Help with Post-Processing Abaqus CAE

[VN1981]

Hi,
I have the following analysis setup.

I have 3 load cases defined (LC1, LC2, LC3). Further, I am performing a two step analysis. The first step is a Frequency (Natural Frequency Extraction) followed by a 2nd step which is a Modal Steady-State Dynamics.

I have a set of interested frequencies defined for the 2nd step. I am still learning on how to set up the analysis properly.

Anyways, for post-processing, I am wondering if the following is possible in Abaqus CAE

1. Organize results by loadcase
2. Further, in addition to the above, it would be better if I could further categorize the results based on Step.

Currently, the outputs are bunched together. It is tedious to go through each result individually. Also, any pointers on how to generate an envelope plot of Peak Stress values per Load case?

Thx in advance.

P.S: Any pointers or references in Abaqus documentation to get me started would work as well.

 

[Mustaine3]

I assume you mean that you have 3 different SSDs after the frequency extraction.

If the settings for the SSD are equal and just the loading is different, then you can use one SSD step and 3 Loadcases within. For large analysis that would be faster than 3 steps. But it makes no difference for postprocessing. See Analysis Users Guide 6.1.4 Multiple load case analysis

There is an evelope plot in postprocessing. See A/CAE Users Guide 42.7.5 Creating field output by operating on frames
It creates a new temporary step with a new frame that contains the result of the operation.

But be careful when working with Mises stress or other invariants when having complex results.

 

[VN1981]

Hello,
Thanks for the response. I think Abaqus treats multiple load cases just like multiple steps if I am correct. Anyways, I am attaching a couple of images to make my model setup clear.

If I could sort outputs based on Step as well as Load Cases for my situation, it would be great.

I will check out USers Guide 42.7.5

>> But be careful when working with Mises stress or other invariants when having complex results

Any particular reason you say the above?

 

[Mustaine3]

1. You can also use 3 SSD steps after the frequency step if you like and ignore the loadcase functionality It's just slower in the solver, but that's all.

2. Your definition of the frequency ranges looks strange. How to calculate 20 solution in a range that is zero (or one?)?

3. This text can be found in the Abaqus support page: "The Mises stress is a nonlinear invariant of the stress tensor, and this implies that this measure is non-negative and not sinusoidal, in contrast to the stress components sij. The components of a complex stress tensor can be out of phase, and in that case, these invariants make little sense, because the phase information will be lost during their calculation. Note that this applies to other complex tensors such as the strain tensor."

 

[VN1981]

Hi Mustaine3,
Thanks again for replying. I have a few follow up questions.

1. If I create 3 SSD steps, would they be independent of each other? My concern is that what if Abaqus thinks that load associated with Load Case 2 acts on the structure after the effects of load associated with Load Case 1. That is, Abaqus applies load 1 to the structure and it deforms a certain way. What if Abaqus applies Load 2 to the deformed structure rather than to the original configuration? I just want clarity. Is there any option on the Steps set up editor where I can specify the above from not happening.

2. Regarding the freq range definition, the input file was supplied to us which was generated using Nastran. Abaqus automatically generated the definition. I manually modified the entries later (not shown in the image I posted).

3. This is how results cases appear in MSC Patran post-processing module.

And this is the format I get in Abaqus.

Somehow, I feel that Patran format gives me better continuity (since I can just scroll through all results for one particular load case frequencies sequentially before I proceed to the next Load Case). I was hoping that Abaqus would let me arrange (using filter or any other means) freq modes for each load case...but using Steps would alleviate I guess.

4. "The Mises stress is a nonlinear invariant of the stress tensor, and this implies that this measure is non-negative and not sinusoidal, in contrast to the stress components sij. The components of a complex stress tensor can be out of phase, and in that case, these invariants make little sense, because the phase information will be lost during their calculation. Note that this applies to other complex tensors such as the strain tensor."

As of now, I don't understand the above in its entirety. If you have the time, I would appreciate if you could explain it better in a simpler way else I will eventually figure it out.

Anyways, thanks man for your responses. Appreciate it a lot.

 

[Mustaine3]

1. The 3 SSD steps would not interact which each other and also not continue things. Each linear step acts independent. Only nonlinear steps with NLGEOM change the state of the model for the next step.

2. Have a look at the 3 result of one frequency. I assume they are equal. When you define only one frequency per line, then you can set the number of points and the bias to one. There is nothing else to calculate. They only matter when you have an actual frequency range per line.

3. With 3 steps and only one result frame per frequency you might have something similar like in Patran.

4. In a SSD the load varies harmonically (cosinus function) and so your result vary per frequency. Because of damping (and maybe a complex load) the results are usually complex. They vary by phase angle and the real and imaginary parts and the magnitude changes with that. In /CAE you can display these different results and also animate them. See references below. And because of these variation it makes little sense to calculate a Mises stress with the components of one phase angle.

See Abaqus/CAE User's Guide
42.6.3 Understanding complex results
42.5.8 Selecting complex results
42.6.9 Controlling the form of complex results
49.1.3 Harmonic animation

 

[VN1981]

Hi Mustaine3,
I just wanted to thank you for your responses. I now have a better understanding on performing SSD using Abaqus.