Energy consumption forecast and model predictive control

Wahrscheinlichkeitsverteilung
© Fraunhofer ISE

Probability distribution of the daily energy consumption taking variations of uncertain model parameters and time series input data into account.

Through the automated and continuous provision of energy consumption forecasts, deficits in building operation can be identified and the implementation of load control mechanisms made possible. We develop models on the basis of physical and statistical modelling methods and provide energy demand forecasts taking into account forecasts for occupancy and local weather. We integrate our models using Functional Mockup Units or machine learning methods with IoT platforms.