The Digital Detective for Messy Blueprints
TONE: Materials/Chemistry (Inventive and Practical)
Have you ever tried to build a giant Lego castle, but realized halfway through that some of the bricks were actually painted marshmallows?
That is exactly the nightmare city planners face when they try to use computers to check if a new skyscraper is safe to build.
The Problem: Noisy Models
Architects use a special computer language called BIM (Building Information Modeling)—think of it like a 3D digital blueprint that knows exactly what every piece of a building is.
In a perfect world, a computer could read these files and instantly say "Yes" or "No" to a building permit. But in the real world, these files are a mess.
Architects often make "noisy" models. This means the digital files are full of "static" or mistakes, like a radio station that won't come in clearly.
Inside these files, a floor might be labeled as a wall, or a giant support beam might be floating where it shouldn't be. These are called semantic inaccuracies, which is just a fancy way of saying "giving things the wrong names."
Scientists in the Netherlands decided to stop waiting for "perfect" blueprints and built a tool that acts like a digital detective to find the truth.
Francesca
Noardo
The issues investigated and addressed in the paper are sometimes part of the experience of professionals and researchers dealing with BIM from practice, but there was very little evidence of this in the scientific literature, bringing to the wrong belief that any IFC-compliant model, exported by any IFC-Certified software could be directly used within any other IFC-based tool.
The Investigation: Math as a Lie Detector
The team tested their "detective tool" on two massive skyscrapers in Rotterdam: the Peak Tower and the Terrace Tower.
Instead of trusting the labels on the digital bricks, the tool used math to "look" at the shapes. It sliced the buildings every 20 cm—about the length of a new pencil—to see how wide they were.
The Logic
The tool was incredibly smart. If one floor was 5% wider than the one below it, the computer realized, "Aha! This is a different section of the building!"
The Findings & The Catch
The Evidence
The digital detective calculated the Peak Tower was 103.47 m tall, and the Terrace Tower was 100.46 m. It even counted 57 parking spots inside the digital model of the Peak Tower.
The Red Herring
The "messy" blueprints almost tricked the computer. Some walls were overlapping, and one building's digital coordinates were so wrong that the file claimed the Dutch tower was actually sitting in the middle of the United States!
Key Takeaway: The scientists realized that while their tool is a great "clutter-cleaner," we aren't ready for fully robot-checked buildings yet. The tool only knows the rules for one city right now, and it still needs to learn how to "see" the actual ground and hills around a building.
Until architects start labeling their digital Legos correctly, we’ll need these smart "detective" algorithms to double-check the work.
Reference:
Noardo, F., Wu, T., Arroyo Ohori, K., Krijnen, T., & Stoter, J. (2022). "IFC models for semi-automating common planning checks for building permits." Automation in Construction. DOI: 10.1016/j.autcon.2021.104097.