AI-based control models for optimising the flexibility of heat pumps to relieve the strain on the electricity grid

The FlexHP research project is pursuing this approach to optimise load shifting on the electricity grid side and increase the share of renewable electricity. To this end, buildings are integrated as thermal storage units and heat pumps as highly efficient heat generators into a dynamic energy management system alongside the familiar components (battery storage, electric cars, photovoltaic systems), and load consumption is optimised for the grid using an AI-supported control algorithm. The aim is to turn the building into an additional source of flexibility for the electricity grid during periods of surplus electricity.
The aim of the dynamic energy management system is to optimise load shifting at building level through the interaction of forecast models and optimised control of individual consumers. This requires modelling of all players involved in the electricity grid in order to predict their behaviour. Using forecast models that have been tested and validated with historical data, an algorithm is being developed that enables optimised interaction between the individual participants, with the objective of load balancing.
The individual models and the control algorithm’s functionality are being tested in a living lab using real system components. Power grid parameters from the low- and medium-voltage levels are used as the primary control variables. The aim is to test the system’s applicability using commercially available system components on a laboratory scale.