Abstract
Heat pumps (HP) are becoming increasingly popular due to their ability to heat and cool buildings efficiently. A comprehensive heat pump performance analysis is critical to ensure they are operating at peak efficiency, which is essential to optimise operational running costs and deliver comfort, longevity of the system, and other environmental benefits. The efficiency of Heat pumps is conventionally measured as the Coefficient of Performance (COP), which refers to the ratio of heat output to electrical energy input. This paper aims to understand the role of various operational factors (e.g. building, demographic, HP operations) in understanding the performance gap and classifying the efficiency of heat pumps. This is achieved by applying Self-Organising Map (SOM), one of the widely applied unsupervised Machine Learning (ML) classification approaches. The proposed methodology investigated 185 dwellings with Air Source Heat Pumps (ASHP) and utilises BEIS-measured datasets and metadata. For assessing the performance gap, the daily mean of heat pump COP, operational time, cycles and part load were included to derive quantitative features using the measured 185 dwellings dataset and then fed to the SOM machine learning model for clustering the HP
efficiency and performance gap relative to their accredited position. The model performs reasonably well in clustering the 185 dwellings across the three specified categories for their performance under the given building, demographic and operational factors.
efficiency and performance gap relative to their accredited position. The model performs reasonably well in clustering the 185 dwellings across the three specified categories for their performance under the given building, demographic and operational factors.
Original language | English |
---|---|
Title of host publication | uSIM2024: Shaping net zero policies with building simulation |
Publication status | Published - 25 Nov 2024 |
Event | uSIM2024: Shaping net zero policies with building simulation - Edinburgh Climate Change Institute (ECCI); University of Edinburgh, Edinburgh, United Kingdom Duration: 25 Nov 2024 → 25 Nov 2024 https://usim2024.org/ |
Conference
Conference | uSIM2024: Shaping net zero policies with building simulation |
---|---|
Abbreviated title | uSIM2014 |
Country/Territory | United Kingdom |
City | Edinburgh |
Period | 25/11/24 → 25/11/24 |
Internet address |