Bending Moment for Sizing Fuselage Frame

[DRDearth]

I posted this same Question on the FEA Forum. I'll repeat it here to see if any Bites.

I've run acoss a number of engineers questioning how to use results from FEA models to Size Fuselage frames for bending moment. I've always thought one typically takes Shear loading in the frame between longerons (q lbs/inch) X distance between longerons (inch)==> this gives total shear force (V lbs). Then multiply this total shear X frame depth (dx)to give total Moment (M in-lbs) i.e. M = V x dx. But when shear loads come out of a FEA model that idealizes the frame with outer & inner caps as Rods and the frame web as a membrane (or shear panel), Then the question arises how one relates these Rod loads (lbs) & panel shear (lbs/in) to estimate total moment in the Frame.

What would be nice is a sample problem that illustrates how one can use the FEA results of Rod axial loads and panel shear loads to compute bending moment in the Frame. Any comments out there?

David R. Dearth, P.E.
Applied Analysis & Technology

 

[gwolf]

Don't build a model like that or you will end up with the sort of difficult questions which you face. You either model the whole structure as shear panels for the skins and rods for the stiffeners, then do the hand calcs, or you model it semi-literally using 100% shells and base your frame assessment on shell stresses rather than rod loads.

 

[rb1957]

on the other hand, build your model like that to get those results !

i've modelled frames as "caps and web", the frame moment is just the moment created by the cap loads ... you don't need to know the moment as the model as resolved this into cap loads, which you can analyze directly. i'd size the cap with an effective area of web (30t, h/6), and model the web as a shear panel. You can always reverse-engineer the moment out of the cap loads if you need a less conservative result.

this way of modelling (caps and webs) has a bigger matrix to solve (than using beam elements for the frames) but i don't think that's much of a problem these days. i guess this style came about for consideration of a built-up frame, with a discrete inner cap rivetted onto the frame web panel, as opposed to a one-piece machined frame; tho' i thnk both approaches are just as valid for both types of structures.

my problem with modelling the frame as a beam is that you need to be careful if you've modelled the stringers as beams. you need to releasethe bending freedoms that are not restrained in the structure.

there're many ways to skin a cat, the trouble is to find one that likes it (or will sit still long enough) !

good luck

 

[rb1957]

oh, btw, remember that the inner cap of the frame has to be modelled as a beam (with nominal properties) to constrain the up-of-plane deflections

 

[DRDearth]

Thanks for the commments.. the approach I was outlining is from the "old school" of FEM models. The "rod & membrane" approach was about all that could be done back then. Today's contemporay FEA programs allow for more realistic approachs. But the questions I posted still come up from time to time...

David R. Dearth, P.E.
Applied Analysis & Technology

 

[DRDearth]

Also as an aside comment... a ROM estimate for the moment in the frame can be estimated from the differences in cap loads (Rods Axial Loads)X the height of the frame. M = (delta rod load lbs) X d (Depth of frame). This moment is balanced by the shear in the panel between the rods (q lbs/in) as M = 2qA. A = panel area... if my I remember my arithmetic correctly.

David R. Dearth, P.E.
Applied Analysis & Technology