kth.sePublications KTH
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Sensor fault detection and data-driven modeling of heat pumps
KTH, School of Industrial Engineering and Management (ITM), Energy Technology.
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Sensor data are widely used for simulation, prediction, and system control, with the goal of improving operational efficiency. However, real-world measurements often contain errors caused by sensor misplacement, incorrect calibration, or faults. Such erroneous data can lead to inaccurate outcomes, and detecting these faults remains a challenging problem. This thesis addresses the issue of fault detection in temperature sensors in building heating and cooling systems. A novel algorithm inspired by the PageRank method is developed and benchmarked against spectral clustering and genetic algorithm approaches. The proposed algorithm demonstrates effective fault detection capabilities, with improved accuracy particularly in small-scale systems.

Additionally, the digitalization of the heat pump sector drives the need for accurate performance estimation and forecasting. This work explores several data-driven modeling techniques—including neural networks and linear regression—to estimate heat pump performance. The best models achieve a mean absolute error of approximately 0.4, highlighting their potential for practical applications.

Abstract [sv]

Sensordata används ofta för simulering, prediktion och systemkontroll, med målet att förbättraeffektiviteten i verksamheten. Mätningar i den verkliga världen innehåller dock ofta fel somorsakas av att sensorn är felplacerad, felaktig kalibrering eller skavank. Sådana felaktiga datakan leda till felaktiga resultat, och att upptäcka dessa fel är fortfarande ett utmanande problem.Detta examensarbete behandlar frågan om feldetektering i temperatursensorer i värme- och kylsystem. En ny algoritm inspirerad av PageRank-metoden utvecklas och jämförs med spektral klustring och genetiska algoritmer. Den föreslagna algoritmen demonstrerar effektiva feldetekteringsfunktioner med förbättrad noggrannhet, särskilt i småskaliga system.

Digitaliseringen av värmepumpssektorn driver dessutom på behovet av korrekta prestandauppskattningar och prognoser. I detta arbete undersöks flera datadrivna modelleringstekniker - inklusive neurala nätverk och linjär regression - för att uppskatta värmepumpens prestanda. De bästa modellerna uppnår ett genomsnittligt absolut fel på cirka 0,4, vilket belyser deras potential för praktiska tillämpningar. 

Place, publisher, year, edition, pages
2025. , p. 91
Series
TRITA-ITM-EX ; 2025:533
National Category
Engineering and Technology
Identifiers
URN: urn:nbn:se:kth:diva-371819OAI: oai:DiVA.org:kth-371819DiVA, id: diva2:2007799
Supervisors
Examiners
Available from: 2025-10-21 Created: 2025-10-21 Last updated: 2025-10-21Bibliographically approved

Open Access in DiVA

fulltext(5106 kB)168 downloads
File information
File name FULLTEXT01.pdfFile size 5106 kBChecksum SHA-512
5e0c76d485b25332cb91aa4a0cbb067a0f7204d93ed53347b2e0bfbe2b68e06e1b1a19192d419f61d49ae0f7b8b0383a4b79472f648d56656a732c0566bb2a0e
Type fulltextMimetype application/pdf

By organisation
Energy Technology
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
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

urn-nbn

Altmetric score

urn-nbn
Total: 1032 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf