OP2 reading: values at integration points

[Ryan Watkins]

In general, quantities that are integrated during an analysis (such as stress and strain) are calculated at quadrature points. If you want nodal values, these quadrature values are then projected accordingly. Unfortunately, this can lead to "misleading" data depending on the coarseness of you mesh. Is there a way to read in the quadrature values? In my dealing with the OP2 reader thus far, it appears that only nodal (projected) values are read in.

[Steven Doyle]

PyNastran's OP2 reader reads the data exactly as written to the OP2 (with a few minor exceptions).  You can think of the OP2 file as being identical to the F06.  The only real differences are that it's binary (so it's faster and has higher precision) and there are a few special tables in the F06/OP2 that are unique to those files.  Other than that they are the same.

One thing to understand is the difference between a CQUAD4 type 33 and a CQUAD4 type 144.  The CQUAD4-33 outputs the results at the center of the element, while the CQUAD4-144 element outputs the results at the center and 4 corner points of the element.  Note that internal to Nastran, these two elements are the same.  They have the same degrees of freedom and the same quadrature points.  The CQUADR introduces the drilling DOF and is fundamentally a different element.  It is not possible to recover the quadrature point values for a CQUAD4-33.  However, it should be possible to recover that data for a CQUAD4-144, though I think it's not possible for other elements like CQUADR or CQUAD8.

If you had a structured planar quad mesh and wanted to get the exact shape function nodal averaged stress per Patran, you'd blindly add up the 4 corner point stresses and divide by number of elements touching that node.  If you also had CTRIA3 elements, the centroidal value is the same as the nodal value.  A oxx, oyy, txy "answer" will be the same with average/derive vs. derive/average, while von mises and the principal stresses will be different.  The derive/average method is typically what's done, but that must be calculated after the fact.  It is not stored in the OP2/F06.

> Is there a way to read in the quadrature values? 

No because the F06/OP2 does not contain the data.  You must calculate it (if it's even possible with just the stress/strain) though I don't think you actually want those values.  The critical stress for a CQUAD4-144 will always be at one of the corners.  If that number is too high, you probably want an averaged stress.

On Tue, Feb 16, 2016 at 12:57 PM, Ryan Watkins <watk...@gmail.com> wrote:

In general, quantities that are integrated during an analysis (such as stress and strain) are calculated at quadrature points. If you want nodal values, these quadrature values are then projected accordingly. Unfortunately, this can lead to "misleading" data depending on the coarseness of you mesh. Is there a way to read in the quadrature values? In my dealing with the OP2 reader thus far, it appears that only nodal (projected) values are read in.

--
You received this message because you are subscribed to the Google Groups "pyNastran Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pynastran-disc...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ryan Watkins]

Makes sense. As always, thanks for the prompt reply.

[Paul Blelloch]

I don't believe that nastran reports results at the integration points.

On Feb 16, 2016 12:57 PM, "Ryan Watkins" <watk...@gmail.com> wrote:

In general, quantities that are integrated during an analysis (such as stress and strain) are calculated at quadrature points. If you want nodal values, these quadrature values are then projected accordingly. Unfortunately, this can lead to "misleading" data depending on the coarseness of you mesh. Is there a way to read in the quadrature values? In my dealing with the OP2 reader thus far, it appears that only nodal (projected) values are read in.

--