errors in doubly embedded beam and portal

04-Unexpected behaviour/errors
Post Reply
bover992
Posts: 3
Joined: 02 Feb 2018, 17:47

errors in doubly embedded beam and portal

Post by bover992 »

hello,
I would have to present some problems: I noticed that even on simple elements, such as a doubly embedded beam subjected to uniformly distributed vertical load, the program provides very different results changing the number of frames in which the beam element is divided (with only one element does not even provide the correct stress and strain outputs in the integration sections). It also provides a normal stress value different from zero, despite the type of load, this means that the breaking load of the beam is much higher than expected.
Finally, from the study of a portal with the concrete "con_ma" there were problems related to the lack of equilibrium at the nodes and also in this case a considerable difference between the results provided using only one frame or multiple for the beam.
in all cases I have modeled the elements as "inelastic frames based on forces" for which it should not be necessary to divide the elements into several parts.
thanks for your time!
User avatar
seismosoft
Posts: 1184
Joined: 06 Jul 2007, 04:55

Re: errors in doubly embedded beam and portal

Post by seismosoft »

First what do you mean that the program "with only one element does not even provide the correct stress and strain outputs in the integration sections"? If you are using the force-based element type the results should be correct.
We cannot also understand what you mean with "It also provides a normal stress value different from zero, despite the type of load, this means that the breaking load of the beam is much higher than expected". If you have restraints in the direction of the beam, you should expect axial loads.
Regarding the lack of convergence, we note that (1) the con_ma is the most stable concrete model of SeismoStruct, and SeismoStruct is one of the most stable programs worldwide. Can it be that you have convergence problems, due to very large loading, and you have lack of accuracy due to loose convergence criteria that you applied in order to make the model converge?
Note that in general you need no more than 5-6 integration sections per structural member. Hence, if you have a beam divided into 4 elements of 5 integration sections each (i.e. a total of 4*5=20 sections) you will probably have convergence problems.
Seismosoft Support
Post Reply

Return to “04-Unexpected behaviour/errors”