Why am I seeing fluctuation in the displacements?

[compe_ad]

Hi everyone! I am working on ABAQUS model which consists of a reinforced concrete wall and has a cantilevered W-shaped steel beam embedded (see attached picture). The steel beam is subjected to a cyclic shear loading whereas the wall has its base fixed. I am using explicit analysis.

When I look at the beam's displacement at the loading point, I am seeing fluctuation at the beginning of the loading. The kinetic energy at those cycles is also fluctuating. However, the kinetic energy is very small as compared to internal energy of the model. Please refer to attached graphs.

Can anyone suggest me what is going on and how can I eliminate this to get the consistent displacements?

Thank you in advance.

 

[FEA way]

It looks like initial transient response of a structure subjected to excitation, typical especially for quasi-static explicit dynamics. Try using the smooth step amplitude at the beginning and maybe add more steps to obtain a smoother response in all cycles.

 

[SWComposites]

Why are you not running it as a simple static analysis? Why the explicit dynamics? Is the shear load frequency that high that it can’t be considered quasi-static?

 

[compe_ad]

I was just showing 100 cycles of full loading (see highlighted section in figure below). In total it consists of 543 cycles (including ramp up and ramp down) with increasing amplitude of shear load. And I am expecting high nonlinearity and excessive material damage so that's the reason I am using explicit solver for my modeling.
img

https://res.cloudinary.com/engineering-com/image/upload/v1698020746/tips/load_tj7gom.jpg

]

 

[compe_ad]

Thanks for reply. But how can I apply smooth step for so many cycles?

 

[FEA way]

I was thinking about using the smooth step amplitude only for the initial excitation. The direct cyclic step would normally be the best for this kind of analysis unless the nonlinearities are so severe that Abaqus/Standard doesn’t converge no matter how hard you try.

 

[IceBreakerSours]

Unless I am missing something obvious, running an explicit analysis for 24 seconds appears ridiculous; you want to be the millisecond scale, not seconds. How long is the analysis running and on how many cores?

Anyway, as suggested above, you need to use a smooth function at each discontinuity in the load profile. That would translate into 5 time points in the gray box. Any abrupt change in force => abrupt change in acceleration => relatively smoother change in velocity => relatively smoother change in displacement. As an alternative to a smooth step, you could pick any nice function to map a vector from 0 through 1 and back to 0 to a smooth output and then scale your load accordingly.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083