 |
| Source:https://www.ml-science.com/k-means-clustering |
Something about machine learning gets me excited because I think one of its greatest applications lies in the engineering design field, especially in the early stages. A true major time saver!
In one of the projects in Africa, a very big hospital, the client required that in each floor, fire alarm system devices be divided into loops and to which Fire Alarm Control Panel they connect.
The straight forward way would be is to open each floor and in a very slow Revit model, due to its size, select devices, count them, filter them, tag them, and then repeat the process for the next 200 devices.
Yet, K-Mean clustering does it in seconds according to X-Y coordinates, and achieves the task seamlessly.
Using the ready library for K-Mean clustering, over 6000 devices were clustered and tagged in seconds!
What's next for machine learning in engineering design?
Let's see, for I believe, it is most suitable for these applications, not for high uncertainty areas.
Peace,
Edmond