integration of surface Normals
 

Load a sphere into Paraview (poly surface data from simpleFoam)
Generate surface normals filter
Integrate Variables filter (and look at the Normal components)

I get the following integrated Normal components (1.2e-5, 1.1e-4, -3.4e-5)
in other words integrated Ny is 10 times larger than integrated Nx.
This has the effect of completely distorting the pressure components.

Say, you have constant pressure 100kPa on the surface. With these normals,
you will get Forces Fx=1.2N Fy=11N Fz=-3.4N which must be wrong for a sphere.

Surely the integrated normal components should all be similar for a sphere.

https://drive.google.com/open?id=0Bx...mZMUDBNY2gtdVU

https://drive.google.com/open?id=0Bx...mZMUDBNY2gtdVU
https://drive.google.com/open?id=0Bx...mZMUDBNY2gtdVU

fixing surface pressure integration errors
 

still not sure what the problem is with integration of pressure over surface, in Paraview but the workaround is to integrate (p-101325)*Normals...
or whatever your freestream pressure value is, (p-pinf)*Normals,
since the error can be seen by integrating the value "101325*Normals" over the surface.

Wierdly the error is also changed/reduced by clipping the surface (such that none of the surface is actually clipped).

sample surface data
 

I have attached sample surface data, in case anyone wants to replicate the issue: Generate Normals > Calc p*Normals > Integrate variable, and then look at the resulting force components.

Then do the same with Calc (p-101325)*Normals or Calc 101325*Normals, and compare the difference force components that are obtained.

https://drive.google.com/open?id=0Bx...3YtWUZqbVlLVnM