Estimating the energy consumption for residential buildings in semiarid and arid desert climate using artificial intelligence

• Published in: Scientific reports Vol. 14; no. 1; pp. 13648 - 13
• Main Authors: Wefki, Hossam, Khallaf, Rana, Ebid, Ahmed M.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 13.06.2024; Nature Publishing Group; Nature Portfolio
• Subjects: 639/166; 639/166/986; Architects; Arid climates; Artificial intelligence; Building energy consumption; Climate change; Construction materials; Cooling; Datasets; Design; Emissions; Energy consumption; Energy efficiency; Green buildings; Heating; Humanities and Social Sciences; Indoor air quality; Literature reviews; multidisciplinary; Neural networks; Prediction models; Residential areas; Science; Science (multidisciplinary); Simulation; Building energy consumption; Artificial intelligence; Energy efficiency
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-024-63843-w

This research aims to develop predictive models to estimate building energy accurately. Three commonly used artificial intelligence techniques were chosen to develop a new building energy estimation model. The chosen techniques are Genetic Programming (GP), Artificial Neural Network (ANN), and Evolutionary Polynomial Regression (EPR). Sixteen energy efficiency measures were collected and used in designing and evaluating the proposed models, which include building dimensions, orientation, envelope construction materials properties, window-to-wall ratio, heating and cooling set points, and glass properties. The performance of the developed models was evaluated in terms of the RMS, R 2 , and MAPE. The results showed that the EPR model is the most accurate and practical model with an error percent of 2%. Additionally, the energy consumption was found to be mainly governed by three factors which dominate 87% of the impact; which are building size, Solar Heating Glass Coefficient (SHGC), and the target inside temperature in summer.