JWACA2023-032: Quantitative Research on Public Building Energy Consumption and Building Envelope Based on Econometric Model

Abstract

Machine learning has achieved rapid development in recent years and is widely used in data recognition and processing. This article aims to explore the strategy of urban layout generation based on neural network algorithm and use the Generative Adversarial Networks (GAN) in deep learning technology to realize the automatic generation design of urban morphology. This article takes the modern European city morphology as the experimental object. First, collect the layout data of European cities; then conduct data processing that is convenient for machine learning, retain and label the main data in the city; use GAN for machine training by adjusting the parameters. With the aid of the trained model, the city form is automatically generated by inputting initial city data. This automatic design tool can help release the burden of architects in the early design stage and quickly provide an effective plan preview for urban design tasks.