Shape properties using VolumeProperties and VolumePropertiesGK

[markkulait]

There are two different functions to calculate the volume properties in the BRepGProp (GProp_GProps, GProp_PrincipalProps) class:

• VolumeProperties
• VolumePropertiesGK

Both calculate the shape volume correctly. However, when calculating the inertia properties of the system, it seems that the former produces the correct results but the latter does not.
For a box and a cylinder:
Principal moments of inertia
Cylinder
Volume: 196349.54084936203
Ixx, Iyy, Izz: (194317337.4428578, 194291128.82151097, 61359231.51543226)
Cylinder GK
Volume: 196349.5408493644
Ixx, Iyy, Izz: (6.122366410532748e-40, -2.2351741790771484e-07, -3.5762786865234375e-07)
Box
Volume: 124999.99999999999
Ixx, Iyy, Izz: (52083333.33333338, 52083333.33333331, 52083333.333333276)
Box GK
Volume: 124999.99999999999
Ixx, Iyy, Izz: (1.6326310428087332e-40, -8.940696716308594e-08, -8.940696716308594e-08)

[tpaviot]

see https://dev.opencascade.org/content/difference-between-brepgpropvolumeproperties-and-brepgpropvolumepropertiesgk

Both methods use different algorithms. GK means Gauss-Kronrod integration. To be rigourous, each algorithm results should be compared to the theoretical solution for volume/Ixx/Iyy/Izz.

[markkulait]

For a cylinder (r = 25, h = 100), Izz = 0.5mr^2, and m = Pir^2h (density = 1).
This gives Izz = 61359231.5154 which is the same as the result given by the VolumeProperties function. The result given by the VolumePropertiesGK function is way off, probably it gives something close to zero.

[tpaviot]

>>> print(brepgprop.VolumePropertiesGK.__doc__)


        Parameters
        ----------
        S: TopoDS_Shape
        VProps: GProp_GProps
        Eps: float (optional, default to 0.001)
        OnlyClosed: bool (optional, default to Standard_False)
        IsUseSpan: bool (optional, default to Standard_False)
        CGFlag: bool (optional, default to Standard_False)
        IFlag: bool (optional, default to Standard_False)
        SkipShared: bool (optional, default to Standard_False)

        Return
        -------
        float

Did you try to tune the parameters for this function?

[markkulait]

No, both algorithms were run with eps = 1e-3 and OnlyClosed = True.