Research activities - Visions and plans TrønderEnergi
Wind power - powered by AI
Research activities - Visions and Plans

What are your thoughts on NorwAI from your perspective?
In the consumer world, much information is already digitized. For industrial applications, the first step is digitization. TrønderEnergi is largely digitized. Next step is to automate these parts using AI methods. For example, all electricity produced by wind farms is handled automatically by AI. Parts of the energy trade already take place in real time by an AI “algotrader”. The technologies are thus already in use but can be further developed. Given the complexity, we can reap the benefits of collaborating with other industries and support each other scientifically.
What are your gains from participating in the NorwAI effort?
TrønderEnergi is involved in work packages dealing with data streaming, transparency and hybrid solutions.
Two examples illustrate possible benefits: Wind power is non-adjustable, which means that hydropower must be used when wind power is low. In addition, power usage for charging electric cars and for freezer counters can be optimized using electricity when the demand for electricity is low elsewhere.
Use of data can improve maintenance. It should only be performed when required. Thus, we ensure being ready to replace parts that are worn out or stop working. Reduced maintenance time gives increased power production as wind and water turbines are not left idle due to repairs
and maintenance.
Many physical models are available for wind and hydropower plants. Combining such models with data from the specific equipment to improve predictions of service life and power production, will gain our business.
Going forward, what are your expectations for NorwAI?
Transparency is about making the results of the AI systems understandable to users who are unfamiliar with the AI systems. Hybrid models based on physical models that technicians already understand, will help prevent the AI system from becoming a black box to not-so skilled.
Weather forecasts, electricity prices, sensor data from wind turbines and from refrigerated display counters are forms of streamed data. In order to be able to forecast power consumption, power production or the remaining service life of a wind turbine, we must use and develop the correct methods for data streaming.
We have several goals. One is, for example, wind and hydropower plants that can determine on their own how much energy they should and will produce. But also, that shops can adjust power consumption according to their need, that the electricity produced or consumed is traded automatically, that the equipment informs when it needs maintenance, what problem needs to be solved, what parts are needed, and how long the repair or maintenance will take. When technicians and others need to make a decision, the system should provide the vital, data-driven information for improved decisions.