Exploring the potential of GC to optimize a sun hood

Hi All

I am looking for a starting point for running an optimization on a sun hood. I haven't utilized GC in this way before but I recall from a number of years ago there are some tools to run through an optimize solution within GC.

Any tips on where to start learning?

To test this I thought I would use a simple problem I have at home where I want to place a sun hood over a weather cam I have installed before the Australian summer melts it!

I would like the sun hood to be optimized to provide the best sun shade between a 2 dates and related to the direction the camera is installed. It can be as organic as it needs to be and then once modeled I will 3d print the result and install.

I can use this as a proof of concept for another highrise project where I would like to optimize both sun shading shape and location on the façade.

Any help would be appreciated

Thanks
Wayne

Parents
  • Hi All,

    I have been thinking about the criteria I would be trying to test and I have come up with a potential method of testing an option.

    Say I set the criteria for the sun hood:

    • Block all (100%) direct sun between say 1st November to the 28th February.
    • Sun Hood had to be attached to the outside of the camera box
    • Resolution of any given tested point would be 10% or 1/10 the length of the camera box. so test for 10 points
    • Time resolution of say 10min (maybe too many I think this could be around 10,000)

    I could manually create a solution by projecting lines at the required angle (based on the sun position at the given time) through a surface that was projected out from the perimeter of the box. I could then generate a point at each of these intersections.

    something like this

    then repeat for the next interval.

    Given enough time (maybe 6months) I could generate a point cloud that would indicate the required position of a sun hood on every day/hours/min tested.

    Then I could mesh the results into a form. (could even follow up with a solar analysis (solar exposure) to check the outcome)

    So now I have a very slow way of doing that manually I assume all I need to do is create a system in GC that will generate lines from a set of points along a line (on the base line) then project in a positive direction until it intersects with a potential shape, generate a point (or skip if an intersection is not found which could happen if the sun comes from behind the unit).

    Then loop this through all the required dates and times

    Maybe this is the wrong approach so happy to hear if anyone has any suggestions Slight smile

    First step I think is to try generating a line based on the Azimuth and Altitude angles or maybe generate a coordinate system at that angle first.

    I can see some great potential in optimizing some sun shading requirements on some of our projects where we could block out the sun during the most intense times of the day and have an understanding of what was the required for sun shades on any given part of the building (which we may choose to ignore based on other requirements).

    Sorry for the long post, thinking as I type.

    Thanks
    Wayne

Reply
  • Hi All,

    I have been thinking about the criteria I would be trying to test and I have come up with a potential method of testing an option.

    Say I set the criteria for the sun hood:

    • Block all (100%) direct sun between say 1st November to the 28th February.
    • Sun Hood had to be attached to the outside of the camera box
    • Resolution of any given tested point would be 10% or 1/10 the length of the camera box. so test for 10 points
    • Time resolution of say 10min (maybe too many I think this could be around 10,000)

    I could manually create a solution by projecting lines at the required angle (based on the sun position at the given time) through a surface that was projected out from the perimeter of the box. I could then generate a point at each of these intersections.

    something like this

    then repeat for the next interval.

    Given enough time (maybe 6months) I could generate a point cloud that would indicate the required position of a sun hood on every day/hours/min tested.

    Then I could mesh the results into a form. (could even follow up with a solar analysis (solar exposure) to check the outcome)

    So now I have a very slow way of doing that manually I assume all I need to do is create a system in GC that will generate lines from a set of points along a line (on the base line) then project in a positive direction until it intersects with a potential shape, generate a point (or skip if an intersection is not found which could happen if the sun comes from behind the unit).

    Then loop this through all the required dates and times

    Maybe this is the wrong approach so happy to hear if anyone has any suggestions Slight smile

    First step I think is to try generating a line based on the Azimuth and Altitude angles or maybe generate a coordinate system at that angle first.

    I can see some great potential in optimizing some sun shading requirements on some of our projects where we could block out the sun during the most intense times of the day and have an understanding of what was the required for sun shades on any given part of the building (which we may choose to ignore based on other requirements).

    Sorry for the long post, thinking as I type.

    Thanks
    Wayne

Children
  • Hi All 

    As listed above in the previous posts I have been looking into the use of generative components to test design options for sun shading devices on our latest projects. 

    I know there are probably firms out there that have highly skilled script writers and probably find the following a bit basic but I thought some might find something that can be applied to their designs. 

    I have approached it in a way that could be used in our office hopefully by most of the architects with a limited knowledge of GC. So here we go. 

    As I mentioned in the previous post I was originally trying to ask rays on a façade panel but a much easier approach would be to start from the panel and check if it is receiving sun or not.  

    So to test this theory I took my DIY weather station and it’s SkyCam as a pilot project. 

    As some of you may know the Sydney summer can be warm and I had a raspberry pi camera mounted in a weatherproof box with a clear front facing approximately north. This worked well for most of the year but then as the really hot days approached I found I had 2 camera fail. I am not sure how hot it was in the box but let’s say it was very warm!  

    So my design goal was to shade the face of the box 100%. I decided the peak time of year was probably between from the end of October to the beginning of March.  So the 1st of March would be the day I would check which should cover the shading through this 4month period. 

    So first to gather some info: 

    • Face size I wanted to shade
    • Downloaded the sun angles for every hour into an excel file. (www.sunearthtools.com
    • Determined exact orientation of the face. 

     

    The last item I determined was the sun shade would be on three sides of the face and extend perpendicular to the face. (I would like to try the optimizer to test if changing angles of the hood would make it more efficient, but small steps). 

    So I setup a script that was relatively simple and projected lines from a point grid on the face I wanted to protect at the require azimuth and altitude. Then created a series of points where the lines intersected with my sunhood surface. 

    I then linked up the excel read node to a slider so I could change the required time of day and then linked this slider to the export node so I could export the points generated at each hour giving a new dgn file for each. 

    Using a bit of Volker magic I then created a transaction to run through the slider(hours of the day) and produce a series of dgn files for each hour.

    It was then simple to reference all the dgn files into a Masterfile then create a mesh from points.

    A bit of cleanup for what I determined was the best shape. And then I 3d printed the results.

     

    From concept into reality!

    And it works perfectly. As you can see in the photo below the sun is just blocked by the side screen. I will rotate the camera soon so you don’t see the sun hood but for this exercise it proves the point.

    Now I just have to work out how to implement the Optimizer to see if the hood would be more efficient if it was not at a perpendicular angle to the surface it is protecting.

    if anyone is interested I will post the script/dgn file and excel files

    Thanks

    Wayne