Bioclimatic classification for building energy efficiency using hierarchical clustering: A case study for Sri Lanka

Abstract

Climate zoning is fundamental to improving building energy efficiency through bioclimatic design principles and passive design strategies. In Sri Lanka, buildings consume approximately 37 % of primary energy and about 33 % of total generated electricity. This study aims to carry out bioclimatic zoning and estimate the passive design potential for each bioclimatic zone. Utilising dimensionality reduction through Principal Component Analysis (PCA) and Hierarchical Clustering on 31 years of weather information, Sri Lanka is classified into three distinct bioclimatic zones, namely Z1 (Hot and humid), Z2 (Cool and humid) and Z3 (Warm and humid). An hourly method was used for each zone to determine HDD and CDD. Accordingly, regression models were constructed (R2 > 0.77) to estimate the CDD for each zone for cooling load. Bioclimatic potential analysis is carried out to link climate parameters with building performance. The results demonstrate that natural ventilation is the most effective passive design method with potential 64 ± 13%, 63 ± 10, and 83 ± 4% in a year for hot and humid, cool and humid, and warm and humid zones, respectively. The results emphasise the importance of using bioclimatic classification to design energy-efficient buildings to achieve national energy efficiency goals in low-income countries like Sri Lanka.