realistically cover? 80%? 90%? Is 100% self–sufficiency feasible – and if so, does it actually make financial sense compared to 70%?

These questions can feel overwhelming when you're trying to make the case for a carbon neutrality roadmap. Scenario–based modelling turns them from guesswork into a calculated, comparable set of options – so you can find the number that makes the most sense as an investment, not just the one that sounds the most ambitious.

Fuel switching and low–carbon fuels: Biomass, green hydrogen, and biogas are increasingly available alternatives to natural gas for industrial heat. The modelling challenge is understanding how these fuels interact with existing combustion equipment, how their costs compare under different market scenarios, and what operational flexibility they require.

Demand flexibility and grid interaction: Energy–intensive industrial loads are increasingly valuable to grid operators seeking demand response. Modelling the operational flexibility available in your production system: which loads can shift, by how much, and at what cost; allows you to value demand response revenue streams alongside cost reduction measures.

The investment case: Modelling uncertainty before you commit

Industrial energy investments are long–lived. A new CHP unit, a heat recovery system, or an electrification project will be on the books for fifteen to twenty years. The economic case for these investments is highly sensitive to assumptions about future energy prices, carbon costs, and production volumes – all of which are genuinely uncertain.

This is where rigorous scenario modelling earns its value most clearly. Rather than building an investment case on a single set of price assumptions, a well–constructed simulation model supports:

• Scenario analysis across a range of energy price futures, carbon price trajectories, and production volume assumptions
• Sensitivity testing to identify which variables most affect project economics – and therefore where risk management focus belongs
• Breakeven analysis that shows under what conditions a proposed investment fails to deliver its expected return
• Portfolio comparison of alternative investment strategies, showing not just expected returns but the distribution of outcomes under uncertainty

For manufacturers seeking internal approval for significant capital expenditure, or for those working with external lenders or infrastructure investors, this kind of stress–tested, scenario–aware analysis is increasingly the standard of evidence required.

Where to start

The gap between recognising the need for better energy modelling and having a working site model can feel large. In practice, it doesn't need to be.

The most productive starting point is usually a model of the core energy system: the primary heat generation assets, the largest electrical consumers, and the main production lines, built with measured data from existing metering infrastructure. This produces immediate analytical value: it establishes a baseline, identifies the largest sources of inefficiency, and provides a framework for evaluating specific investment proposals.

From that foundation, the model can be extended incrementally: adding heat recovery potential, testing storage configurations, introducing renewable generation, or modelling alternative fuel scenarios. Each extension builds on validated work rather than starting from scratch.

For energy–intensive manufacturers facing both cost pressure and decarbonisation obligations, the question is not whether to invest in this kind of analytical capability. It is how quickly it can be put to work.

How energyPRO supports industrial energy modelling

energyPRO provides the MILP–based dispatch optimisation, integrated financial modelling, and scenario calculation capabilities that industrial energy modelling requires. The platform is used across energy–intensive manufacturing sectors to model complex multi–energy systems, analyse decarbonisation pathways, and build investment cases that hold up under scrutiny.

Want to know more about how energyPRO can help you model, stress–test, and de–risk your industrial energy projects?