I performed harmonic analysis (Forced response analysis) in Ansys using "Full" method. At resonant frequency i see a finite value in the response instead of an infinite value. Theoretically it should be infinite. How does ansys approximate the response at resonant frequencies (i.e from infinte to finite). Please note that I have not defied any damping.
Regards,
Kumara
Dear Claudio,
I am using APDL and applying force (in Newtons) NOT displacement (base excitation). I am using FULL method not mode superposition. I did not check amplification factor. Actually i'm trying to compare numerical results obtained using mode superposition method with ansys (APDL) results and observing difference in the magnitude by a scale of 0.2*frequency. (i.e if I multiply analytical response corresponding to one particular frequency by a factor of 0.2* frequency , it is correlating with Ansys result). One difference I would like you to note is that I'm using model superposition for obtaining numeerical method and FULL option in APDL, i don't know how could this effect the correlation. I kind of suspect there is some frequency dependent scaling happening in Ansys.
Thanks for your quick response.
Regards,
Kumara
Yes. This is more of an academic exercise. I agree that any physical system would have damping. But, here I would like to understand how does ansys deal with the case of zero damping definitely it would not give infinite response at resonance, which means it has some approximation involved. I want to know whatever the approximation used is applied at resonance is used across the spectrum or resonant freq are treated differently?
As Mr. Claudio Pedrazzi has written, It is the assumed numerical damping that brings the response from infinite to finite. In the transient dynamic analyses, by default Ansys assumes 0.5% damping ratio as the default even if no damping defined. While it is not explicitly mentioned in the harmonic analysis interface, I believe that the same 0.5% additional damping is assumed in the harmonic analysis also as the transient dynamic analysis appropriately defined for the harmonic loads and the harmonic analysis both give similar results. While in the transient dynamic analysis, this numerical damping can be set to zero, I could not find any such option in harmonic analysis. Hope this helps.
@ Mr.Claudio: I've gone through the theory manual already and could not find anything significant. I shall definitely post the progress here.
@Mr.Aamir,
Thanks for the information. I couldn't find too how to specify numerical damping in harmonic analyis (FULL) using APDL.
Dear Claudio,
Sorry for the late reply, was busy with course assignments over the weekend. Thanks for taking efforts and providing insight. At this point I agree with you that Ansys doesn't add small amount of damping.
If I may ask for more favor, can you please try implementing following example with and without damping and provide your comments.
https://www.sharcnet.ca/Software/Ansys/17.2/en-us/help/ans_vm/Hlp_V_VM86.html
I tried both the cases, for the NO damping case, results (displacement of mass) given by ansys doesn't compare with the analytical results at all. (Not at the resonant frequency even at the other frequencies too)
Analytical formula used: (Fo/K)/(1-omega^2/omegaN^2);
omega:Forcing frequency; omegaN:Natural frequency;
Does it mean Ansys results should not be considered for zero damping case at all.
Regards,
Kumara
I solved the same problem given in
https://www.sharcnet.ca/Software/Ansys/17.2/en-us/help/ans_vm/Hlp_V_VM86.html
as is still don't get the correlation between analytical result and ansys result. Analytical result is correct (0.0833), can be obtained from (Fo/K)*(1/2*zeta). But result from ansys is 0.053 as shown in the attachments.
Also, I've attached *.cdb file. Appreciate if anybody can help me the reason for the discrepancy. Thanks.
Regards,
Kumara
- Badges
- Science topic