Change search
ReferencesLink to record
Permanent link

Direct link
Fault Detection using KPCA analysis applied to building efficiency
KTH, School of Industrial Engineering and Management (ITM), Energy Technology, Applied Thermodynamics and Refrigeration.
2015 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Over the past few years, due to global warming and the depletion of energy resources, the reduction of buildings energy consumption and control systems in buildings have received more attention than ever. In particular, the management of faults affecting buildings is a field that could enable huge energy savings. This report investigates the coupling of a statistical analysis with a building simulation software with the aim of detecting faults. To do so the data collected from the buildings is compared to the simulation results of the model using IDA ICE. Then a Fault Detection algorithm, that has been developed based on the Kernel Principal Component analysis, processes those differences complemented by several variables that help the algorithm to understand the behavior of the building. Since data with identified faults was not available, this methodology has been tested on simulation versus simulation comparison based on the model of a well known building. Several faults were then simulated and were successfully detected by the algorithm.

Place, publisher, year, edition, pages
2015. , 51 p.
National Category
Engineering and Technology
URN: urn:nbn:se:kth:diva-166733OAI: diva2:812024
Educational program
Master of Science - Sustainable Energy Engineering
Available from: 2015-08-27 Created: 2015-05-14 Last updated: 2015-08-27Bibliographically approved

Open Access in DiVA

fulltext(2164 kB)54 downloads
File information
File name FULLTEXT01.pdfFile size 2164 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Applied Thermodynamics and Refrigeration
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 54 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 163 hits
ReferencesLink to record
Permanent link

Direct link