Interesting Joint Stiffness Method

[norman_neo]

So I was doing some FE work (w/ NASTRAN) to determine the stiffness in all 6 DOF of a model bolted to another model. To characterize the bolted interfaces I attached surface nodes within the bolt frustums to rigid body elements (RBE2). Then, these RBE2s were connected with a directional stiffness element (CBUSH) to model the bolt. To determine the directional stiffnesses I used a method suggested by another person with much more experience.

Using Shigley's method, I found the bolt stiffness as so Kb = E*At/L'. Where At was the tensile area and L' was the effective grip length.

Then, to approximate the surrounding material stiffness, I multiplied Kb by 2. This was the longitudinal directional stiffness (2*Kb).

The transverse directional stiffness was found by multiplying the longitudinal joint stiffness by G/E ~ 0.4 to get a short beam in shear vs. tension stiffness.

This is a method that I had never heard of, or thought of. But because it was based on industry experience, I used with it. Does anybody have any comments or thoughtful input on this method?

 

[Navin A S]

It seems reasonable to assume that the stiffness of the surrounding material should be greater than that of the bolt itself, and multiplying by 2 seems to account for this. However, it is difficult to determine the exact stiffness of any material without testing it, so this method can only offer an approximation of the stiffness. Additionally, multiplying by 0.4 to approximate the transverse stiffness of the surrounding material may not be accurate in all cases, as many materials can have different shear-to-tension stiffness ratios. It may be better to conduct tests to determine the shear-to-tension stiffness ratio of the material, then use that ratio to more accurately calculate the transverse stiffness.