Occupant related household energy consumption in Canada: Estimation using a bottom-up neural-network technique

A national model of residential energy consumption requires consideration of the following end-uses: space heating, space cooling, appliances and lighting (AL), and domestic hot water (DHW). The space heating and space cooling end-use energy consumption is strongly affected by the climatic conditions and the house thermal envelope. In contrast, both AL and DHW energy consumption are primarily a function of occupant behaviour, appliance ownership, demographic conditions, and occupancy rate. Because of these characteristics, a bottom-up statistical model is a candidate for estimating AL and DHW energy consumption. This article presents the detailed methodology and results of the application of a previously developed set of neural network models, as the statistical method of the Canadian Hybrid Residential End-Use Energy and Greenhouse Gas Emissions Model (CHREM). The CHREM estimates the national AL and DHW secondary energy consumption of Canadian single-detached and double/row houses to be 248 PJ and 201 PJ, respectively. The energy consumption values translate to per household values of 27.8 GJ and 22.5 GJ, and per capita values of 9.0 GJ and 7.3 GJ, respectively.

Research highlights▶ Neural network for the prediction of occupant energy consumption. ▶ Separation of appliance and lighting loads by heat allocation and energy source. ▶ Application of end-use profiles for inclusion of occupant energy consumption within an hybrid model. ▶ Estimation of national residential occupant energy consumption. ▶ Comparison with other modeling techniques and per-capita estimates.