Automatic direction for objects issued from revolution

[Anonymous]

Hello,

I'm having troubles with the sollicitation map in cartographic view. It seems that the automatic direction of the objects does not match with the local axis.

The structure is a water tower, the fem of the dome and wall of the tank is obtain by revolution of lines and arc:

Here is a screenshot of the software output (load: water pressure, automatic direction, wood and armer Nx):

 

Here is another screenshot of the software imput, but in a polar axis, Ny:

 

I obviously have troubles to obtain the results in the local axis of the objects.

Am I doing something wrong? What can I do to solve the issue?

Many thanks for your answers!

[Pawel.Pulak]

Rafal has shown in the video how to set local Z axis.

As concerns local X for automatic direction setting by default it is set in such way that local X axis in specific location is rotated about local Z axis in this location to minimize the angle between it and the local X axis of the whole panel (displayed when the local coordinate system of the panel is displayed - see your screen captures above.

For revolved and generally curved/non-planar panels this behavior results in "preferring" some direction close to local X of the panel - the effects for conical surfaces are visible on your second screen capture.

There are 2 possibilities to obtain the output of results in more user-friendly (radial + circumferential) directions:

1/ setting polar directions for results

or

2/ using Geometry > Properties > Local Panel directions and setting local X in vertical direction (as on the screen capture below)  - in such case X for Automatic direction of results will correspond to radial (polar) direction. In such case Y will be circumferential.

---------------------------------------------
If this post answers your question please click the "Accept as Solution" button. It will help everyone to find answer more quickly!

Regards,

[Pawel.Pulak]

If my explanations above are not clear enough here is the video showing it:

---------------------------------------------
If this post answers your question please click the "Accept as Solution" button. It will help everyone to find answer more quickly!

Regards,

[Anonymous]

Thank you for the answers,

To sum up, the stress map in local axis is not possible for some geometries.

In my opinion the manipulation to get the circumferential stress matching with local axis (changing to polar axis in the results or "changing the local x in vertical direction") leads to an other problem for shear stress (/local z) and compression (/local x); for exemple, shear stress shall be expressed in a perpendicular axis of the section. So the solution is to change the axis to analyse circumferential stress, and then coming back to automatic direction for Nx and T?

[Pawel.Pulak · Accepted solution]

Hello,

it seems that I was not clear enough

I will try to explain the way of setting directions for displaying results in surface finite elements:

1. For any setting of directions (including polar) local Z is always perpendicular to the surface of finite elements
2. Then local coordinate systems of finite elements are rotated from their default position (resulting from order of nodes defining specific finite element) about their local Z to minimize the angle between local X and some direction. This "some direction" can be the property of the panel in case of "Automatic direction" or it can be defined as "Cartesian" or "Polar" in "Selection of direction" window. So this operation sets  the direction of local X for displaying results. (Additional note: if specific finite element or whole panel is perpendicular to selected direction then it is not possible to select the direction of local X minimizing the angle (because it is always 90 degrees) so results are not displayed for such element or panel and "incorrect direction" is displayed in the table of plate and shell results)
3. The direction of local Y results from the direction of local Z (see point 1. above) and local X (see point 2. above)

As concerns my previous explanations and setting the direction to polar in results or using automatic direction with local X set according to vertical direction for the panel the above procedure will result in the following directions for results for vertical cylinders and cones composing your structure: X will be radial (or along generating line of cylinder or cone), Y will be circumferential, Z will be perpendicular to the surface.

Some hint - the "automatic" directions of results can be seen for finite elements when "FE local system" is displayed TOGETHER WITH "Local system of panels". When "FE local system" is displayed WITHOUT "Local system of panels" then the default positions of local coordinate systems of finite elements (see point 2. above). By default blue corresponds to X, green to Y and red to Z. See the video below:

---------------------------------------------
If this post answers your question please click the "Accept as Solution" button. It will help everyone to find answer more quickly!

Regards,

[Anonymous]

Hello,
It was clear, I wanted to be sure that local Z stays perpendicular to the surface of finite elements.
Thank you for your help!