Real Estate Planner
Can Yang, Jialu Deng, Qingyangyu, Wenxuan Xu, Yinlei Pang| Generative Design |Spring 2022 | GSAPP
1.Introduction
In real estate design, site plan design is tedious and highly repetitive work. Whenever the area index, single building plan and building floor height change, it often means the revision of the site plan. At the same time, the rapid economic and social development has led to the increasing shortage of urban land, and the urban development model has gradually changed from a plane extension to a three-dimensional expansion, and the intensive utilization of land resources has become an inevitable trend. The design of the strong discharge scheme in the residential area is helpful to save the land for construction projects and achieve intensive construction. In the existing strong emission reduction design, the designer subjectively makes strong emission reduction design decisions based on the results of sunshine simulation analysis. However, in high-density residential areas, the sunlight and shadows of buildings are seriously blocked by each other. The method of manual trial-and-error adjustment makes the scheme design inefficient and is limited to the architectural design cycle. More intelligent and efficient design methods are urgently needed. The original design intention of our project is to use procedures instead of manual work to carry out these cumbersome works so that the site plan can carry out real-time feedback according to the changes of individual buildings. With further discussion and thinking, we try to explore whether this project can not only replace human labor but go further to achieve the goal that cannot be achieved by human labor and provide better choices for design. Taking the sunshine duration, an important evaluation standard of residential design, as the starting point, we try to complete a tool that can be applied to different sites through this project, which can provide the optimal solution through multiple simulations according to the specific conditions of different sites.
2. Methodology
2.1 Design Generation Logic of Model
The project aims at generating solutions facing various different building sites while achieving the best daylight conditions and profits. According to this idea, the generation of building volumes on a given building site can be divided into four steps.1-Generate general logic for the layout of the building complex.2-Adjust the building complex as a whole, including positions of the complex, different rotation angles, total number of units needed, etc.3-Add the shadow overlap area of the building complex under certain daylight conditions as a measure of daylight conditions.4-Generate the best options and let the user choose the one that better suits The option that better suits the needs.
2.2 Input Parameters
Site model — Inputs from the site include the boundary of the site where the building is generated and the massing of the surrounding buildings.
Parameters for generating buildings — Inputs which determine the basic condition of the buildings, including the footprint of the building unit, distance between each unit, floor height of the buildings, total numbers of levels and the maximum and the minimum numbers of the levels in each building. Setting these parameters can help us to initially filter the possibilities that need to be calculated and to more efficiently generate the design later.
Solar Condition — Inputs which determine the site location for positioning the sun position and sun direction, including the latitude and longitude of the site, and time zone.
2.3 Iterations Processing
Layout of the building complex — by offsetting the building footprint with half of the distance between each unit, we get the cell for Box Array.
Position the building complex — the building complex is first moved to the center of the site, then shifted to the two directions of the complex in random distances, and at last rotated in a random angle for a final position.
Number of floors in each building unit — with inputs of total numbers of levels and the limits of the maximum and the minimum numbers of the levels in each building, python will help offer a set of numbers showing how many floors are in each of the building.
Massing of the building complex — units are first sorted by their positions on north-south direction, and the numbers from the last step are applied to them based on the principle that the unit closer to the north have more floors so the complex will have a better daylight condition.
2.4 Performance Metrics
We use daylight conditions as a performance indicator for filtering the generated models to help produce better building layouts.
Daylight Condition (Objectives) — When analyzing the daylight conditions, we need a quantifiable data to determine the good or bad daylight conditions. We use the shaded area of the obtained model and the area of the building repetition as a parameter for the determination. By comparing the sum of the repeated shaded areas at three different times (9:00 am, 12:00 pm, and 3:00 pm), we can determine which solution has better daylight conditions.
In this part of the generation process, we first perform the input of the longitude and latitude of the site and the time zone and determine the parameters of the sunlight in the rhino model by writing a Python script. We then overlap the boundary of the generated shadows with the boundary of the building itself by using the mesh shadow battery group to get the area of the overlapping part.
Subsequently, we use cell group operations for each of the three different times to obtain three overlapping area parameters, and add them together to obtain the overlapping partial area parameter of the final layout as a quantifiable index for judging the daylight conditions of each layout.
The objective can then be calculated to find the possible layouts for optimization. The optimized design options are visualized by connecting the grasshopper batteries to Discover.
3. Results
We got the optimized designs after running 20 generations with 20 designs generated per generation. With the X-axis representing different building groups and y representing the total exposed area at 3 different times of a day, the optimal building groups appear when the exposed area is largest. Exposed area represents the difference of building groups’ footprint area and the shadow area. The larger the exposed area, the better light conditions buildings get.
The results reach our goals that the grasshopper and Discover could generate multiple building groups’ arrangements and calculate each arrangement’s total exposed area, letting the designer choose the optimized arrangement with the best light conditions.
4. Conclusion
In conclusion, the optimization process demonstrated that it is possible to generate multiple residential layouts by Rhino, Grasshopper, Python, and Discover and to optimize plan layouts with daylight conditions. We use the software to generate a series of layout schemes in a short time by adjusting the Angle, spacing, width and height. Then, the light conditions in three time periods of each scheme in a day are calculated to get the optimal solution. Through these calculations, designers can save a lot of time in simulating residential light conditions after design, and developers can maximize profits under limited conditions. At the same time, because all variables are artificially input, this system can be applied to different sites and different projects. It is universal and adaptive. While there are still areas where we need to explore further, the overall design has met our intent and produced practical results that are useful in the real world.
