Dexter and the future of renewable energy

When it comes to energy, renewable is the future. Moving away from the current fossil fuel economy can make our communities healthier, reduce pollution, and create more and better jobs. This means that the quest for green, renewable and profitable energy sources and distribution models is a hot topic in our modern society, especially since renewable energy is going to act as a disruptive force on energy systems and the energy market.

Dexter Energy tries to tackle these challenges of the future by using brand-new technology to better forecast future energy needs and flows.

Changing energy systems and demands

We used to have big, fossil-fueled energy plants that strictly followed the demands of consumption. The new system paints a completely different picture and is characterized by the emergence of large amounts of power generation that are not plannable, but actually dependant on weather patterns. Due to this whole change of the energy system, we are in dire need of new technologies to meet the accompanying challenges. 

What Dexter Energy does

To meet the demands of the future, Dexter lends energy companies a helping hand. We do this by forecasting energy needs and the availability of energy through artificial intelligence (AI).

Energy companies have a lot of customers that consume or produce energy. To properly balance electric grids, the companies have to be able to make accurate predictions about the amount of energy that all consumers are going to use at a certain moment. But they also have to know what all of their producers (wind, solar power) are going to produce at any given moment. Dexter helps energy companies to draw up accurate forecasts of the aforementioned energy flows to keep grids stable and safeguard the constant availability of electricity. 

Weather data

“We need a lot of data to run our models and perfect the accurateness of our forecasts”, says Luuk Veeken, founder and CEO of Dexter. “Amongst the most important data sources are weather data. To get the best out of these data sets, we process seven different models in our cloud and combine this information with grid data, time tables of our customers and 400-500 other features that go into our models.” Subsequently, these models combine all the data and learn to recognize patterns. This enables them to make really accurate predictions on future developments and energy flows. 

Pilot

Dexter starts every project with a pilot. This means that the customer gives us a bit of data from energy producers and/or users that they have in their portfolio. Following up on this, we tune the models accordingly and run a demo in which we actually perform a forecast in a live setting over a given period of time to show the customer how good and reliable we really are.

Data optimization

At this point in time, many energy companies use weather data that are already processed. The problem is that this information is not yet precisely optimized to their needs and portfolios. To give an example: one of our customers just used one of the simplest weather models available to forecast their entire solar portfolio.

Dexter, on the other hand, actually combines different weather models. We also have a layer on top of our weather models that actually learns which weather model performs best on a certain area and timeframe. Furthermore, we combine different weather models and optimize the necessary grid size for these models. The result is that you actually get much better data to feed your models. 

“So far the feedback we have received from our customers is really positive”, says Veeken. “We have been able to reduce the imbalance costs of many a customer significantly. Additionally we have also helped clients in further optimizing their portfolios.”

Developing new products

Dexter is currently developing a new nowcasting product. This novelty proves the value of short-term forecasting for wind power and solar power generation. We do this by combining all the different weather models with satellite and radar data. Advanced deep learning is used to predict how clouds and winds are going to move. The result is a better and more detailed forecast that increases the accuracy for our customers.

Crossing the border

“Since I founded the company 2,5 years ago, we have been able to close many deals in the Netherlands. We are also about to close our first two deals outside of the Netherlands, something that makes us very proud. Furthermore, we are also looking at other markets now. We already have quite a presence on the Dutch market and hope to extend that visibility to Belgium, Austria, Germany and the UK. We want to be the forecasting partner for energy companies across Europe.

I’m really grateful for the fact that Dexter has been a success story so far. It’s great that we can help energy companies tackle the energy transition by getting more renewable energy into the system due to our software. That makes me really proud”, Veeken concludes.

Play Video