Heat Pumps Aided with AI Consumes Much Less Energy

Researchers at EPFL have developed a way that makes use of synthetic intelligence to design subsequent-technology warmth-pump compressors. Their methodology can lower the pumps’ energy requirement by around 25%.

In Switzerland, 50 — 60% of new homes are outfitted with warmth pumps. These techniques attract thermal vitality from the encircling setting — resembling from the bottom, air, or a nearby lake or river — and switch it into warmth for buildings.

Whereas at the moment’s warmth pumps typically work correctly and are environmentally pleasant, they nonetheless have substantial room for enchancment. For instance, by utilizing microturbocompressors as an alternative of typical compression methods, engineers can scale back warmth pumps’ energy requirement by 20-25% (see inset) in addition to their impression on the surroundings. That is as a result of turbocompressors are extra efficient and ten occasions smaller than piston units. However, incorporating these mini elements into warmth pumps’ designs is just not simple; problems come up from their tiny diameters (<20 mm) and quick rotation speeds (>200,000 rpm).

At EPFL’s Laboratory for Applied Mechanical Design on the Microcity campus, a staff of researchers led by Jürg Schiffmann has developed a way that makes it simpler and quicker so as to add turbocompressors to warmth pumps. Utilizing a machine-studying course of referred to as symbolic regression, the researchers got here up with easy equations for rapidly calculating the optimum dimensions of a turbocompressor for a given warmth pump. Their analysis simply gained the Greatest Paper Award on the 2019 Turbo Expo Conference held by the American Society of Mechanical Engineers.

The researchers’ methodology drastically simplifies step one in designing turbochargers. This step — which includes roughly calculating the perfect dimension and rotation velocity for the specified warmth pump — is extraordinarily essential as a result of the preliminary estimate can significantly shorten the general design time. Till now, engineers have been utilizing design charts to dimension their turbocompressors — however, these charts turn into more and more inaccurate the smaller the tools. And the tables haven’t saved updated with the newest expertise.