Abstract

Domestic hot water (DHW) consumption in dwellings can play a key role in the development of policies that are focused on energy poverty, and in improving energy efficiency, among other aspects. There is an important variability observed with DHW among different countries due to technical, sociological, climatic, and economic factors. Most studies that deal with DHW pre-dictions are based on stochastic models, and only a few apply time series or statistical methods. In the case of Chile, the country is undergoing a policy development process, and there is little information about DHW consumption. As a result, it is fundamental to have DHW consumption prediction models that are focused on dwelling. For this reason, the study analysed the possibility of using time series models to make future estimations about monthly domestic hot water (DHW) consumption. To this end, consumption data obtained from 98 apartments between 2015 and 2021 were used, and 3 approaches were applied namely, exponential smoothing, basic structural model (BSM), and state-space model (SSM). The results showed that exponential smoothing and state-space methods allowed to obtain satisfactory results with regard to percentage error and confidence levels. Therefore, these models could be used to make future estimations of domestic hot water (DHW) consumption.