Jungle AI - Predicting wind turbine performance with past data

News |
TKI Wind op Zee

Jungle AI has developed a smart tool that reveals and resolves the underperformance of wind turbine components. What is special is that the system learns what normal behaviour is in order to make predictions about future behaviour. This provides free extra sustainable energy.

Currently, many wind turbines are performing less than they could. This leads to lower yields and unnecessary downtime. For the founders of Jungle AI who combine knowledge of wind turbines, machine learning and electrical engineering, this was an eyesore. Pim Breukelman: “We notice that there is an increasing discrepancy between the enormous jungle of available data and the ability of people to understand what they mean.” In 2017, they decided to develop the world's most effective tool to increase this performance. And that with data that is already available, but not used.

Canopy uses sensors in wind turbines that now register, for example, a wrong orientation to the wind, constant switching off or rotor imbalance. This sensor data is processed with a self-learning algorithm (machine learning). As a result, the system learns what normal behaviour is and learns to recognize deviant patterns. This allows operators of large offshore wind farms to compare the actual behaviour of their turbines with how they should operate. Breukelman: "The most special thing is that you can look at all components from behind your screen and, thanks to a historical data analysis, you can see in the future when something breaks." This allows operators to repair components before causing downtime.

Operators and asset managers of wind farms benefit from this. They benefit directly from higher energy production, less unnecessary downtime and lower costs due to more efficient wind turbines. They benefit indirectly because they can work more factually and purposefully. Moreover, they do not have to make large investments because the sensor data is already available. Its use yields profit from day 1. According to Breukelman, the revenue for the customer is a factor of 10 higher than the fee that customers pay for the system.

Canopy also helps the energy transition move forward. “Our tool provides extra sustainable energy for free at a lower cost. For every 100 wind turbines, you get two free. After all, every extra MWh of electricity generated by a wind farm that is already operational does not have to be generated with fossil fuels. This makes sustainable assets cheaper than fossil power stations. You can't make it much simpler.”

In 2021 Jungle AI launched Canopy and deployed it to eight players in the (offshore) wind sector. These customers are now in the process of transitioning from the first wind farms to the rest of their portfolio.

Jungle AI is one of the three nominees for the Offshore Wind Innovators Award 2021. Would you like to attend the presentation of the winner of this Award? Register for our live (and online) event on March 22 in Rotterdam.

Register here