How to find an interior space with the best view
Purpose : The view changes depending on where you are in the existing building. The goal is to visualize which locations have the most views and the widest views.
Initial Process
1. Insert a floor plan of an existing building.
2. Place a uniform point inside the building.
3. Count the number of lines going out from a point toward windows, doors, etc.
4. Visualize the color of the dots based on the number of outgoing lines.
5. See the field of view change in real time as you add walls.
Process
1. Create a grid and delimit only those grids that are inside the building’s outline.
2. Use Isovist to measure the number of points touching the outer and inner walls from the center of each grid cell.
3. Visualize the number of walls touching the cell center point in a gradient from lowest to highest.
4. To emphasize the thickness of the walls, visualize the single lines of the outer and inner walls with a thickness.
Next, we assumed that the space above is an exhibition space and wanted to apply a generative design to maximize the viewing space as walls are added inside.
Galapagos leveraged a program called Discover to automate the process, which was simple but slow to compute and difficult to store and visualize data.
5. Create an algorithm to connect two random points inside to create five walls.
6. We’ve utilized discover to randomly generate seed values for Random reduce. We generate one of the numbers from 0 to 100 for each generation.
And Capture screenshot helps automatically capture every generation image.
7. The objective is to maximize the sum of the number of points that do not touch the wall when Isovist on the points inside. The higher the number, the wider the field of view.
As a constraint, we set that the walls inside must have a total sum of lengths greater than 10 meters. Otherwise, the graph shows hollow squares. They are automatically excluded from the calculation.
If I click the circle, I can see the number of file and image.
OUTPUT
Upper 3
Lower 3
Struggle
Initially, we measured the points from the center of the cell to the wall by connecting them with a line segment. However, it took a long time to calculate and the computer was heavily loaded, so I found a component called Isovist, which helped me to reduce the calculation speed and CPU load.
Limits
The current limitation does not address the point where the wall meets the cell, causing the cell to encroach on the wall in the visualization. I think this is caused by setting the size of the cell first and then trying to fit it to the fixed wall.
In the optimization process, the logic for creating internal walls is not clear, so there are limitations in areas such as interference with existing walls and movement. If this part is improved, more realistic logic will be created.
Plugins
Discover (*Activate upto Rhino7)
https://colidescope.github.io/discover/
Discover Youtube Guide
https://www.youtube.com/watch?v=tazRo4mMBm4
Downloads