How clean energy modelling is key to the energy transition

Colin McKinnon: The IEA has been around for seven years. It was originally set up to help organisations that are facing environmental challenges and as a centre of expertise around data discovery and around developing software tools and products to help people navigate weather and climate change.

We spend most of our time working with organisations that have these challenges and using the data to see whether we can provide solutions.

The genesis of EnergyMetric in many respects was a bid that we won from the UK Space Agency about four and half years ago to help developing countries – in particular, small-island developing states – in a specific themed area, energy. We helped seven of these countries – including Mauritius and the Seychelles – on this journey to a higher penetration of renewable energy. They all have different targets, [and] fairly ambitious ones.

In some of these places the weather is quite changeable and there is not that much data available. What we try to do is pull together the data and information and bring it altogether in a packaged tool, a piece of software, to help them navigate the journey and make more integrated decisions. EnergyMetric takes different forms of information from satellites, weather models, climate information and bundles that up into a piece of technology that enables people to make more informed decisions.

We are selling the software to make money, but we are not a company; the IEA is part of the University of Reading. We now employee around 25 people and we employee two types of people: many are data scientists, with physics and maths type backgrounds, and the others are software developers.

Andrew Groom: EnergyMetric is an online application that enables you to model renewable generation potential. So, if you take an energy transition use case – say, country X is trying to get to 20%, 30% or 40% penetration of renewables by a particular time – they require a plan to be able to achieve it, a policy or roadmap to be able to follow. They want to explore different options for how they might achieve their particular target.

They need to understand their resource – solar and wind, typically – and they would like to be able to say, “we would like this project in this location and this project is this location, what kind of generation would we get from that combination of generation assets?”

It’s a scenario planning tool, a strategic planning tool, which enables you to understand renewable generation potential.

There are two key parts to how it works: the first is this idea of scenario, so we have a map-based interface, your particular geography in which you are doing planning. You characterise assets —such as a solar farm or wind farm – with a set of technical parameters and financial parameters. You can have different specifications of assets in different locations. You can decide where the locations are based on the available resource.

We are also able to integrate geo-spatial layers into the application. It’s not just resources that determine where you might want to put a renewable generation facility but it’s also [questions regarding] environmental information, land ownership data sets and where the existing transmission and distribution infrastructure is.

These kinds of datasets also feed into the strategic site selection thought process. What we are producing is a detailed time series of weather data [from] up to ten years ago.

AG: We have done some validation work with the islands that we have been working with and the results that we have achieved are strong in those areas but each time you work in a different geographic context you will have slightly different challenges in terms of air modelling and these kinds of things.

Suriname is just one example of a country that has ambitions to increase the amount of renewables generation it uses in its generation mix. It already uses lots of renewable generation; it has a large hydro resource but there is a lot of variability associated with it. You can have a dry year or a wet year; production capability can move up and down as a result.

They would like to understand what options they would have for integrating more solar or wind into that mix. This kind of planning work is fundamental for developing a strategy for the energy transition. The fact that we produce our own weather data is unique; other software providers are typically just consuming the weather data. Wind is more difficult to model. There is a lot more variability in the wind resource. Solar does not vary wildly from year to year or month to month.

We adopt a probabilistic approach to the way we do the modelling. We try to quantify that uncertainty. We don’t just give you a number; we give you a range around that number. You might have an optimistic view or a pessimistic view, depending on who you are and what you are trying to achieve. 

CM: All weather and climate modelling is modelling to some degree; no matter how sophisticated you might be in terms of observations of what has happened, it’s still a network with partial coverage. In the UK, there is not an observation point at every square metre in the country.

You have this issue of interpolation and working out what goes on between the various points. Computationally, it’s a big heavy lifting job. We are running the application EnergyMetric in AWS, but cloud infrastructure has made a big difference.

Wind is a lot more intermittent. It is probably the most important work in the whole energy transition; the whole point of what we have been trying to do is quantify as accurately as we can this intermittency. Whether a project is successful or not is largely driven by the amount of energy it is producing.

// Main image: Wind turbine in rain. Credit: Riccio da favola via Shutterstock