Nastran Incad hydrostatic loading on curved surface.

[dave.stensby]

Greetings,

I'm new to this program, but I can't get the correct hydrostatic loads on a curved surface.  I made a simple test which is shown below. It has a 10 inch radius and is 10 inches wide.  The water is applied to the curved surface and the water surface is at the top of the piece.

Hand calcs show that the lateral force on the piece should be 18.1 lbs and the vertical force 28.35.  The vertical force can also be found from the displace volume in water.  The results from Nastran are 19.1 lateral (close)and 24.5 lbs vertical.  Constraint 2 has very low reactions as it should since the resultant should be through the center of the arc surface, which is the location of Constraint 1. This is using the Quadratic Option.  When I run it with LInear equations it has no apparent applied loads.  When I run it with the Linear option it gets lateral reaction of 18.7 (not too bad) and vertical of 21.8. (Way off).

Am I doing something wrong?  It seems like this should be an easy to accomplish loading. The documentation does not seem to discuss a vertically curved surface for pressure loads.

Thanks,  for any advice.

Dave

Test model 10" radius 10" wide in water.

[g.ceruti]

Hi Dave,

for Hydrostatic pressure distribution load, I think you should select "Linear (equation based)" as interpolation method, and specify 3 points on the curved surface.

Take a look at this:

http://help.autodesk.com/view/NINCAD/2018/ENU/?guid=GUID-AA14E0BC-95DD-4AC0-A92A-A1AEB75EE1E3

regards

Giorgio

[dave.stensby]

Giorgio,
Thanks for the quick reply. I already read the article you referenced. It doesn't really address the issue of a curve in the plane of the of the load variance. There are also YouTube videos that show the method I tried. Are they getting the wrong answer?
I tried your solution of adding another point. Results still significantly wrong. I then added two additional points. The linear equation option still produces no load (not even arrows on the screen). The linear and quadratic options still produce wrong answers. I am pretty sure I am working with the right solution. Had another engineer here look at it.
We are civil engineers and thought we would try Inventor. The project is a large steel gate. The inventor part actually worked pretty well for us. Nastran also seems pretty solid. However for this to be useable for us, a reliable hydrostatic load is a requirement. It should be easy. The distance below the surface should provide the pressure load to the element. Only the surface elevation and density need be input. Gravity if you wish, but we generally work here on the surface of the earth.
This product seems aimed primarily at mechanical engineers. Are there other firms using it for civil type structures? Is this an target market sector. We working for a large client, and they are very interested in how using Inventor works on the project.
I should also say that we are on the early version of 18. The program is fluky and hangs, crashes, etc. I doubt that this sort of issue explains the computation problems we are having. We are working on updates, but encounter organizational issues at every turn. Maybe next week.
Thanks for your help, Dave

[g.ceruti]

Hi Dave,

I' m sorry, but yesterday I suggested you something without trying it.

That article seemed to me to be clear, however InCad  does not work as should do.

I have just build something similar to your model, with the following boundary conditions:

• TZ constraint on the top square surface
• Sym X constraint (TX, RY, RZ) on lateral square surface
• Sym Y constraint (TY, RX, RZ) on front quarter circular surface

If you follow that article, using 2 points for a linear, 3 points for a equation based and 2 point for a quadratic interpolation method, InCad always provides unexpected load distributions.

They change in someway also in widthwise direction, rather than only along height.

I tried also to select the four vertexes of the curved surface with all interpolation method, then all load distributions change, however all of them are again unexpected load distributions.

Again, they change in someway also in widthwise direction, but now in a symmetrical way.

Furthermore, now linear and equation based methods seems to be the same, and provide the same SPC summation results.

I also tried creating a quadrilateral shell mesh on the curved surface, fixing all four edges, applying the same pressure load distribution with the three methods, but again I gain unexpected load distributions.

I don't know if the latest version work fine. I'm using NIC 2018.0.1.189 (with HotFix1)

We have to wait someone from Autodesk to solve this trouble.

[dave.stensby]

Giorgio,
Thanks for spending the time to confirm this issue. Has this been passed on to Autocad? I'm not familiar with the system here.
Dave

[g.ceruti]

I think that this open topic on the forum, is itself a warning to Autodesk team.

Maybe @John_Holtz  can confirm this.

[John_Holtz]

Hi,

The variable load does need some improvements:

• The load distribution or magnitudes are only approximately correct according to this article: Variable load gives wrong result in Autodesk Nastran In-CAD.
• The dialog is hard to use because it is not clear how to use it. I am working with the developers to re-design the workflow.

[dave.stensby]

John, Thanks for the reply. Is there a non-forum email address I can reach you at?
Thanks, Dave Stensby

[David_Kind]

@John_Holtz how many developers does the Nastran-in-CAD team have?