Modelling of rivets in thin plate

[garymostyn]

Hi all,

What is the best way of modelling rivets in relatively thin plates in order to obtain the correct load transfer for joints such as a multi-rivet lap joint?. I have previously used various methods such as beam and celas/rigid elements and compared these with hand-calcs but the best I could achieve was about a 20% difference. Any ideas would be most appreciated.

Cheers,

Gary Mostyn
Aerospace Stress Engineer.

 

[kilbchoi]

Hi Gary,

I would like to share my experience.

<1> I alwas roughly model it as rigid element. Obviously it provides higher data than the real one due to the assumption of rigidity.
<2> If there is a too high load exceeding its allowable, I remodel it as BAR element. Normally I ended here since the more precise analysis is not required based on conservatism.
<3> In very rare case, I model it with gap with ROD element. Please be advised it's not nonlinear gap element provided by NASTRAN. If the module is avaliable to you, it's much better. In this case, you have to do take off the interest area from your global model with Grid Force Balance Data, and you completely remodel it with very very fine mesh.

Actually I have a few of references dealing with this topic as follows.
[1] Multi-Spring Representation of Fasteners for MSC/NASTRAN Modeling, Boeing -> it uses CELAS for bearing stiffness and uses BAR element for the fastener.
[2] Local Analysis of Fastener Holes using the linear gap technology of MSC/NASTRAN, BELL Helicoter

Both are available in MSC online library, I remember.

I guess we, aerospace stress guy, have to deal with both conservatism and accuacy all the time.

Good luck,

kilbchoi