A hybrid simulation approach to predict cooling energy demand for public housing in Hong Kong

The residential sector accounts for a significant and increasing portion of the national energy use. Cooling energy reduction in the public housing sector is one of the key success factors for sustainable building development measures especially in the sub-tropics. This study proposes a hybrid EnergyPlus (EP) - artificial neural network (ANN) model, which is more flexible and time-efficient than the conventional cooling energy simulation methods, for simulating the cooling energy consumption in the Hong Kong public housing sector and evaluates the cooling energy impacts related to building materials, window sizes, indoor-outdoor temperature variations and apartment sizes. The results show that climate changes and temperature set-points have the greatest impact on cooling energy use (-19.9% to 24.1%), followed by flat size combinations (-13.4% to 27.9%). The proposed model can be a useful tool for policymakers to establish sustainable public housing development plans in Hong Kong.