Predictive maintenance

© Fraunhofer ISE
Absorption chiller at Milan Malpensa airport.

On the basis of measurement data from existing control and building automation systems, we develop methods that proactively determine the condition of building systems. Thus, facility managers can organize their maintenance processes in advance and on the basis of the real condition and save costs. We use methods from the field of machine learning as well as methods for tensor decomposition of the data sets in order to process very large amounts of data promptly and efficiently.