Seam when Revolving Sketch with Spline

[DRoam]

I made a part using the revolve tool and found that even though the extent was set to Full, there was a seam in the revolve feature. After some experimenting I found that any time I revolve a sketch containing a spline, this seam appears.

Has anyone else seen this, and does anyone know of a way to fix it? It's not a huge deal, except the line looks ugly and detracts from the presentation of the assembly.

Thanks!

Version: Autodesk Inventor 2013 Build 176

[JDMather]

---

@DRoam wrote:

... It's not a huge deal, except the line looks ugly and detracts from the presentation of the assembly.

 ...

---

Turn off Shaded Edges.

or

rotate the view till you can't see it.

[Oesh01]

This is an issue when it comes to exporting out the model and using it in ANSYS... it causes errors in meshing.

[JDMather]

Attach the file here - maybe there is some way to replace the surface with a Boundary Patch.

[Oesh01]

Please see attached file

[JDMather]

How precise is your manufacturing process?

What is the tolerance on your design parameters?

See attached "solution".

[Oesh01]

I know where you are going with this... and an arc is not an acceptable method for our design intent.  I must be a controlled spline to get the profile that is required.

[DRoam]

That's an interesting alternate method, JD. It creates a clean, seamless surface. Unfortunately, it does not at all use the spline profile to define its shape. Using the Boundary Patch method only lets you define a sort of ratio for the slope of the profile, it doesn't let you select a specific profile that you can force the surface to conform to.

Looking closely at the surface profile and the desired sketch profile, you can see the difference:

So this may or may not be a solution for you, @Oesh01. I know it isn't for us.

[Oesh01]

Unfortunately this will not work.  

[JDMather]

I obviously knew the difference before I posted a file.

My experience working out on the shop floor, and my experience with analysis software gives me enough confidence to wager that it is "close enough".

I guess you will need to find some other software to meet your criteria.

[DRoam]

---

@Anonymous wrote:

I obviously knew the difference before I posted a file.

---

I assumed so, was just clarifying in case Oesh01 wasn't aware, because I wasn't either until doing some investigating.

---

@Anonymous wrote:

My experience working out on the shop floor, and my experience with analysis software gives me enough confidence to wager that it is "close enough".

---

You're probably right in this case, but in some cases, such as with ASME dished heads, an elliptical assumption (which I think the modeler is using to create the boundary patch) can have a deviation from the required profile by as much as half an inch, maybe even more.

[DRoam]

I remember in our case, we had used a FARO arm to get the real-life profile of our dished head, and then wanted to use a spline to aproximate the real-life profile we measured, so that the profile wasn't composed of dozens of little lines and arcs as created by the FARO measurement.

I've attached a quick example I made up (v2016) that's similar to how we'd like to model our dished heads. The seam in the inner and outer curved faces is obviously an issue with the modeler and shouldn't be there. It would be nice if there were some way to do a "stitch" and remove the seam.

[JDMather]

My next thought is,

the mesher is going to mesh the face anyhow, so maybe something as simple as Splitting the face in a symmetrical way to help the mesher is all that is needed.

Most any part is multiple faces anyhow.  I don't think it would be possible to design much of anything without multiple faces.  Try the idea of Splitting the face to help the mesher achieve a symmetrical solution.

[DRoam]

---

@Anonymous wrote:

My next thought is,

the mesher is going to mesh the face anyhow, so maybe something as simple as Splitting the face in a symmetrical way to help the mesher is all that is needed.

Most any part is multiple faces anyhow. 

---

True. You could try something like this, @Oesh01 (see attached part). It's a 180-revolve and then an mirror so that internal and external dome faces are each split into two surfaces, rather than one split surface like a full revolve makes. This may allow you to mesh the part.

You could also try just splitting the face, but since there's an unwanted split already, I think doing a 180+mirror is less likely to result in dirty geometry.

Still not as nice as the seamless surface that a revolve without a spline would make, but will hopefully allow you to mesh and perform your simulation.