Finding Patterns of Construction Systems in Low-Income Housing for Thermal and Energy Performance Evaluation through Cluster Analysis

Energy consumption in buildings depends on many factors, including envelope materials. This paper aims to propose a method based on cluster analysis for finding reference models based on actual construction systems of low-income housing. Such reference models can be used in future thermal and energy performance studies. Data on the envelope composition of a sample of 106 dwellings were obtained through a field survey in Florianópolis, southern Brazil. Cluster analyses were performed to group similar materials and construction systems together, and therefore, a reference model was obtained for each cluster. Computer simulations and hypothesis tests were performed to verify whether the reference models represented the sample. Three reference models were obtained from the cluster analysis. Cluster 1 comprised houses with ceramic-brick walls, concrete floor, and concrete slabs. Cluster 2 comprised houses with ceramic-brick walls, concrete floor, ceramic tiles, and wooden ceilings. Cluster 3 comprised houses with wooden walls, wooden floor, cement tiles, and wooden ceilings. Cluster 1 performed better than the other clusters in the cold season (mean degree-hour was 1299 for cooling and 1361 for heating in the reference model). Cluster 2 performed better the other clusters over the hot season (mean degree-hour was 1014 for cooling and 1451 for heating). Cluster 3 showed the worst performance (mean degree-hour was 3793 for cooling and 2988 for heating). Thus, the hypothesis tests have shown that the three reference models differ from each other and can represent their clusters properly. Cluster analysis was a practical and objective method for obtaining reference models.