I'm currently re-analysing changes to a system that has employed cold spring to reduce loads on a vessel nozzle and a guide. This system has a 2" gap on the guide (introducing non-linearity). The cold spring has been modeled using the modified alpha tolerance method because of springs in the system. The load cases have been set up according to the COADE technical reference manual as follows:
(note T1=max temp, T3=min temp, T4=cold spring
Load case 1: W+P1+F1+T1+T4(OPE)
Load case 2: W+P1+F1+T3+T4(OPE)
Load case 3: W+T4+P1+F1(OPE)
Load case 4: W+P1+F1(SUS)

To verify code compliance for the expansion stress range between T1(max T) & T3(min T) is the following load case valid (i.o.w. does it effectively eliminate the effects of cold spring to comply with B31.3):?
Load case 5: L1-L2(EXP)

Due to the non-linearity in the system the code stresses in this case are not the same as if a seperate run was made on the same system with no cold spring modeled. In the latter case (run with no cold spring) the code stresses are significantly different, and in some cases over allowables. In reality, the gap on the guide will exist (although the magnitude of it may be questionable depending on the accuracy of implementing the cold spring).

Would someone like to comment on the correct approach (i.e. is the seperate run necessary to comply with the code)?. I'm aware of the inherent risks of cold springing, but eliminating it in this case is not an option.

For the expansion stress range, yes, your load case "L1-L2" is valid. It does "effectively" eliminate the effects of cold spring.

As soon as you have non-linear boundary conditions (directional restraints, friction, gaps, or large rotation rods), you really can't make load case comparisons. This is because the behavior of the system is dependent on what loads are in the load case. This is why

1) W+T1+P1
2) W+P1
3) L1-L2 (EXP)

is not the same as

1) W+T1+P1
2) W+P1
3) T1 (EXP)

Even though the first set of load cases effectively removes W and P1 from the expansion case, the results are different from a T1 only case due to the non-linear effects of the boundary conditions.

Richard,

Regarding this response example to an old inquiry, When it is said "the results are different from a T1 only case due to the non-linear effects of the boundary conditions"

Is this a result of addional non-linear effects of W & P1 left in the L1-L2 case? In the T1 case, Are only non-linear effects of T1 are applied?

I was wondering if non-linear items are present in the model, are they always applied no matter what load case you use?

thanks!

The only thing (possibly) "non-linear" in CAESAR II are the boundary conditions. Non-linear boundary conditions would be:
- directional restraints (such as +Y)
- restraints with gaps
- restraints with friction
- large rotation rod restraints.


If your system has non-linear boundary conditions, then in all probability an "L1-L2" load case will not yield the same results as a "T1" only load case, where:

L1 = W+P1+T1 (OPE)
L2 = W+P1 (SUS)

This means caesar only could solve linear problems,yes?

No. By setting up your load cases as

1)W+P1+T1 (OPE)
2)W+P1 (SUS)

3) L1 - L2 (EXP)

the variation in the stiffness matrix due to the non-linear boundary conditions is accounted for.

Hi,

Good Day!

I will share my little understanding to this topic... PLEASE correct me if Im mislead...

It goes like this... In evaluating the expansion case, I have to consider the effect of non-linear condition from the operating and sustain condition... thou the process would cancel the value of W,P1 but the individual effect in the non-linear after running the condition still declare it's non-linearity..
unlike L3=T1(exp) evaluating this will not include the effect of the restrain, gaps, friction and etc. that make it a non-linear because the force, moment and other reaction of things that can make it non-linear cannot be considered.

PLEASE SAY SOMETHING!

Pretty much.

Read these other posts on this subject:
http://www.coade.com/ubbthreads/ubbthrea...=true#Post10211

http://www.coade.com/ubbthreads/ubbthrea...h=true#Post4820

http://www.coade.com/ubbthreads/ubbthrea...h=true#Post2766

http://www.coade.com/ubbthreads/ubbthreads.php?ubb=showflat&Number=2789&page=0&fpart=1