Paper Conference

Proceedings of eSim 2010: 6th Conference of IBPSA-Canada

     

EVALUATION OF DAILY ENERGY SIGNATURE MODELS FOR HOUSES HEATED AND OR COOLED BY ELECTRICITY: CASE OF OSHAWA, CANADA

Demba Ndiaye, Kamiel Gabriel

Abstract: Energy signature models are static identification models that can be used to normalize energy consumption with respect to weather, to predict the energy consumption, to estimate savings after building retrofits, or for building parameter estimations. In this paper, the main energy signature models available from the literature are tested using data on all-electric houses in the city of Oshawa (Canada). The models tested are: (i) a two-parameter model that can be used for all houses; (ii) a three-parameter model called PRISM that is appropriate for heated-only houses; and (iii) a five-parameter model called Extended PRISM that is usable with airconditioned houses. The results of the tests reveal that the PRISMtype models have a better prediction performance than the two-parameter model. Variants of the three models using the outdoor wet-bulb temperature in place of the outdoor dry-bulb temperature, or adding the solar gains as an explicit input to the model, are also tested. It is found that replacing the outdoor dry-bulb with the outdoor wet-bulb temperature in the calculation of the cooling degree-days for the extended PRISM model is advantageous. The extended PRISM model is found to be the best predictor. However, it can be improved and an avenue for doing so is proposed.
Pages: 72 - 79
Paper:
esim2010_3A2_072_079