a Non-Linier analysis (Earthquake and wind)

Hello...

I have read a discussion in this forum about Earthquake acceleration and I have a case about a non linier condition with seismic acceleration and wind loads and i have do exactly in two way:

A)First way
Case 2 = W+T1+P1 (OPE)
Case 3 = W+T2+P2 (OPE)
Case 4 = W+T3+P3 (OPE)
Case 5 = W+P1 (SUS)
Case 6 = W+P2 (SUS)
Case 7 = W+P3 (SUS)
Case 8 = W+T1+P1+U1
Case 9 = W+T1+P1+U2
Case 10 = W+T1+P1+U3
Case 17 = L2-L8 (OCC) ALGEBRA
Case 18 = L2-L9 (OCC)
Case 19 = L2-L10 (OCC)
Case 20 = L17+L18+L19 (OCC) SRSS
Case 21 = L2+L20 (OPE) ABSOLUTE
Case 22 = L5+L20 (OCC) ABSOLUTE

B)Second
Case 17 = L2-L8 (OCC) ALGEBRA
Case 18 = L2-L9 (OCC)
Case 19 = L2-L10 (OCC)
Case 20 = L5+L17 (OCC) SCALAR
Case 21 = L5+L18 (OCC) SCALAR
Case 22 = L5+L19 (OCC) SCALAR
and so on

Which way is CORRECT?,i think ABSOLUTE way give me an analysis in worst case rather than SCALAR. About Wind Case, i do not write it here cause it will be to long but i do the analysis just like the seismic, in first way(A) i try to find the MAX Wind in X and MAX WIND in Z and then add it with SUS case to get the OCCASIONAL case, in second way(B) just the same way in SCALAR manner.
Am i missing something?

Both methods will give you the same stress results. Only the forces and moments and displacements would be different, but we don't review those results for combination load cases. There is another difference in your cases however. In the first set you have SRSS'd your results prior to adding them to SUS. In the second set you did not do this, but considered the different seismic directions separately.

Ups i forget to put the SRSS in second set, but for analysis the worst case don't you think i better do with the first set?

If you had SRSS in both sets, then the OCC stresses will be the same whether you use ABS or Scalar combination methods provided you reference the same load cases.

Yes the stress are in the same value, but the force and moment are more big than SCALAR so i can use it for analysis the worst case for pump and flange and Max Restraint loads too ((OPE+SEISMIC) + (SUS+SEISMIC)).

No! You use L8 through L10 for that instead. As I said in my first post above, we don't use the forces and moments or displacements in combination load cases.

oH..so you mean sir i have to use L8 through L10 in pump analysis and others for the worst case rather than Case 21.

Thanks Mr.Loren, i understand about what you say now.

Mr. Loren there are one more thing that i want to ask to you, last night i had read COADE article/newspaper for July01, i read there for MAX Restraint Loads it is OPE+OCC(Hot load) + SUS+OCC(Cold load) in MAX way. Instead you said yesterday and i saw at CAESAR Tutorial: MAX Restraint Loads was the sum between OPE+OCC case with other OPE+OCC case in MAX way, it is so different, isn't? can you explain to me? Or one is for Linear Boundary condition and the second was for Non-linier condition?
I'm sorry my question is so basicly, i just want to develop my skill in using CAESAR II.

OPE+OCC = W+T1+P1+U1 (OPE)

The load case above contains both OPE and OCC loads. You can perform a MAX combination on all your OPE cases to see which one is your worst-case for each restraint/equipment connection in your system. But combination load cases do not apply real loads or real displacements unless you use the Algebraic Combination Method. However, for stress consideration you should not combine algebraically (the exception is EXP), but rather scalar or absolute value. If you really want to check SUS+OCC then you need it as a native load case such as W+P1+U1 (OPE). Note I call it OPE, because we don't want to perform a code stress check here. This case is only for restraint loads.

I'm sorry, it was the January 2001 COADE Newsletter/article, there you can see for MAX Restraint Loads it was given as OPE + SEISMIC (Absolute) add with SUS + SEISMIC (Absolute). But for what you say yesterday:



So you didn't mentioned about the SUS + OCC, but only the OPE+OCC where case 8 is W+T1+P1+U1, Case 9 is W+T1+P1+U2 and Case 10 is W+T1+P1+U3.
And from the Animated Tutorials it is also gives for checking Max Restraint Loads as:

Case 4= W+T1+P1+Win1 (OPE+OCC)
Case 5= W+T1+P1+Win2 (OPE+OCC)

so Max restraint loads is L4+L5 combine in MAX.

My question here which one is correct for check max restraint loads, is it OPE+OCC add with SUS+OCC in Max, or only sum the OPE+OCC case together in Max?

When collecting maximum restraint loads search all load cases that might actualy load your system. Be sure to check both +tive & -tive.

Using your first example (without a LC1), I would add the following two load cases:
LCn: L2,L3,L4,L5,L6,L7,L8,L9,L10 (SIGN MAX)
LCm: L2,L3,L4,L5,L6,L7,L8,L9,L10 (SIGN MIN)

But you probably want to add a few more cases such as W+T1+P1-U1, where your uniform (seismic) load is in the opposite direction.

Ok Mr. Dave, but which combination that we use to add LCn+LCm ? is it Scalar or Absolute ?
and then is case 20 in my first example is it right to do the combination again like that ?
One more thing if i consider the seismic loads in opposite direction (-), it mean i have to do the same thing with the wind load case, don't i?
and why in article january 2001 it was only add the Hot condition (OPE+OCC in abs) with Cold condition (SUS+OCC in abs)in MAX way to get the MAX restraint loads ? why don't you do like you said to me to ?
Thanks ...

Dylan,

There is no reason to add LCm+LCn. From all the load cases compared, one will display the largest component load on the restraint and the other show the smallest component load. I was focused on load and not stress.

If you have nonlinear conditions in your system +tive and -tive occasional loads may produce different structural responses so you might want to run both.

That January 2001 article looks like it assumes a linear system response as far as the seismic loads are concerned.

Well I get it right now Mr. Dave, Lcn and Lcm are for Restraint loads in maximum and minimum. So for non linier it is necessary to consider both (+) and (-) load.
I just read a threaded about Occasional Stress and others that discuss about Scalar and Algebraic, instead of that my question is:
In Caesar II when do we have/need to use ABSOLUTE and when do we have/need to use SCALAR to combine SUS+OCC or OPE+OCC ? since for stress they have same value, but one of my senior/mentor said that ABSOLUTE is more conservative,is it true?

Thanks for your time.

Best Regards

Tengku Syahdilan

In terms of stress - Absolute and Scalar results are identical. They add unsigned values from the combined load cases.

In terms of component forces and moment - Absolute summation adds unsigned values from the combined load cases; Scalar summation calculates the loads using the displaced position of the component and the component stiffness. (I claim that Scalar and Algebraic are the same here.)

In terms of displacements and rotations - Absolute summation adds unsigned values from the combined load cases; Scalar summation adds signed values from the combined load cases (just like Algebraic).

The relationship between Absolute and Scalar depends on the category of response (displacement, load, stress). Absolute will produce larger numbers in some categories. Some may call bigger numbers conservative, others might call them wrong.

Thank you very much Mr.Dave, your explanation has give me a light for this case. I appreciate for all your time and responses. Iti s People like you, Mr.Breen, Mr.Luff and Mr.Ay that give much more light in this field. Once again thank you very much.

Best Regards

Tengku Syahdilan

Tengku, you missed the point over and over again. You should not care, nor should you review, in fact, you should go to the load case options tab and show ONLY stress as the Output Type for your combination load cases (this include EXP as well). Displacements, Forces and Moments are NOT REAL in combination load cases!!! I don't know how to say this more clearly.

Mr. Loren i don't miss the point... I KNOW that for COMBINATION CASE we only look for the stress. My point is if we want to check Pump, Flange, and MAX restraint loads which combination is much better or that COADE recomended, is it ABSOLUTE or SCALAR? Because for stress ABS and SCALAR are given the same result, they only sum the unsigned value. But for checking LOADS (not STRESS) in Pumps, Flange or Restraint Loads, ABS and SCALAR will give different result (I want to check for the worst case in pump and flange verification where in this case is OPE+SEISMIC), and one of my mentor said ABS is more consrvative than SCALAR, so i just want to verify what he said to me...and Mr. Dave has already give me a clearly answer.
Ok. Thank you once again for your time and attention, i really appreciate that..

Best Regards

Tengku Syahdilan

To Mr. Loren and Mr.ay

I have a question: like Mr.Loren said before that Ok, for Non-linier condition i can view the Loads and Displacement for checking Pump and Flange Verification in W+T+P+U case but what about in Linier Condition when i want to check Loads (Force & Moments) and Displacement for Pump Verification? Which Case should i look for? because if i look in OPE + SEISMIC SRSS case it is not given the real Loads like Mr.Loren said doesn't it and COADE recomended not review Force, Moment and displacement for Combination Case ? Am i Missing something ?

Best Regards

Tengku Syahdilan

I am not considering nonlinear response in this message.
I imagine that if I was not concerned about any Code-defined stress evaluation, if I was only interested in loads and displacements (F&D), I probably would not have to work with any so-called 'combination cases'. I would simply create as many load cases as I needed to evaluate the F&D response.
But some of those loads that I wish to evaluate do not have a direction defined. A wind in the X direction could also be in the -X. I could evaluate this in at least two ways (remember, I'm linear):
One:
L1:W+T1+P1
L2:W
L3:WIND1
L4:W+T1+P1+WIND1
L5:W+T1+P1-WIND1
Two:
L1:W+T1+P1
L2:W
L3:WIND1
L4:L1+L3
L5:L1-L3
Your seismic load is more complicated since you are creating an amalgamation of three components through SRSS summation. Your seismic load, now, cannot be manipulated through approach One above, but approach Two works fine. L3 would be your U1+U2+U3 (SRSS) and L4 & L5 would be the operating load plus/minus this hypothetical seismic load.

Mr.Dave, my main point question actualy are:

1)why for combination case we do not look for the Force, Moments and Displacement? In TOTAL piping stress design basis there are a point that said:



it is clearly ask for the displacement check for SUS+OCC limited to 20 mm in horizontal and verical including weight effect.

2)From my previous question that you have not answer is for case 20, is it true or false if i combine the OPE again like in case 20 below:

Case 2 = W+T1+P1 (OPE)
Case 3 = W+T2+P2 (OPE)
Case 4 = W+T3+P3 (OPE)
Case 5 = W+P1 (SUS)
Case 6 = W+P2 (SUS)
Case 7 = W+P3 (SUS)
Case 8 = W+T1+P1+U1
Case 9 = W+T1+P1+U2
Case 10 = W+T1+P1+U3
Case 17 = L2-L8 (OCC) ALGEBRA
Case 18 = L2-L9 (OCC)
Case 19 = L2-L10 (OCC)
Case 20 = L17+L18+L19 (OCC) SRSS
Case 21 = L2+L20 (OPE) ABSOLUTE
Case 22 = L5+L20 (OCC) ABSOLUTE

I am not sure how many different ways we can say this: For nonlinear conditions combination case displacements, forces and moments are not real. Don't use them!!!!

If you need to limit the sustained combined with seismic displacements to less than 20 mm then I would do the following:

W+P1+U1 (OPE)

and do this for as many sus + occ cases as you have. Then for the output type use Displacement. You see? This is not a combination case, but it contains all the loads you are asked to quantify for your displacement check.

Your cases 21 and 22 do not make sense to me. Why don't you do this instead?

W+T1+P1+U1+U2+U3 (OPE)

Check anything here except stress. This will give you true displacements, forces and moments.

Dylan,

My guess is that your case 22 can be used to check that deflection criteria.

The problem is that SRSS combination of the three uniform loads. Since you must do that before adding the sustained effects, a combination is required. Loren says the structural results from combination cases for systems with nonlinears conditions are not "real"; I venture that they are "real enough" with qualification.

I trust we are all having a good day...Please I need to know if the effect of wind load is enough to overcome friction load during stress analysis,And please I would like to know the reasons for either answer...

thanks, i look forward to your reply...

1. We have signed restraint results from static load case.
2. Then we have +/- restraint result from a dynamic load case.

Now we must combination static result with dynamic result.
So we become 2 combinationen load cases.
Call in german HSmax and HSmin.
H mean OPE loads
S mean OCC loads

Correctly we must then two stress analysis for asymetric restraint structures made.

Example : Mx HSmax > Mx HSmin , but Mz HSmax < Mz HSmin
and possible Mz give a higher stress an one Point in the structure as a bigger Mx.

Or we must the max Fx,Fy,Fz, Mx,My,Mz from HSmax and HSmin take.

Herr Ohliger,

It looks as if you are mixing loads from different solutions or "states" of the piping system - for example, mixing loads from the operating+wind state and the operating-wind state - in order to develop a maximum stress level.

I do not agree. The calculated stresses must caused by a consistent load set - stress is calculated from the Fx,y,z & Mx,y,z that occur together - all from a single "state" of the piping system.

Please provide additional description if I misunderstand your point.

Mr. Diehl,
Let me answer your question with this example :

Hsmax HSmin
Fx Fy Fz Mx My Mz Fx Fy Fz Mx My Mz
10 50 30 70 30 40 20 30 90 30 20 90

Now you cant say HSmax or HSmin give the max stress in a asymetric restraint structure. Think we have weak and strong axis in the structure beams.
Also we have Mc Ml.. in the vessel branch (anchor in pipe calculation).
A lower Ml can give a higher stress as a higher Mc in the shell, beause the
branch/shell structure not symetric.
A spheric shell is symetric, here can you approximalety this load case take, that the higher resultant moment give, because the effect is here negligible.
My suggestion (max Fx,Fy,Fz, Mx,My,Mz from HSmax and HSmin take), have only the mind conservativ only made 1 structure/vessel stress calculation
instead one for HSmax and one for HSmin.

I think it best to calculate a state of stress with a set of forces and moments that work together. I am not certain that calculating stress using a group of maximum load components from several states is valid. The interaction of terms will complicate this. For example, start with a nozzle with a moment load. That moment will cause a local stress on the vessel wall at the nozzle intersection. Now consider the shear load on the nozzle flange. Depending on its direction, that shear load may add to or reduce that local stress due to bending.
If moment arms are significant, signs are important. And combining maxima ignores the significance of signs.
That said, with occasional loads it may be possible to think up additional local load combinations for your nozzle/attachment.
I would work out extreme sets of occasional loads (Fx,y,x & Mx,y,z)and keep these numbers together when calculating local stress.