fbpx

Thermal predictions – an environmental ML success story

Using deep learning and weather data we have succeeded in calculating days ahead the power a thermal power plant needs to produce. This has allowed calculating and optimizing the use of supplementary fuel leading to a reduction in overall use of expensive oil burners.

I want to show how that machine learning is locally applicable, for «small» Norwegian businesses by talking about how we brought value to the customer and improved the environment (a bit) at the same time.

I will present a business case from the district heating company BKK Varme, where we created a forecast model for predicting required power output up to 48 hours ahead.

This forecast gives BKK Varme the ability to better operate their incinerators leading to less wasted heat and less use of supplementary fuel. I hope that after my lecture the participants will be inspired to look at how data and data analysis can be used to improve the environment.