Old Residential Balcony Renovation in Xi ‘an 

Based on Deep Learning Network



| Teams |

| Tutors |

Zhong Ximing, Li Kun 



| Abstract |



Xi’an has a large number of old residential areas that need to be transformed, and the addition of balconies is an economical and effective reconstruction measure.

Although this measure increases the use area, it also creates the problem of sunlight and line of sight. How to meet the demand of sunshine and the preference of different residents is a difficult contradiction to balance. In the traditional design, it is necessary to investigate the residents ‘use preferences, combined with the site lighting analysis, a lot of artificial designers work subjectively to meet the lighting needs and residents’ preferences in different environmental conditions. So whether the machine can learn the preference of residents in different living locations for balcony forms, and learn how to meet the needs of sunshine and sight is our main research problem. When previous researchers faced similar problems, they only considered the building spacing and building contour texture, and the influence relationship within the overall layout. For this project, we start from the vision of the residents, and consider the influence relationship between the vision of the residents and the residents’ use of the balcony in the house type. We entered 9,000 pairs of data as samples. The machine inputs the surrounding environment and the resident questionnaire, and outputs the corresponding house type of different locations and different environments. The house type meets the needs of residents’ preferences as far as possible and maximizes the light duration, lighting preference and sight preference.



| Research |



This research project is a series of outcomes from the AI architecture course launched by Archiford and Xi’an University of Architecture and Technology. The course aims to use deep learning methods to design and advance architectural design solutions.

The main research questions in this project focus on how to choose the right form of balcony according to the resident’s preferences and site environment. As well as how to meet the residents’ preferences and environmental constraints from the perspective of balcony design in the face of renovation problems.

It was found that previous researchers have considered similar problems at a larger scale and from a broader perspective, mainly considering the relationship between influences at a larger scale.For this project, however, we considered smaller-scale influences, primarily from the perspective of the residents.




| Designing Process |



The design process was divided into two steps in total, data collection in the first step and model building in the second step.

In the first step, we conducted a questionnaire survey. We created three questions about residents’ preferences for balconies to generate a dataset on preferences.In the second step, a model was built.

In rhino, the spatial relationships between residents and windows and surrounding buildings and trees were modeled.



Leave a Comment