How to get buckling factor for variable load?

[netsonicyxf]

I have 2 load cases, G for permenet load, and Q for variable load.
Both load cases act on the structure at the same time.

I want to do a linear buckling analysis to get the buckling load = 1.0*G + buckling factor * Q,

How to achieve this in Sap2000?

 

[271828]

I would try computing the buckling load and then subtracting the permanent load. That would leave the part of the buckling load that comes from the variable load.

 

[netsonicyxf]

Thank you for the suggestion.

AFAIK, the "subtracting" method works for simple structure and simple load pattern only.
E.g. Both G and Q are point loads and act on top of the column.

But for complex structure or complex load patterns, I don't think the "subtracting" method works.

 

[netsonicyxf]

Linear Buckling Analysis – Constant + Buckling Factor * Variable Case
These are descriptions of the feature in Strand 7 R3 and Midas Gen / Civil

Strand 7 R3
The linear buckling solver calculates factors applied to a load case to cause the structure to buckle. The load case is referred as the Variable Case in R3 and the buckling factors scale the loads in that load case.
R3 also offers the option to select a Constant Case together with the variable case. If a constant case is selected, that result case is assumed to be applied to the structure with a constant load factor of 1.0. The linear buckling solver then calculates the load factor that scales only the variable case to cause the structure to buckle while the load factor on the constant case remains as 1.0.

Midas Gen / Civil
Buckling Combination
Enter the load cases for the buckling analysis. Different Load Cases with the corresponding Scale Factors may be entered.

Load Type

Variable: Consider increase or decrease in loads (live loads, etc.)

Constant: Do not consider increase or decrease in loads (self-weight, dead loads)

 

[JoshPlumSE]

I don't think a linear buckling analysis will work the way you want it to. The CSI Analysis Reference Manual has the following to say on Linear Buckling Analysis:

That suggests that the way the program is solving for it is dependent on the current load vector. And, I don't see an obvious option for doing the type of Base Load + Variable Load * Buckling factor type of analysis.

You could certainly create a series of load cases where you use Base Load + 1.1*Variable, Base Load + 1.2*Variable Load ..... until you start converging on your buckle factor.

That's not a totally clean way to do it, but if you're stuck then it seems a realistic way to move forward.

Alternatively, you could try a nonlinear buckling analysis with large displacements to see if that gives you some more loading options. Or, a time history analysis where you Ramp up the variable load with time.