interpolation of strain/stress at nodes when pressure applied by surf. elements

[JoGe]

Dear all,

I would like to solve a problem where a beam which is clamped at one end experiences a pressure distribution at its top surface.

In the first case the structure is meshed with TET10 elements. A constant pressure load is applied using TRI6 surface elements. In the second scenario HEX8 and QUAD4 elements are used.

Finally I am aiming at solving larger problems in parallel as well as with multiple load steps. So I use this test case to see if the parallel FEAPv8.6 solves this problem.
In both cases the displacement field is computed whereas the strains, stresses and principal stresses are all zero!
In contrast to that these values are given when the problem is solved with serial FEAPv8.6.

[JoGe]

For further investigations I solved a very similar problem where nodal forces were applied instead of pressure. I made sure that the sum of all nodal forces equals pressure * area of the previously posted problem. The parallel solution was done with 2 load steps and the forces were applied proportionally.

For this type of load the results of the serial and parallel run are identical. So for this type of load the nodal strains/stresses are computed.

Is there any other information I may provide that helps you to find the root of the issue?

[Prof. R.L. Taylor]

The problem is caused by missing sets of 'pstyp' in ./parfeap/pform.f

At about line 474 after 'else' and before 'call passble' add:   pstyp = ie(1,ma)

At about line 486 after 'else' and before 'call passble' add:   pstyp = ie(1,ma)

[JoGe]

I could insert the missing statements at the corresponding lines such that my test cases are solved as expected.

Thank you so much, Prof. Taylor!