Open Menu
Dübendorf ZH – Empa researchers are devising ideas to help buildings save power. In their experiments, they fed a self-learning heating control system with temperature data from the previous year and the current weather forecast. The first tests proved successful.

Office blocks and factory halls are often equipped with anticipatory heating systems that afford energy savings for the operators. However, such individual programming is too expensive for private homes, explains the Federal Laboratories for Materials Science and Technology (Empa) in a statement.

Empa researchers are now testing a method that does not require the skills of programming experts. Instead, heating and cooling systems can learn how to save energy by themselves, using data from the previous weeks and months and the current weather forecast. In their experiments, the researchers fed this data into a self-learning heating and cooling control system in the Empa research building NEST – with success. The system adhered much more closely to the pre-set comfort specifications while using around 25 per cent less energy.

Further tests are in the pipeline. Project manager Felix Bünning is curious: “I think that new controllers based on machine learning offer a huge opportunity. With this method we can construct a good, energy-saving retrofit solution for existing heating systems using relatively simple means and the recorded data.”