An integrated building energy performance evaluation method: From parametric modeling to GA-NN based energy consumption prediction modeling

Abstract

Building energy performance evaluation, as an important process in a sustainable building design, has important consequences for global energy conservation and environmental protection. The traditional methods to perform this evaluation are usually time-consuming and computationally complex, and have high requirements for designers’ professional knowledge on architectural physics and software operation skills. To solve these problems and provide rapid, user-friendly, and more accurate prediction results, this study presents an efficient building energy performance evaluation method which integrates building information modeling, energy simulation, and energy consumption prediction together. This method follows a three-stage research framework: Stage 1 proposes a rapid 3D building energy modeling process according to the parameterized setting, Stage 2 generates numerous simulation results automatically by EnergyPlus, and Stage 3 develops the user-friendly building energy consumption prediction model with the help of the Genetic Algorithm-Neural Network (GA-NN) and provides the energy performance level of the building design after the prediction. A case study is carried out to present the overall process and verify the accuracy of the proposed three-stage building energy performance evaluation method. This study contributes to the improvement of both the extensive dataset establishment and the operational efficiency of building energy consumption prediction. It can provide designers with a real-time, user-friendly, and reliable building energy consumption prediction tool and an energy performance assessment basis in the design phase of construction projects.