Dynamic shock analysis results

[MW_9654911]

Dear Creo user,

i have question regarding interpretation of my results i got from a Dynamic Shock analysis in Creo 4.0 m120. I was also getting sth similar in Creo 5. Can anybody tell me from where such "corrugated" results may be coming? I am using here shell elements for each component. The modal analysis shows some modes and everything is ok (photo, right side). When i run the shock analysis with a defined by me spectral acceleration, modal combination SRSS, the result is as in the left part of the photo. 

so far my observations are as follow:

- when i use the default Uniform spectral values the result is "smooth"

- when i run this assembly with its parent everything is ok

- tried once to run with the supports having solid elements and the results were ok.

I don't think it is a matter of shell/solids. It looks like the flat areas would be getting some extra waves as mode for the area itself, but it doesnt make sense as the modal analysis doesnt show such modes. 

Any hint will be appreciated highly, thx

BR,

Marek

[JXBWk]

image 1. disp =  3675mm !! It this from you dynamic analysis right? so are these peak values? or at a given time

what is your "spectral acceleration" ?

the right picture is an eigenvector?

applying the correct results file? May not be applicable to software like simulate but with a proper FEM tool if you attach the wrong result file something you get that pattern (node ID "conflict)

[MW_9654911]

my spectral acceleration is max 1,5g. it is not much. (file Spectral.png attached). THe units i have are mmKgSec and i was checking already if the value i type for spectral acceleration is 1000 bigger. FIrst of all, when i have a full assembly, not only this barrell i am getting proper results with this input. i played with units and spectral acceleration 1000smaller and the results are also nonsense. Small displacement looking relatively ok but the stresses are at the level of Pascals, not MPa. Extremely small - that is impossible.

The right photo shows displacement for all modal nodes combined. 

The results file is taken directly from the analysis. No mistake here possible from my side. 

[JXBWk]

this is not a shock profile! some sort of sine sweep I guess

What is the lowest eigenvalue (1st natural frequency) you are getting for your assy?

If one uses mm then usually one has the mass in tonne to be unit consistent

What does the "all modes combined" suppose to show?

[MW_9654911]

i am interested in reaction for a seismic loading at a power transformer. I followed Creo help suggesting this simulation for seismicity. as an input i have a acceleration response spectrum in m/s2 vs Hz as in the attachment. 

here the natural frequencies -  rigid mode search is not ticked

Mode  Frequency (Hz)
     ----  --------------
        1   1.703986e+00
        2   3.132876e+00
        3   8.275412e+00
        4   1.177347e+01
        5   3.299463e+01
        6   3.552031e+01
        7   3.575230e+01
        8   6.346721e+01

What does the "all modes combined" suppose to show? - in the results for modal analysis one can pick which modal shapes to show. it is also possible to tick all of them.

You mention that is not a shock analysis. For simulating seismic activity with the inputs i showed, are you suggesting to use a different type of analysis  ?

[JXBWk]

How about figuring out what is the main mode (1st one !), decide the damping level (your curve gives g-level for 2%,5% & 10% damping) and from that get the g-level? Once you have that multiply the stress level for 1g by this factor. Need to make sure that the 1g direction is appropriately selected i.e. it corresponds to the "direction" of the main mode. If 1st mode is a vertical bouncing for example don't use a lateral 1g condition. Just a thought!

[MW_9654911]

Hello JXBWk,

the damping level is decided and for my case it is 5%. As a results, combining the damping + input accelerations i get the table which you called "semi sine function". That is an outcome from certain standards that led me to the final "g" i should use.

My seismic accelerations can appear all at once so i take the 1g for all axes. in fact vertical are 1/2 of horizontal but that is another story. I could calculate each direction separately with only one of the axis and combine results using SRSS but that Creo can do for me. 

I am not sure if direction of 1g will make any difference. Modal shape is related to the body not to the accelerations. So if proper frequency appears any acceleration should excite the body to its form from the modal shape.

At the same time, i am not really interested in the 1st one only, as for seismicity the mass participation factor should be 90%. Here in the example, i reach it with 4-5th mode. For bigger structures i did it could be between 100-150th natural frequency for those modes the damaging effect can be bigger than for the first ones when only small piece is shaking. 

anyway, going to the primary topic, i still cannot find a reason why the results look like in the fist post photo. the way how i get to the final accelerations is another story. the questions is still, why the result is so corrupted/corugated ?