Data-driven models for estimating heat pump power consumption

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Insinööritieteiden korkeakoulu | Master's thesis
Degree programme
Nordic Master Programme in Innovative and Sustainable Energy Engineering (ISEE)
The number of installed heat pumps has been rapidly increasing in recent years, accelerating the decarbonisation of the heating sector. The impact of the increasing deployment of heat pumps on the grid can be evaluated with the help of models estimating the heat pump power consumption. This thesis contributes to the field of data-driven heat pump modelling by developing regression models based on data from field installations to reflect the heat pump operation in real conditions. The developed models estimate the heat pump power consumption using a limited number of input features (parameters) measured during heat pump operation. This thesis analysed anonymised data obtained from the monitoring system of domestic ground-source heat pump (GSHP) and air-source heat pump (ASHP) installations to develop GSHP and ASHP regression models. Prior to developing the regression models, the data were pre-processed and the most important features (measured parameters) used as independent variables in the regression models were identified. Further, various regression models were proposed ranging from simple-linear, multiple-linear to non-linear (up to the fourth-degree) regression models, with and without the interaction terms and with the varying number of the selected input features. The identified most significant input features for developing the regression models based on the obtained datasets involved supply, source and outdoor temperatures and compressor frequency. The results in this thesis showed that regression models can estimate the heat pump power consumption with a satisfactory accuracy (up to R2 88 % and mean absolute percentage error 13 %). Furthermore, it was proven that non-linear regression models performed with higher accuracy compared to linear regression models and the accuracy was increasing with the increasing number of statistically significant input features. This thesis also revealed the importance of outlier detection and feature selection prior to developing heat pump models when data from field installations are used.
Lahdelma, Risto
Thesis advisor
Larijani, Hatef Madani
Natiesta, Thomas
heat pump, data-driven modelling, modelling, regression model