Using the past to visualize the future
Published on 03rd of March 2020
Mariel Alexandra Garrido Urena
Head of Fleet Optimization at Vestas
When asked about Vestas’ digital transformation, I’m reminded that this is a journey that we started many years ago - and that perhaps it’s as much evolution as transition: We were there right at the first tentative stages of digital data gathering and analysis and have always stayed at the forefront of advancing digital adoption in the renewables industry.
Here’s a bit of context: In the 90s we connected individual turbines into a central server; we invested in what was then one of the planet’s most powerful supercomputers; In 1998, we connected the first turbine to the internet. We then established the first online service centers, SCADA systems, Data Center and cloud servers. We’ve always believed in the potential of data and connectivity to add value.
The evolution is that now, in addition to using the advanced statistical and engineering methods we had in place, we are considering data in a more systematic way: We’re investing in
machine learning and neural networks; developing patterns. And this is where it gets really interesting: Data has always allowed us to draw conclusions about our world, but now data visualization of these patterns can help us better understand and use our data. Engineers can use increasingly sophisticated models to react upon. This is why we’re now focusing on refining our suite of data visualization tools.
And again, this is evolutionary; we’ve been working with data visualization since 2012, and have created visualization software that we’ve developed for internal purposes – it was raw then, but using internally developed tools we began to develop report-based platforms for our customers.
In 2015 though, we realized that to remain at the leading edge of the data visualization frontier, it would make sense to partner with best-in-class professional specialists. So, having acquired the company, for the last few years we’ve been collaborating with Utopus Insight’s highly experienced data scientists.
We’ve always collected live data; each turbine is effectively a computer on its own and each park has a server. This data is sent live to a centralized physical server or to the cloud. All our data is consolidated in 10 minutes - this is the industry standard for good data. Based on this live performance data we’ve now released two products to help our customers manage their portfolio of energy producing assets – including solar – we are asset agnostic, after all!
The first of these products, PowerForecast is based on the largest meteorological database in the industry and a climate library containing 20 plus years of data. It gives our customers site-specific power forecasts every 10 minutes – although we’re now proud to able to offer shorter intervals of five minutes in many cases. In China and Australia, for example, we’ve demonstrated that our algorithms are the best in the industry, due to our unique correlation between live stream, stored data and experience.
Our second product, WeatherForecast lets our customers analyze on-site weather conditions remotely. This allows them to improve seasonal and short term planning, reduce operation and maintenance costs, and increase on-site safety. Incorporating the top 10 alerts needed to plan ahead safely, the app gives customers a comprehensive overview of weather variables from a single screen visualization.
But what does it look like? We use the Scipher platform that’s interactive and customizable for a variety of apps. This could be anything from a computer screen to a tablet or iPhone. In essence it looks like a map and is designed to be very intuitive and user friendly. For example, a turbine will show a green signal if it’s working to its potential but shift to red to signal a performance or service issue. Users can then simply click on the icon and drill down for more data about specific issues.
Looking forward, from a market perspective, we have a hands-on roadmap of imminent product releases lined up, including a refined visualization app, advanced predictive maintenance models and new tools based on user cases. What we’re seeing generally is that in our journey to become even more customer centric, we are broadening our scope and developing new solutions that will encompass everything from logistics to plant design. We’re taking data visualization very seriously right across the entire value chain for our customers.