Research on Building Form Energy Saving Design Based on Neural Network

With the continuous advancement of urbanization, the thermal environment quality and energy consumption of urban residential space have been paid more and more attention, which makes energy-saving planning and design and energy-saving building design become the focus of current research. However, due to the limited computing power and insufficient consideration of form factors, there are relatively few studies on the impact of building group form on building energy consumption. Therefore, this paper adopts the machine learning framework based on convolutional neural network to analyze the influence of building group shape on its overall building energy consumption under the action of various form factors, and verifies that different building group shape factor layout in residential areas can be obtained by changing the combined control factors under different plot ratio and building density control conditions. The results show that building energy consumption can be significantly changed by changing the form factors of building groups. The research in this paper provides an energy-saving oriented design method and analysis process of residential building group form, and provides corresponding methods and suggestions for energy-saving building design, energy-saving planning and design, which guides building energy-saving design.