Although heat pumps already include sensor technology and communication interfaces, these are only queried sporadically and not analyzed thoroughly enough to detect developing faults. It is still very cost- and labor-intensive to carry out fault detection and diagnosis by acquiring and analyzing process data.

Fraunhofer IIS/EAS has joined forces with other project partners as part of the Smart Heating System Optimization (SHANGO) project with the aim to find solutions that will reduce the energy consumption of heat generators in the private sector sustainably and in the long term. As part of this group, Fraunhofer IIS/EAS is responsible for the topic data science, while other project partners contribute expertise in relation to heating technology and energy management software.

In the future, the aim is for installed systems – with their existing sensor technology and control and communication hardware – to be monitored and optimized during operation with minimal installation and operating effort. Using artificial intelligence (AI) and machine learning, missing information will be added and datasets generated that also incorporate additional data such as weather, building data, and the observed operating patterns. Later in the project, an AI-based algorithm will be trained and will provide the basis for the operation in the field.

A laboratory test stand is being set up at Fraunhofer IIS/EAS in order to carry out hardware-in-the-loop (HiL) emulations of the heat generator within a virtual building environment. The process data obtained are to be used for the AI algorithms in order to detect problematic operating states. Moreover, the load-shifting potential of such heat pumps is to be investigated in order to determine their suitability for dynamic load management in the power grid.