Learning Forum

[BALA]

While doing Modal analysis of one assemby I came across some doubt related to Participation Factor (PF) and Ratio of Effective mass (REM now onwards).. 

1. In my modal results in one direction the REM is coming low even after extracting 60 modes , so i just want to know if any structure REM is coming low does it indicates that there is some issue or mistake in our FEM model or just to consider that these structure have very high frequency may be?
2. as number of modes extracting limit is up to we will get a ratio near to 90% but right now after 60 modes i am getting only a 20 %. I am suppose to do Random vibration on this which has frequemcy range 20 to 2000 and my 60 modes are going very high almost 8000 HZ.
3. as PF is can be used to indicate significant mode , but to decide this what value we should consider as threshold value that above or equal to that i will consider a significant mode, bcz mostly the PF values for me are in the range of 10^-2 to 10^-5. As per knowledge the mode which having PF value 0.1 can be consider as significant mode, but in my case no values are close to this?

does all these issues indicates that the my fem model is wrong? I am missing anything ?

[Dave Looman]

There are certain structures where the modes in one direction are very high.  For example, a thin flat plate will have low frequency modes in the direction perpendicular to the plate, but only very high frequencies in the plane of the plate.  Even a cantilever beam will have low bending frequencies, but very high axial frequencies.  So it's natural for models to have a low REM in one direction and that doesn't invalidate the PSD analysis, particularly if the direction of the PSD is not in the direction producing the low REM.

[BALA]

Thank you very much for your reply, Could please tell me about participation factor like what value i should consider for deciding a significant modes?

[Dave Looman]

Usually it is effective mass (participation factor squared) that is used to determine significant modes.  Since the effective masses eventually add up to the total mass you can use the ratio of effective mass to total mass as an indication of significant modes.  Another consideration is the excitation time-history.   Rapidly varying excitation will have a higher dynamic load factor for high frequency modes making them more important.   

[BALA]

Thanks. Do we have any threshold value like for this ratio of effective mass to total mass we can say this is significant mode.I am actually trying to quantify because we usually say that modes which is having higher PF or ratio effective mass to total mass is a significant but what is the criteria for considering that high value. 

Also could you please elaborate what is the excitation time-history, how does it work in Ansys?

[Dave Looman]

Each industry has their own guidelines for that.  When actually using the modes like in a response spectrum analysis, you can using "missing mass" to make up for the mass not included in the modes.  Excitation is usually ground motion, like an earthquake.

[BALA]

Thank you very much for your response. Could you please suggest standard or guideline for Ratio of effective mass to total mass. 

As i told you i heard about 0.1 as PF value for deciding significant, but in my case it is coming very less. 

Also as far as i know missing mass and residual method only work for shock analysis not for Random vibration?

[Dave Looman]

Ansys can't provide information like.  For one, we don't have it, but also, we don't know what code would apply to your analysis.  I don't recognize 0.1 PF as a criterion for significance.  It would produce an effective mass of 0.01, but for a very small structure a mass of 0.01 could be very large.  So it's the ratio of effective mass to total mass that should be considered.  Missing mass is just for response spectrum (not just shock), but residual vectors (RESVEC command) can be used in a random vibration analysis.

[BALA]

I am really thankful for your valueable answers. Could you please tell me how can I use this Residual vectors option in Random vibration in case of respone sepctrum there is direct option in Ansys Mechanical , but i didnt find it in Random vibration analysis. Do i need to use APDL command for this, if it so please guide.

[Dave Looman]

Do a string search on 'residual vector' in the mechanical help.  The APDL command is RESVEC,On

[BALA]

Thanks, I found.