We’ve all heard of ‘big data’. We might not understand exactly what it means, or how we can use it to improve our lives, but it’s got ‘big’ in the name so must be important!

That said, the principle of ‘big data’ is quite simple. The idea is that we can collect information from large data sets and then use it to identify patterns and trends, in areas including human behaviour. Businesses can then use it to sell more products and services.

That is the theory. Doing it is rather more complex, hence the old joke comparing big data to teenage sex: everyone talks about it and thinks everyone else is doing it, but nobody really knows how to do it properly if they’re even doing it at all. That punchline might have lost something in translation.

One way businesses in the wind sector are increasingly looking to make use of big data is through the use of ‘digital twins’. This is where wind farm owners use data from a range of sources – turbine specifications, wind speeds, site layouts – to build a digital replica of the project. Graeme McCann, DNV GL’s head of turbine engineering, calls it a “virtual representation of a physical asset”.

By doing this, McCann says wind farm owners can gather useful insights on how to optimise their performance of their projects. This could include making use of weather forecasts to determine how well a scheme is likely to perform on a given day, and how hard each turbine should work to maximise its performance.

Gathering information on how hard turbines have been working can also help owners to develop proactive strategies to manage anticipated maintenance problems before they happen; and test out these strategies using the information gathered.

In addition, owners can benefit if they can make better decisions on the risks and rewards of using turbines beyond their operational life, for example, or whether it makes sense to repower projects.

McCann says that producing ‘digital twins’ of wind farms can play an important role as project owners plan their strategies: “Those models then become living things,” he says. “They don’t stay static. They learn from the operation of the asset, and update themselves based on the information of how the asset is really performing.”

This can then enable owners to make a series of marginal improvements in their wind farms. For example, McCann says predictive maintenance strategies can help to improve the availability of turbines by a number of percentage points; and advanced wind farm control systems can similarly add 1%-1.5% to energy yield. If an owner can make improvements like these on a 50MW project, the effective gains could enable them to add an extra turbine to the site.

“There’s no quantum leap here, but this can facilitate a large number of incremental gains,” he says. “When you add them up you can have a big effect.”

The idea of ‘digital twinning’ is not a new one: the aerospace sector has used mathematical modelling to similar effect since the 1970s to help it optimise its planes. Likewise, wind turbine makers have used comparable technology to improve their turbines’ performance.

But using it to improve the performance of whole wind farms, and pulling in data from a wide range of sources, is relatively new.

General Electric came out with its ‘digital wind farm’ concept two years ago, to use both hardware and software to optimise the performance of its projects. There are other companies doing similar, and DNV GL launched a platform called Veracity in February to help firms make use of big data. This brings together data from a range of sources to model projects’ performance.

McCann says a good digital twin would allow lots of disparate simulation models to be brought together from a range of stakeholders from across the industry.

This will then raise a further challenge: companies need to decide not only what data they need to come up with an accurate model, but also make sure that they are comfortable sharing performance data that could be commercially sensitive.

These relationships will need to evolve if wind companies are to make the most of big data and digital twinning. In that regard, big data isn’t just like teenage sex. It’s also about the complicated relationships that come with it.