Umeå University's logo

umu.sePublications
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
Application of Regression and ANN Models for Heat Pumps with Field Measurements
Umeå University, Faculty of Science and Technology, Department of Applied Physics and Electronics.ORCID iD: 0000-0003-3115-4195
Umeå University, Faculty of Science and Technology, Department of Applied Physics and Electronics.
Umeå University, Faculty of Science and Technology, Department of Applied Physics and Electronics.ORCID iD: 0000-0001-6400-7744
Umeå University, Faculty of Science and Technology, Department of Applied Physics and Electronics.ORCID iD: 0000-0002-8704-8538
2021 (English)In: Energies, E-ISSN 1996-1073, Vol. 14, no 6, article id 1750Article in journal (Refereed) Published
Abstract [en]

Developing accurate models is necessary to optimize the operation of heating systems. A large number of field measurements from monitored heat pumps have made it possible to evaluate different heat pump models and improve their accuracy. This study used measured data from a heating system consisting of three heat pumps to compare five regression and two artificial neural network (ANN) models. The models’ performance was compared to determine which model was suitable during the design and operation stage by calibrating them using data provided by the manufacturer and the measured data. A method to refine the ANN model was also presented. The results indicate that simple regression models are more suitable when only manufacturers’ data are available, while ANN models are more suited to utilize a large amount of measured data. The method to refine the ANN model is effective at increasing the accuracy of the model. The refined models have a relative root mean square error (RMSE) of less than 5%

Place, publisher, year, edition, pages
MDPI, 2021. Vol. 14, no 6, article id 1750
Keywords [en]
heat pump, artificial neural network, regression model, modeling, field measurements
National Category
Energy Engineering
Identifiers
URN: urn:nbn:se:umu:diva-181746DOI: 10.3390/en14061750ISI: 000634408500001Scopus ID: 2-s2.0-85107949313OAI: oai:DiVA.org:umu-181746DiVA, id: diva2:1539386
Available from: 2021-03-23 Created: 2021-03-23 Last updated: 2023-09-05Bibliographically approved
In thesis
1. Utilization of a GSHP System in a DHC Network: modeling and optimization
Open this publication in new window or tab >>Utilization of a GSHP System in a DHC Network: modeling and optimization
2021 (English)Doctoral thesis, comprehensive summary (Other academic)
Alternative title[sv]
Användning av bergvärmepumpar i ett fjärrvärme- och fjärrkylanätverk : modellering och optimering
Abstract [en]

The ground source heat pumps (GSHPs) of customers connected to the district heating and cooling (DHC) network can benefit both the customer and the energy company. However, operating the GSHP to minimize the cost of providing heating and cooling to the customer while ensuring the long-term stability of the ground temperature is a challenge. This thesis addresses the challenge by developing accurate models of GSHP and optimizing the operation of the GSHP system using these models.

The models presented in this thesis use field measurements to develop accurate models with low computational time. The main components of a GSHP system are the heat pump and the borehole heat exchanger (BHE). This thesis presents two approaches to use measured data to improve the accuracy of analytical models for BHE. The first approach is the calibration of the model parameters using this measured data. The second approach combines the analytical model with an artificial neural network model resulting in a hybrid model. The calibration approach reduced the relative RMSE of the analytical model from 21.9% to 13.9% in the testing period. The relative RMSE of the hybrid model for the testing period was 6.3%.

We compared different data-driven models for heat pumps and determined that artificial neural network models have an advantage over traditional regression models when field measurements are available. The artificial neural network model was refined to better utilize the measured data. The refined models of heat pumps had a relative RMSE of less than 5%.

The hybrid BHE model and an artificial neural network model for the heat pumps were used to model the GSHP system. The model was validated using four years of field measurements. The relative MAE for the compressor power and BHE power were 7.3% and 19.1% respectively.

The validated model was used to optimize the operation of the GSHP system. In optimal operation, the cost of providing heating and cooling to the area was minimized from the perspective of the energy company while maintaining a stable temperature in the ground. In optimal operation, the annual cost of operation was shown to reduce by 64 t€ and the annual CO2 emission was shown to reduce by 92 tons.

Abstract [sv]

Bergvärmepumpar som är anslutna till fjärrvärme- och kylnät kan vara till fördel både för användare och leverantörer av energi. I detta sammanhang utgör emellertid driftstrategin en stor utmaning för att minimera energikostnaden för att tillhandahålla värme och kyla och samtidigt säkerställa att marktemperaturen långsiktigt blir stabil. En viktig målsättning med denna avhandling har därför varit att förfina och utveckla modeller för driftoptimering av ett bergvärmepumpsystemi samverkan med ett fjärrvärmenät.

Huvudkomponenterna i ett bergvärmepumpsystem är värmepumpen och borrhålsvärmeväxlaren. Denna avhandling presenterar två metoder för att använda verkliga driftdata med syftet att förbättra noggrannheten hos analytiska modeller för borrhålsvärmeväxlaren. Det första tillvägagångssättet är kalibrering av modellparametrarna med hjälp av uppmätta data. Det andra tillvägagångssättet kombinerar den analytiska modellen med en artificiell neural nätverksmodell som resulterar i en hybridmodell. Kalibreringsmetoden reducerade den analytiska modellens standardavvikelse från 21,9% till 13,9% under testperioden. Standardavvikelsen för hybridmodellen för testperioden var 6,3%.

Vid jämförelsen av olika datadrivna modeller för värmepumpar konstaterades det att artificiella neurala nätverksmodeller har en fördel jämfört med traditionella regressionsmodeller då fältmätningar är tillgängliga. Den artificiella neurala nätverksmodellen förfinades för att på bästa sätt nyttja uppmätta data. Med de förfinade modellerna erhölls en standardavvikelse på mindre än 5%.

Borrhålsvärmeväxlarens modell och en artificiell neural modell för värmepumparna användes för att modellera bergvärmepumpsystemet. Modellen validerades med driftdata från fyra års fältmätningar. Det relativa medelfelet för kompressoreffekten och borrhålsvärmeväxlarens effekt var 7,8% respektive 19,1%.

Den validerade modellen användes för att optimera driften av bergvärmepumpsystemet. Vid optimal drift minimerades kostnaden för att tillhandahålla uppvärmning och kyla sett ur energileverantörens perspektiv, samtidigt som en stabil temperatur i marken bibehölls. Vid optimal drift visade sig den årliga driftskostnaden minska med 64 t€ och det årliga koldioxidutsläppet visade sig minska med 92 ton.

Place, publisher, year, edition, pages
Umeå: Umeå University, 2021. p. 60
Keywords
Ground source heat pump, district heating and cooling, optimization, borehole heat exchanger, artificial neural network, hybrid model, field measurements, calibration, heat pumps, prosumer
National Category
Energy Engineering Energy Systems
Identifiers
urn:nbn:se:umu:diva-187781 (URN)978-91-7855-649-6 (ISBN)978-91-7855-648-9 (ISBN)
Public defence
2021-10-19, Triple Helix, Samverkanshuset, Umeå Universitet, Umeå, 13:00 (English)
Opponent
Supervisors
Available from: 2021-09-28 Created: 2021-09-21 Last updated: 2021-09-22Bibliographically approved

Open Access in DiVA

fulltext(2979 kB)585 downloads
File information
File name FULLTEXT01.pdfFile size 2979 kBChecksum SHA-512
8e440b1fe14ad3d54b8508f9ead050891b6dc48a85d3b04832d20eec33ab2e799f70c3cb23096611044adc35f3f3799d7de7e0c7fe7fd2e4ea09b46ca977ade5
Type fulltextMimetype application/pdf

Other links

Publisher's full textScopus

Authority records

Puttige, Anjan RaoAndersson, StaffanÖstin, RonnyOlofsson, Thomas

Search in DiVA

By author/editor
Puttige, Anjan RaoAndersson, StaffanÖstin, RonnyOlofsson, Thomas
By organisation
Department of Applied Physics and Electronics
In the same journal
Energies
Energy Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 591 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

doi
urn-nbn

Altmetric score

doi
urn-nbn
Total: 367 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