Learning Forum

[jbettinger]

Hello.

When performing modal analyses, I have generally tried to make sure that I have enough modes in the analysis to get an effective mass ratio of ~0.9 in each of the important directions of the downstream analyses. However, I generally have to guess how many modes that is. What I've found myself doing is running a modal analysis seeking hundreds of modes, sometimes up to 500, then I will go through the Ratio of Effective Mass to Total Mass table and figure out how many modes I need to hit ~0.9 in each direction. Then I can re-run the modal analysis with that found number.

I was wondering if there was an easier way to do this? Like, can I tell Ansys Mechanical to just keep solving for modes until it reaches the mass ratio, or perhaps run something to predict the number of modes needed?

Any assistance in this regard would be greatly appreciated. Thank you.

[Deepak]

Hello,

Ansys Mechanical does not have a direct feature to automatically continue solving for modes until a specified mass ratio is reached, there are some strategies you can employ to streamline the process.

One approach is to initially run a modal analysis with a reasonably high number of modes, as you have been doing. After reviewing the Ratio of Effective Mass to Total Mass table, you can identify the number of modes that give you the desired mass coverage. This initial analysis can serve as a guideline for future analyses on similar structures, potentially reducing the number of modes you need to start with each time.

Thanks,

Deepak

[dlooman]

Generally, capturing a high percentage of the total mass is only important when doing a response spectrum analysis.  If that's the context, one thing to consider is using "missing mass" and not worrying about the ratio of effective mass to total mass.  What's the ultimate purpose of the modal?  As Deepak mentioned, there's isn't a capability in Ansys to extract enough modes to reach a certain percentage of the total mass.

[jbettinger]

So for us, the goal of the model and others like it is to subject the unit to shock and vibration. The vibration is almost always random vibe (every so often harmonic/sine vibe but not as frequent), and the shock usually splits the middle between being a time domain (transient structural "sawtooth" shock) or a frequency domain (response spectrum). I guess it makes sense not to need too many modes for rando vibe as long as I have enough to cover the frequency range and a bit above. But still, since the modal analysis feeds into the shock and vibe I figured it would be important to make sure as much of the system is described as reasonably achievable. The projects I work with vary pretty wildly in shape and size. My most previous project needed ~150 modes before I hit a 0.9 ratio, while my current one (which is larger and goofier in shape/mounting) only needs ~35 modes.

[dlooman]

OK.  Does missing mass help?