An efficient surrogate model for building energy consumption estimation considering global warming scenario: a case study in Tra Vinh

Climate change has emerged as a global challenge, profoundly affecting various sectors, including building energy. Rising temperatures lead to an increase in energy requirements, and the choice of building materials and orientation influences a building's energy efficiency. This paper aims to build a surrogate model for predicting energy demand through building materials. A case study of a two-storey residential house in Tra Vinh is conducted based on the RCP4.5 climate change scenario for 2050. Firstly, training datasets are generated under the assumed global warming scenario RCP4.5, considering various materials for walls and ceiling slabs. Then, a machine learning model is developed to forecast building energy consumption, with the number of hidden nodes, hidden layers, and training algorithm optimized based on a root mean squared error objective function. Additionally, building orientations are analyzed concerning the movement of the sun. The findings demonstrate the simplicity, efficacy, and application potential of the suggested approach in enhancing the efficiency of building energy analysis, offering valuable insights for sustainable building practices in the face of future climate change.