Hourly occupant density prediction in commercial buildings for urban energy simulation

Abstract

Building occupant density is a key factor influencing urban energy consumption. However, it is difficult to predict and thus often simplified to be an assumed and fixed number in urban energy simulation. In this study, the hourly occupant densities of ten representative commercial buildings in Nanjing, China were measured. The pattern of the hourly occupant density was analyzed and the key parameters defining the pattern were identified. To expand the measured hourly occupant density pattern to thousands of commercial buildings in Nanjing, five predictors, namely function, accessibility, population, business diversity, business density, were proposed. Big data technique was used to obtain the value of the five predictors for more than 3000 commercial buildings. A regression analysis was conducted to establish a model linking the five predictors with the parameters defining the hourly occupant density pattern. The methodology developed provides an effective means to predict the hourly occupant density of buildings and thus substantially improves the accuracy and reliability of urban energy consumption.