Authors
Sugiyama, Genta; Honda, Tomonori; Itsubo, Norihiro
Abstract
Air conditioner usage makes up around 20% of the total electricity used in buildings worldwide, and the IEA report shows that around two-thirds of the world’s households will be able to own an air conditioner by 2050 and the energy demand for air conditioners will increase triple from current status. The Life Cycle Assessment(LCA) of air conditioners shows that the use phase of air conditioners accounts for about 90% of the environmental impact, and the results of the calculation vary greatly depending on the local climate, operating conditions and usage time, etc. In this study, by using Home Energy Management System(HEMS) data and considering regional characteristics and housing attributes, we calculate energy consumption and greenhouse gas emissions in line with the actual usage conditions in regional scale of Japan, which have not been fully reflected in the past. In addition, in order to improve the accuracy of LCA for air-conditioners, we have used artificial intelligence to create a model that estimates electricity consumption based on the regional and residential attributes of HEMS data. In the LCA based on HEMS data, the primary data for the use phase and the results of GHG emission using HEMS data are calculated respectively, and by looking at the differences between the results authorized by industrial association and those based on actual values, the key parameters required to carry out detailed analysis in order to generate model to estimate reliable results of GHG emissions are extracted. The model for estimating electricity consumption was created using a neural network based on artificial intelligence, and the differences in forecasting accuracy between conventional regression analysis methods and selected variables were compared.