In this study, I utilized Python, DIVA, Flux, and Genetic Algorithm(GA) to develop shading device and optimize the panels for the shading device. The following figure shows the rendered shading device.
Rendered shading device
This project consists of four parts, which are Cosine wave surface, the skin of the shading device, using excel to real time modify the panel of the shading device, and daylighting analysis and optimization.
The structure of the shading device
1. Cosine wave surface
In this study, I used the grid of points and individual anchor point to create the cosine wave surface. The number of the points can be modified in order to change the surface. The amplitude and period could also be changed to modify the height and period of the cosine wave surface. The simple python script only controls the Z value of the points for the grid.
The grasshopper nodes for the cosine wave surface
The simple Python script
2. The skin of the surface
The skin of the surface consists of two parts, which are the developed surface and the rectangle structure. For the developed surface, the original cosine wave surface is divided into multiple sub-surface by different U and V count and deconstruct each of the Brep into four points. Three of the four points are treated as the anchor points and find the closest points on the surface. The other point is treated as the modify point. The offset of the modified point can be changed to control the opening for the panels on the surface.
The grasshopper nodes for the skin of the surface
3. Using excel to real-time modify the surface
In this study, Flux is used to real-time modify the surface by using Excel. Flux is a plug-in for grasshopper, Revit, Dynamo, and Excel. The following figure illustrates the workflow of Flux.
The workflow of Flux
In the grasshopper, I send the flattened list panel data and geometry data to the Flux server using Flux node. In the Excel worksheet, I download the flattend list panel data from Flux server and update the panel data with 0,1,2,3 number pattern. I sent the updated number pattern back to Flux server. In the grasshopper, I set up a recive node to recive the updated data. Thus, the excel can real-time control the panel. In this study, I used different color to reprsent 0,1,2,3. You also can use different material or different layer to real time change the panel.
The grasshopper nodes for the Flux
4. Daylighting analysis and optimization
In this study, DIVA is utilized to analyis the daylight factor in grasshopper. The daylight factor is the ratio of the light level inside to the light level outside:
DF=(EI/EO) x 100%
The daylight factor will be projected to the surface under the shading device. The flowing figure shows the structure of the the daylighting analysis by using DIVA
For the daylighting optimization part, Galapagos is used to optimize the daylight factor for the shading device. The fitness of the Galapagos is the minimized diference between target daylight factor and the averge number of the daylight factor from DIVA. In study, daylight factor 5 is defined as the target daylight factor. In order to reduce the calculation load for optimization in this study, two surfaces with different UV count(5x5, 8x8) are tested. Both of the test have the result with flat surface and the result show restrain tendency.
The genome for the optimization
The fitness of the optimization
The result of 8x8 UV count panel
The result of 5x5 UV count panel
The optimized offset number and amplitude for 8x8 surface
The optimized average daylight factor for 8x8 surface
