Controlling power consumption thanks to artificial intelligence

This means that the time margins can be identified and used to reduce and smooth peak loads. Andrew Paice, head of the iHomeLab at the Lucerne University of Applied Sciences and Arts, notes, "This requires intelligent and efficient energy and load management, which can be used to shift energy." To this end, a team from the iHomeLab and the Lucerne University of Applied Sciences and Arts' Thermal Energy Storage Competence Center, together with partners ASGAL Informatik GmbH, Semax AG and Vilters-Wangs Electricity Works, developed a prototype system that uses artificial intelligence to help utility companies with load management. The Swiss Agency for Innovation Promotion Innosuisse supported the project.

Need more energy in the short term, which can be saved again afterwards

In order to cover the greater demand in the evening, the surplus energy produced during the day, for example from photovoltaics or other renewable sources, should be stored for a short time, and this without the need to install additional storage. Boilers, heat pumps or e-mobiles can be used for this purpose. Solutions already exist for large-scale distributors such as refrigerated warehouses: If excess energy is produced, they can be cooled a little more. Subsequently, the cooling is reduced again; the energy is thus available elsewhere. The team from the Lucerne University of Applied Sciences and Arts and its research partners developed a solution thanks to which single- and multi-family houses or commercial enterprises can also be used in a similar way.

For example, the boiler is brought to the maximum temperature when needed and thus used as a buffer storage. The project focused on houses with heat pumps, because electricity companies can access them with a small adjustment and control them accordingly.

Making better use of energy thanks to artificial intelligence

In order to reduce or smooth load peaks in this way, it is first necessary to ascertain where heat pumps are installed at all and where photovoltaics are generated - information that is only partially known to the electricity companies. The more difficult question, however, was: How much time is available to reduce and smooth peak loads without the users feeling a loss of comfort because the apartment becomes uncomfortably cool or the water too cold?

To answer the two key questions, the researchers took advantage of the increasing digitization of electricity distribution networks with smart meters, which electricity companies use to read electricity consumption for billing purposes. This data is available, but calculating the required thermal model of a building from it is very complex. That's why artificial intelligence came into play here. "The algorithms analyze smart meter data and identify individual electricity-consuming devices such as heat pumps, boilers or e-mobiles and electricity-producing devices such as photovoltaic systems from the total electricity consumption," explains Andrew Paice.

Valuable information about the consumers would be determined, such as their maximum power consumption, the time they are switched on and off, and the energy consumption per day. Paice adds, "When this data is combined with temperature and weather forecasts, predictions can be made about energy consumption on a given day."

Added value without sacrificing comfort

The project resulters open up new opportunities for the project partners ASGAL Informatik GmbH and Semax AG: Thanks to the automatic identification of electricity consumers and the calculation of their so-called load shifting potential, they can offer utilities a service that helps them to save network costs without the need for additional investments in their distribution networks.

For consumers, the innovation means no loss of convenience; nor do they have to disclose any additional information, because the evaluation is carried out exclusively on the basis of standard smart meter data and without any additional hardware installation. In addition, the data is automatically synchronized permanently with the changing circumstances in the buildings. Thus, in line with the Energy Strategy 2050, added value can be generated for the electricity utilities and for building owners.