Bounds of physical model for dynamic analysis of structural component 1

[krt222]

Hello,

Here's the scenario:

1) A large complex dynamic structure of numerous components has been modeled in MSC Adams as a system of simplified spring-mass-dampers representing each of of the structural components

2) System level loads have been estimated via above Adams model

3) The above estimated quasi-static loads have been run as linear static load cases on fully-featured FEMs of each of the structural components

4) To check that the initial simplified spring-mass-dampers representing each of of the structural components used in Adams did not miss any important smaller scale dynamic effects in the structural components, we are seeking to use the structural component FEMs to run modal / harmonic / transient dynamic analyses

So, that said, the question is: for the structural component FEMs for these dynamic analyses, how do we determine the bounds of the analytical problem? My intuition is that we use the same component bounds that were effectively assumed for the Adams model of the full system. So for a given component, the bounds would be the interface to the other structural components on the up-load path and down-load path ends. However, others have suggested that more of the system be modeled - to me this feels unnecessary for the purposes highlighted above, but I've been wrong before!

Thanks in advance for your input.

 

[IRstuff]

Both lumped model and your preferred approach make implicit assumptions about the components, i.e., that they have no interacts other that what's implicitly assumed in the interfaces, and that the interfaces are essentially fixed and infinitely stiff. That may, or may not, be the case.

Ultimately, I think that you must run the entire system as a whole in the modal analysis. Optimized sub-components do not necessarily make an optimized whole. And, the weakess portion of a system is often the joints that hold everything together.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize