"Center Point Loop Edges" is not a centroid?

[pcrawley]

Can anyone help me out with this problem, please? 

In a sketch, I found the centroid of the triangle using construction lines - see point 1 in the image below.

Using the Workpoint > Center point of loop edges finds a different point - see point 2 below.

(The Inspect > Region Properties centroid agrees with the point I constructed - not the Workpoint)

In the API help under "WorkPointDefinitionEnum" it says: kCentroidWorkPoint - 12811 - Edge, Edge Loop or Edges centroid work point definition. 

Does anyone know how the Workpoint > Center point of loop edges works?  It doesn't appear to agree with the centroid.

Sample part attached.

[Xun.Zhang]

Hi Peter,

This looks incorrect to me, I am sending to project team for a better understanding and to confirm the root cause.

Thanks a lot!

[pcrawley]

I don't suppose there's any news on this one is there please?  

[johnsonshiue · Accepted solution]

Hi Peter,

This one has been discussed before. What I was told is that the centroid of a profile (area) is different than centroid of a loop (wire). So the profile centroid can be found by using the medians from each side. But, the loop centroid, which loop workpoint is using, is called Spieker center (https://en.wikipedia.org/wiki/Spieker_center).

The two are indeed different. One is about the center of mass for an area but one is the center of mass for a loop.

Many thanks!

[pcrawley]

Thanks Johnson.  At least that solves the mystery.

I think what threw me is the naming in the API: centroid work point definition

Perhaps it needs renaming to spieker centre work point definition

[ian]

I have similar problem with curved slot that i want to constrain to, so I thought placing a work point in the center would work for me....not the case, apparently 'Center Point Loop Edges' works off a bounding box, therefore the more a slot is curved then the work point moves away from the center, our CAD reseller sent me the maths behind this which is this....if you can follow.....

If the middle part is a rectangle, the result will be simple, as you can get 2 symmetric axis, the center is just the their intersection point. In your case, the middle part is curved, have to do a calculation.
The X-coordinate of center of the slot should be,
X=(S1*X1+S2*X2+S3*X3)/(S1+S2+S3), S1 is the area of one part, X1 is X-coordinate of the center of that part, and so on, and same for Y.

The result for me is "Center Point Loop Edges" is no use, for curved slots. I can work around this by sketching, not ideal though

[johnsonshiue]

Hi! The center point of a loop found by Inventor is the centroid of the loop. The center of the curved slot you are looking for is more like geometric center (or midpoint of the symmetry line). Inventor is not yet able to detect it automatically.

Many thanks!

[ian]

Can we have a command for finding center of curved slots? Think of constraining a tapped hole to the center, or a bolt?

Regards

[pcrawley]

The geometric centre of a curved slot would be useful - but the geometric centre of any loop would be even more useful.  It would have solved both your slot problem and my triangle problem.  Time to head over to the developer's forum...