Prediction of City-Scale Building Energy and Emissions: Toward Sustainable Cities

Building energy use estimation relies on building characteristics, its energy systems, occupants, and weather. Energy estimation of new buildings is considerably an easy task when compared to modeling existing buildings as they require calibration with actual data. Particularly, when energy estimation of existing building stock is warranted at a city-scale, the problem is exacerbated owing to lack of construction drawings and other engineering specifications. However, as collection of buildings and other infrastructure constitute cities, such predictions are a necessary component of developing and maintaining sustainable cities. This paper uses Artificial Neural Network techniques to predict electricity consumption for residential buildings situated in the City of Gainesville, Florida. With the use of 32,813 samples of data vectors that comprise of building floor area, built year, number of stories, and range of monthly energy consumption, this paper extends the prediction to environmental impact assessment of electricity usage at the urban-scale. Among others, one of the applications of the proposed model discussed in this paper is the study of urban scale Life Cycle Assessment, and other decisions related to creating sustainable cities.