Shell elements: local RS, IP location and numbering

[EricsonEncinaZ]

I need some help with the local reference system (RS) on shell elements. I defined a 3D domain with 6DOFs and I modelled an elastic cantilever t=10cm, 1mx1m shell (various mesh sizes) with an orthotropic material (E 2E 3E 0 0 0 G 2G 3G) assigned to a PlateFibre section. The load was just axial load so to isolate the behaviour in each direction. A sketch is shown here: https://drive.google.com/open?id=0B3yHw ... llhb2RzaUE. Based on the models, I figured out that the X local axis is along the line joining the nodes 1 and 2 shown in the bottom of this link: http://opensees.berkeley.edu/wiki/index ... 2D_Command, and the local Y is along the line joining the nodes 1 and 4. But I'm not sure whether the positive values increase from 1 to 2 and from 1 to 4, for each axis respectively. The reason that made me doubt came from the following observations:
1. When I used the recorder "recorder Element -xml forcesEle2.out -time -ele 2 forces" the IP 1 and 2 had positive values and 3 and 4 negative ones for the Model Y; in the case of Model X, IPs 1 and 4 had positive values and 2 and 3 negative ones (I assumed that from the 24 Ps, 6 Ps are for each Gauss IP). From force equilibrium, the numbering of the IPs should be CCW as those shown in the bottom LHS of the sketch provided above.
2. But then, when I used the recorder "recorder Element -xml stressesEle2.out -time -ele 2 stresses" I saw that the location of the IP are 1(-.577,-.577), 2(.577,-.577), 3(.577,.577) and 4(-.577,.577). The numbering is still CCW but their location is as shown in the bottom RHS of the sketch.

Could someone tell me how the IPs are numbered, or in case I made a mistake interpreting the results, could someone point out where the mistake is.

[luky]

Hi, which shell element are you using to create the mashes? can you please post the analysis options (including eigen analysis if done) as I'm having troubles to model a slab using shellDKGQ element with orthotropic material assigned to platefibre section?
My suggestion to be sure about local reference creating the block mash is to try creating nodes manually (the model seems to be quite simple) and compare results.

[EricsonEncinaZ]

Hello Luky, thanks for your interest. I was doing element tests with shellMITC4, ShellDKGQ and ShellNLDKGQ. I'm not doing any modal analysis as I just wanted to determine the local reference system and the location and numbering of the integration points for a shell element.

The analysis options I'm using are:
integrator LoadControl 1.0
algorithm Linear
numberer RCM
constraints Plain
system ProfileSPD
analysis Static
analyze 1

[EricsonEncinaZ]

Adding to question 1 in the original post when using "recorder Element -xml forcesEle2.out -ele 2 forces", a file with 24 columns named P1 to P24 is provided. They are the forces (Fx Fy Fz Mx My Mz) in the global reference system at each node of the shell.