Abstract
Reliable knowledge of domestic hot water (DHW) and space heating (SH) energy use in buildings is essential for dimensioning heating systems, evaluating building performance, and modelling energy demand and flexibility. However, sub-metering that distinguishes between SH and DHW is rare, particularly in older buildings. This article presents a case study comparing five methods for disaggregating DHW and SH energy use and generating average daily load profiles for DHW from total heating measurements. The case study covers 19 Norwegian apartment buildings spanning three locations with both district heating and electric heating systems. A key limitation is that 17 of the 19 buildings belong to the same condominium. All methods are evaluated against measured DHW demand. The results show that DHW energy use has clear seasonal patterns in the case study buildings that may be beneficial to consider when developing average load profiles for building energy simulations. The current standard load profiles for Norwegian buildings do not capture the general shape of the average daily DHW load profiles seen in the case study buildings. A categorical boosting regression model (Method 5) and a seasonally corrected Building-Profet method (Method 3diss,corr) performed at a statistically comparable level for both DHW disaggregation and daily load profile generation and were the best performing methods overall. However, Method 5 performed noticeably weaker on the two buildings outside the main condominium while achieving higher performance within it. A broader and more diverse dataset would be needed to validate the generalisability of the results.