Applying cyclic loading in DIRECT CYCLIC step 2

[yaston4]

Hi all,

I have been through the relevent sections in the manual and understand the parameters to apply in a direct cyclic analysis using the Abaqus/Standard DIRECT CYCLIC Step, however. I am not sure how I apply my amplitude to my pressure load.

I want to apply 1000 load cycles, the time period of each cycle is 0.25s.

Here is how I have set up the step, however I am not sure on how I set the amplitude to apply to the pressure load using the amplitude module.

https://dl.dropbox.com/u/47274064/Screen1.JPG

https://dl.dropbox.com/u/47274064/Screen2.JPG

I would really appreciate any input. Thanks.

 

[IceBreakerSours]

Amplitude, as the term suggests, is just a scaling factor to the amount of load you wish to apply. So, instead of ramping up the load to the final value, if I wanted to have a triangular waveform, I could enter a table in the amplitude definition as follows:

Time | Amplitude
==================
0.0 | 0.0
_______________
0.5 | 0.5
_______________
1.0 | 0.0

So, if my ultimate load value is, say, 100 N, then at time=0.0, load=0.0*100=0.0. At time=0.5, load=0.5*100=50, ..

You can even assign a user defined loading protocol, if that's needed.

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

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

 

[yaston4]

Thanks IceBreakerSours,

I understand your explaination and understand that this is how it would work with a standard static step, however I am not sure if it is different for a DIRECT CYCLIC step in abaqus?

Thanks

 

[yaston4]

Hi all,

Still confused on how to properly utilise the DIRECT CYCLIC step option, any further input would be really helpful, thank-you.

 

[burningDiesel]

What is your problem exactly? The amplitude works the same as in a normal static calculation. The beginning and end value have to be the same obviously, but that's the only difference.

 

[yaston4]

Thanks again sdebock, my problem is that I don't understand how how the number of increments and iterations relates to the number of cycles to be defined. I know you can define the cycle time period (i.e. cycle time for one complete cycle). See screen dumps below

https://dl.dropbox.com/u/47274064/Screen1.JPG

https://dl.dropbox.com/u/47274064/Screen2.JPG

The example in the abaqus manual 1.1.21. Direct cyclic analysis of a cylinder head under cyclic thermal-mechanical loadings seems to suggest that the number of cycles are controlled by the number of incrementation's:
"In the second step of the mechanical analysis cyclic nodal temperatures generated from the previous heat transfer analysis are applied. The direct cyclic procedure with a fixed time incrementation of 0.25 and a load cycle period of 30 is specified in this step, resulting in a total number of 120 increments for one iteration. The number of terms in the Fourier series and the maximum number of iterations are 40 and 100, respectively."

This is where I am confused.

 

[burningDiesel]

increments = increments needed to solve the problem, this is a solver variable because of the (modified) Newton method needed for non-linearity.

iterations = the iterations performed to achieve the stabilized response. It can be that it is stable with less than the maximum you set here. The higher this number, the more accurate your result.

 

[yaston4]

sdebock, thanks for your post, Yes exactly and this is why I don't understand how the example in the manual relate these to the total number of cycles? Do you know how this is done?

 

[burningDiesel]

total number of 120 increments
They use a fixed time incrementation of 0.25, the period is 30, so 120 increments.
In a general case, you can just use the automatic time incrementation.

 

[yaston4]

Thank-you sdebock, my mistake, apologies and thanks for the clarification.

 

[shloktanmay]

Hi yaston4,
I am also having the same doubt regarding how to define number of cycles and I am not able to understand what sdebock told you. Please help me out on number of cycles...Thank you

 

[Amirha]

Hi Friends,
I am really stuck in applying cyclic loads to a simple 2-3D beam in Abaqus. The material is defined just as elastic. I want to apply half-sine loads. So I have defined Tabular data in Amplitude curve like this instead of using Periodic option of Amplitude (as it gives the whole sinusoidal load and not the half-sine):
0 0.00
0.05 0.31
0.1 0.59
0.15 0.81
0.2 0.95
0.25 1.00

On the other hand, I have defined my pressure load equal to 10MPa. The problems are:
1. I cannot see any displacement stresses after completing the job, seems there is a problem with reading the Tabular data because when I run the program without assigning Amplitude and just the ramp loading, it gives outcomes!
2. Can I define for example 1000 half-sine cycles in Tabular data in Amplitude curve section? or I need to go through otehr options.

Please consider my question as I really need help. Thanks Pals!

P.S.:I defined a crack with XFEM at the bottom of the beam.

 

[IceBreakerSours]

a) Read the documentation *and* consult some of the examples.
b) Check the archives or use google; questions related to cyclic loading are common.
c) Run the cyclic loading on a *very* simple problem (1D/2D) and apply a cyclic load and watch appropriate output variables.
d) Once you've gone through the two steps above, you should be in a position to make it work for your specific problem.

In general, this is the approach you should follow whenever you are trying something new (not after something isn't working).

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

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

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

 

[Amirha]

Thanks for the feedback IceBreakerSours!

I have already followed all the procedure you mentioned. I have ran many of the input files including cyclic notched beam, notched unreinforced concrete beam under 3-point bending, cracking models, etc. Also, I checked the web extensively.
The most similar one to mine is the Notched beam under cyclic loading (cyclicnotchedbeam.inp). This example is for 3 1/2 cycles ramped from zero to a certain value. When I use my own half-sine Tabular data for this example for 1000 cycles (T=1s for each cycle), I cannot see any displacement stresses, seems there is a problem with reading the Tabular data because when I run the program without assigning Amplitude and just the ramp loading, it gives outcomes! Is it really OK to apply such loading case through Tabular data?
Do you have any idea for handling such a problem (for half-sine loading) rather than using tabular data?