collisions between objects, one of which may be located inside another (nested).
This method is usually slower than the " By Intersection " method, and the results of canada email list determining the intersection distance may differ.
What does "intersection distance" mean?
There is a " distance " parameter in the collision report .
It may seem that it means the distance to which the volume of one element "enters" the volume of another. But this is not so!
Tangl control
As we wrote above, working with triangulated geometry as physical volumes is mathematically non-trivial.
This means that calculating the distance at which one volume “entered” another is also not trivial, since no volumes, in essence, exist. There are only sets of triangles.
Let's take a very simple but illustrative example. Two intersecting spheres. On the one hand, these are very simple objects, and on the other, they have curved surfaces with many small triangles.
In our example, two spheres “enter” each other at a distance of 3 meters:
Tangl control
After transferring to Tangl we got triangulated geometry. Measurement approximately along the equator shows an intersection of ~2.995 meters :
Tangl control
However, if we perform a collision analysis “ By intersection ”, we will see a completely different number in the report: 0.209 meters.
Tangl control
And if " By volume ", we will see a distance of 2.897 meters , which is already much closer to the real value, but a little less:
Tangl control
So what is this quantity?
It all depends on the type of check.
In the " By Intersection " mode, Tangl detects triangle intersections. It calculates the intersection value of two specific triangles. It then checks the resulting values (since usually, within two objects, intersections are in different triangles) and finds the largest of them.
In other words, the intersection distances of not the spheres themselves are calculated, but the intersection of the triangles of which these spheres consist.
Tangl control
In the " By Volume " mode, Tangl no longer operates with individual triangles, but with derivatives of groups of triangles that can make up a volume or part of it.
In other words, the intersection distances are calculated not of the spheres themselves, but of the assumed simple volumes that Tangl tries to reconstruct from groups of triangles.
Tangl control
In this case, each object can result in many assumed simple volumes, and not just one, especially if the geometry of the object is complex - with protrusions, bends and holes, as in the picture below:
Tangl control
Therefore, even the intersection distance " By Volume " may differ greatly from the actual measurements.
In any mode, when testing with a negative offset (proximity testing), Tangl first checks for obvious intersections of objects. If they intersect, it returns the distance between them, but with a minus sign.
If the objects do not intersect, Tangl finds the distance between the closest triangles or simple volumes depending on the type of check.
But in Navisworks it’s not like that!
In fact, it is very much so! In everything.
In the vast majority of cases, the results of collision detection in Tangl correspond to the results obtained in Navisworks, and this concerns both the fact of intersections themselves and the determined intersection distances.
Our two sphere example in Navisworks showed a distance of 0.211 meters, which is very similar to the distance in Tangl: 0.209 meters:
Tangl control
When using the " By intersection (conservative) " mode, which is similar to the " By volume " mode in Tangl, we also get a very similar result: 2.996 meters (in Tangl the distance was 2.897 meters ):
Tangl control
These small differences are due to the pre-processing of the model in Navisworks, in particular the triangulation of curves and surfaces.
What if we take a more complicated example?
Here, for example, is the building frame. This is how it looks in Revit :
Tangl control
And like this in Navisworks :
Tangl control
And in Tangl :
Tangl control
During a standard error check of Navisworks and Tangl systems, the following results were obtained:
Tangl control Tangl control
In both products you can see the same number of collisions - 43. In fact, all these collisions are identical. If you look at the distances between the collisions, they will match in both products. ( 0.074m ):
Tangl control Tangl control
With the second type of check (Conservative in Navisworks and By Volume in Tangl) the results are not quite the same:
Tangl control Tangl control
The most noticeable thing is that in Navisworks, collisions are present almost throughout the entire model.
Tangl also found more collisions than the standard mode check, but their number was smaller than in Navisworks (61 versus over 5000 in Navisworks).
The high number of collisions occurs because these algorithms can register many false collisions, where elements touch but do not overlap. It is important to note that the Tangl algorithm registers far fewer false collisions.
If you compare the distances that are specified in the conflicts found by both products, you will notice that they reflect a certain average conventional intersection length. (in the examples below - 1 meter in NavisWorks and 2.5 meters in Tangl). In this particular case, both distances correspond to the thickness of the structures:
Tangl control Tangl control Thus, we have become convinced that although there are differences in Tangl and Autodesk Navisworks, they are due to the way the model is pre-processed, i.e. it is a matter of triangulation of curves and surfaces, which I mentioned earlier.
This example demonstrates how close the principles of operation and calculation results of Tangl and Autodesk Navisworks are. This means that there is no talk of the latter being the only alternative today.
We hope that you now have a better understanding of how clash checks work in Tangl and Navisworks and why they produce the results you see in your reports.