Investigation on spatial distribution and thermal properties of typical residential buildings in South China's Pearl River Delta.

• A series of typical residential building models were defined through statistic method. • The envelope's size, shape, thermal properties and opaque feature were determined. • The indoor space partition and function area distribution were also identified. • The models could serve as a standard platform for building energy simulation research. • The method can be applied to other regions and building types with sufficient data. With the rapid growth of urban population in the Pearl River Delta (PRD) region, the research on residential energy saving has attracted increasing attention. Typical buildings are useful tools for building performance simulation of residential district, which is of great significance for local green building-related research, policy formulation as well as market orientation. In this paper, the spatial information (building size, shape, area and amount of dwelling unit area) and the thermal properties of the envelope (such as window-to-wall ratios, shading coefficient, and the heat transfer coefficient) of about 120 residential districts were collected and statistically analyzed. Meanwhile, 120 residential districts are the actual data in the PRD region since 2010. The analysis method involved K-means, histogram and bubble diagram, box plot - bin mean method and slope method, which aims to remove outlier points and find representative value for each parameter, and define a series of typical residential building models. From the analysis result of 120 residential districts, which involved around 350 types of 1300 residential buildings and over 900 types of dwelling units, 7 typical residential building models were determined to be used in other research. [ABSTRACT FROM AUTHOR]