Learning Forum

[Mikkel Dreyer]

I am currently doing a transient response, and view the Timehistory strain in Ansys APDL based on a shell181 element. 

The shape is a hollow cylinder, and it is fixed in one end. When i take the transient z-direction strain i notice that the value does not correspond to the difference between the z-displacement difference divided by the length between the nodes (see formula in picture #1 + see transient strain response in picture #2). 

However, when a similar analysis is done with the same forces in a static analysis, then the strain correspond to the upper and lower z-displacement difference divided by the length. It seems like the transient response time history has a scale difference on 3.4, but it still has the same z-displacements, when the static state in picture #2 is reached.  What causes the strain the strain formula to be correct for the static case and not the transient case? Maybe i use the wrong transient settings?

The transient response uses these settings in APDL:
/SOL        ! Enter the solution module
ANTYPE, 4    ! Analysis type: Transient
TRNOPT,FULL ! Solution method
LUMPM,0     ! Lumped mass approximation

OUTRES,ERASE
OUTRES,NSOL,LAST

DELTIM,1        ! Time step size
TIME,StartTime    ! Start time
MIDTOL,0,0,1    ! Include response frequency
TINTP,0.005     ! Amplitude decay option, 0.5% damping
LSWRITE,1,  ! Write 'LoadStep' file 1

[dlooman]

It seems odd that the graph of strain in post26 is in such a narrow range.  Are you saying the values are all off by a factor of 3.4 or the difference, (3.703e-8 - 3.695e-8), is off by 3.4?  There could be a round off error (some times called subtraction of elephants) in calculating strain by the difference in the uz values if they are both large compared to the difference.