Alternative measures of accuracy to MPF when performing Modal Analysis

[AndreasKarl]

I am in the process of designing a railway bridge using modal analysis in LUSAS.

Throughout my short career I have been told to use the Modal Participation Factor as a measure to obtain desired accuracy. It has worked fine until now when even MPF=99.5% gives substantial errors of up to a factor 3. Other colleagues have experienced the same phenomenon but are unable to explain why?

Who started relating a high MPF to good accuracy? Are there any other parameters that could be used as a compliment to it?

Does anyone have advices of reference literature touching this topic?

Thanks,
Andreas

 

[GregLocock]

Can you describe /how/ you use MPF?

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[AndreasKarl]

Only modes with a MPF larger than a set ratio is used in the analysis.

I was under the impression that modes with a low MPF has little or no influence of the acctual behavior of the structure. So by eliminating these modes the simulation/analysis is reduced w/o losing any accuracy.

Andreas

 

[GBor]

I'm with Greg on not knowing how you use MPF, but LUSAS should indicate a convergence criteria for each mode of the modal analysis. I have seen in some programs where that convergence criteria has failed to be met in one or more of the individual mode calculations, but has still reported what appears to be a reasonable mode. When you look in the output files, thought, the convergence column reads something like "1". Convergence should probably be on the order of 0.0001.

A high MPF generally indicates a global mode. I would suspect that a global mode for a bridge means that the roadway is involved, but the accuracy should be indicated, I would think, by the convergence criteria.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group
Magnitude The Finite Element Analysis Magazine for the Engineering Community

 

[GregLocock]

OK, if I understand it then you are just eliminating modes that are just local. That seems a reasonable approach... but tells you absolutely nothing about the quality of the analysis.

So the usual checks I'd use are:

1) is the mass correct?

2) is the deflection under self weight correct, or reasonable? Does it agree with the frist vertical mode? (Do you know about that one?)

3) is the stress distribution under self weight reasonable?

4) does the stiffness in response to a static load agree with hand calcs, or similar sturctures?

5) ditto stresses

6) are the modes in the right order?

Then it gets tricky.





Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.