PDELTA ANALYSIS 2

[IJR]

Could any fellow structural designer share with me his/her experience on how to replace buckling length factor K by direct application of P-Delta analysis. Please consider a 3 dimensional steel frame model to analyze using for both vertical and lateral loads and utilizing a finite element analysis software.

Thanks in advance.

 

I don't think you can replace the buckling length factor, K, by application of P-delta analysis. They are two separate issues. Column buckling is a first order effect that depends on the unbraced length. P-delta is a second order effect that depends on the displaced shape of the column under loading. Of course, the value of the K factor will be affected by whether the frame is braced or will sway. Maybe I'm missing something. Can you can clarify your question?

 

[cv]

I think the scope of IJR ( Visitor ) is to forget the using of K buckling length factor once he is using a FEA with P-Delta and P-delta analisys. This subject concerns me too because after my point of view, when we are using a FEA with P-Delta and P-delta, we are working on the real behaviour of the model. It would be better if we load the model not in one step but in several steps of load ( NL analisys ) because in each running the geometric matrix would be updated.

Also I would like to have someone to show a point of view about this subject.

CV
carlosvalinhas@netcabo.pt

 

[Qshake]

Most commercially available analysis software allows either a first order analysis or a second order analysis and a non-linear analysis mode. To account for P-Delta, the second order analysis is sufficent for most designs. In fact, in some software the switch for second order analysis is "PDELTA" rather than "ANALYSIS". Beware that some software a tolerance is given while in others a number of iterations is given. Make sure that you've used the proper number of iterations to achieve the tolerance necessary.

In summary, use PDelta or use moment magnification not both. See the response above.

 

[rakbaba]

ijr

eurocode 3 allows k=1 if second order analysis is made,but you have to take some fictitous lateral loading into account.also check ssrc guide on this subject.

 

[Hariharan]

Conceptually, k-factor is a design related feature and p-delta effect is an analysis related feature. The two are not directly comparable. The k-factor correlates the
'effective buckling length' to the physical length, or, in other words, is a factor for determination of axial buckling strength of the member. The p-delta analysis, on the other hand determines the additional stresses (primarily bending) induced on account of deformations in the presence of high axial stresses. If you are performing a design or verifying a member adequacy, you cannot ignore k-factor.

If you are verifying the member adequacy based on an analysis which includes p-delta
effect as well as other nonlinear effects (tangent/secant stiffness - stability functions, large deformation effects etc, you may use the interaction formula

(fa/Fa) + (fb/Fb) less than or equal to 1.0.

When you do a conventional linear analysis you would be using the interaction equations as per AISC

(fa/0.6Fy) + (fb/Fb) less than or equal to 1.0 and

(fa/Fa) + (Cm. fb / (1 -fa/Fe') / Fb) less than or equal to 1.0.

The commentary on these equations in AISC (I remember the 8th Ed) describes the basis. The moment magnifying factor (1-fa/Fe') was the easier way of incorporating the p-delta effect approximately.

Hope this helps.

M. Hariharan