Fault Detection and Diagnosis (FDD) in building systems

The project at a glance

Start date:

15 Oct 2020
Duration in months:

48
Funding:

Own funding
Principal Investigator(s):

Sebastian DIETZ

Frank SCHOLZEN

About

Modern building automation systems collect more and more operational data. But this potential of information is seldom fully exploited, as it requires a lot of expert knowhow and time to extract the valuable information out of tons of recorded data. So automatic Fault Detection and Diagnosis (FDD) would be appropriate to take over this work. However, FDD methods are rarely used in building systems due to their individual design. A comprehensive engineering process shall be developed, which is highly automated and applicable by non-machine learning experts. The development process will be supported by using real building data at our disposal while the transferability of the developed method will remain a key exigence of the method.