Hello, good afternoon everybody.
I'm working in a LNG project where seismic accelerations and seismic anchor movement should be adressed. The applicable code is ASME B31.3, and also NFPA 59A.
My question is related to the load cases and specially to the loads on restraints and equipments.
I have obtaiend U1, U2 and U3 using the criterion (W+D1+T1+P1+U1) - ((W+D1+T1+P1). Then I have combined SRSS of U1+U2+U3 (it is requested by client to consider the 3 accelerations acting simultaneously).
I have done the same for seismic anchor movement, obtaining D4+D5+D6.
Then two stress cases have been checked:
W+P1+(U1+U2+U3)<1.33SH
W+P1+(U1+U2+U3)+(D4+D5+D6)<3SA, according to ASME III.


DO you find this correct??

And my MAIN QUESTION:

For loads on restraints, I don't know which of these solutions is applicable:

Option1:
Max. load of the following load cases:
W+D1+T1+P1+U1
W+D1+T1+P1-U1
W+D1+T1+P1+U2
W+D1+T1+P1-U2
W+D1+T1+P1+U3
W+D1+T1+P1-U3
W+D1+T1+P1+D4
W+D1+T1+P1-D4
W+D1+T1+P1+D5
W+D1+T1+P1-D5
W+D1+T1+P1+D6
W+D1+T1+P1-D6

Option2:
Max. load of the following load cases:
W+D1+T1+P1+U1+D4
W+D1+T1+P1-U1-D4
W+D1+T1+P1+U1+D5
W+D1+T1+P1-U1-D5
W+D1+T1+P1+U1+D6
W+D1+T1+P1-U1-D6

NOTE: U1,U2,U3: Seismic acceleration in X,Y,Z directions
D4,D5,D6: Seismic displacements in X,Y,Z directions

Many thanks in advance! Regards

Hy
I think it a good idea to make including this combination (it is requested by client to consider the 3 accelerations acting simultaneously):

W+P1+U1+U2+U3
W+P1+U1+U2-U3
W+P1+U1-U2-U3
W+P1-U1-U2-U3
W+P1-U1+U2+U3
W+P1-U1-U2+U3
W+P1+U1-U2+U3
W+P1-U1+U2-U3

identical with displacement :

W+P1+U1+U2+U3 +D1+D2+D3
W+P1+U1+U2-U3 +D1+D2-D3
W+P1+U1-U2-U3 +D1-D2-D3
W+P1-U1-U2-U3 -D1-D2-D3
W+P1-U1+U2+U3 -D1+D2+D3
W+P1-U1-U2+U3 -D1-D2+D3
W+P1+U1-U2+U3 +D1-D2+D3
W+P1-U1+U2-U3 -D1+D2-D3

AND after that you cam make :
MAX first group : acceleration
MAX second group : acc + disp

Hi mariantrifan, thanks for your quick response!

I will consider your suggestion.

What about my main question? Do you find more accurate or "correct" to consider the seismic inertia and the seismic anchor displacemts acting simultaneously for giving loads on restraints and equipments? Or maybe this is too "conservative" and these two effect should be consider separately, as it happens with wind and seismic for example? Thanks!

But the wind and seismic is something else: the probability to have wind and seismic loads in the same time is verry low.
The probability to have seismic inertia and seismic anchor displacement in the same time is very high.

Maybe is too conservative but I thing it the best solution .

HELLO sir,

what do you mean by seismic displacements?
may i interrupt and ask?
i have used U1 ,U2 and U3 for seismic static analysis but not considered D1,D2 and D3.
Are these displacements are from wind effects or some equipment displacements?

please explane!

Hi

We are talking Seismic Anchor Movement: imposed diferential seismic displacements in equipment connections and piping supports.

B31E states that, when analysis is required, the seismic bending stress used in calculating longitudinal stress is the elastically calculated resultant moment amplitude due to seismic load, including inertia and relative anchor motion.

Hi Javgarloz,

We always consider the seismic displacements 'D' (when applicable, and not always just at the anchors either) at the same time as the seismic coeficient U.

They are inter-related, you will not get D unless you get U. So combining them in the same loadcase is not really being conservative.

Hi

The following is my idea

W+P1+(U1+U2+U3)<1.33SH(ASME 31.3)

T1+D1+(D4+D5+D6)< SA (ASME 31.3)

U1+U2+U3)+(D4+D5+D6)x2<3SA, according to ASME III.

Please review and advice me.

The first two are defensible but the devil is in the details. How do you combine U1, U2 & U3? How do you combine stress in (W+P1) and (U1+U2+U3)? Are D4, D5 & D6 ranges or amplitudes? Do you count cycles in (D4+D5+D6)?
I have no comment on the third. I don't think an owner of a B31.3 facility would care to reference a nuclear code.

Hi Dave,

Thanks once more for your helpful information but I'm quite lost...

As you you have explained, is defensible the second ecuation but...is intention of ASME code to analyze in the same ecuation such different concepts as thermal stresses and seismic displacements? ASME code specifies different allowable stress for thermal and occasional loads so, is it neccessary to "mix" them? That is what it seem to be stated according to interpretation 20.45 of the code, dated october 2005.

Could it be more "logical" to isolate the seismic problem, taking into accout both inertia and seismic displacement in the same ecuation, as it happens in the third ecuation?
On the other hand I understand that this ecuation does not "belong" to B31.3 code.

You also have talk about B31E code, where both seismic inertia and relative anchor motion are considered for longitudinal stress. I'm sorry but my company has not buyed yet this code, I'm trying to convince them. In the meanwhile please let me ask you:
Does this means that for the computed stress for seismic, which is compared with 1,33Sh, should we consider both seismic inertia and relative anchor motion effects?

I apologize for asking so much, I'm only trying to clarify the most accurate way to study stresses due to seismic anchor displacement in compliance with ASME code.

Many thanks in advance.

Thanks to all for the helpful discussions.

I have two other questions regarding the response spectrum option when using the B31E approach. I would appreciate any input on this.

In section 3.4, there is an equation that includes the terms PD/(4t) and 0.75i x (Msustained+Mseismic)/Z, etc...

1st question: Is this equation to be used at allowable stress levels? I believe I read in an earlier forum that this was the case.

2nd question: If that is the case (i.e. allowable stress), is that true for both the response spectrum approach and the equivalent static approach? Typically we have a site specific spectrum from our geotechnical engineer, but following structural engineering practice, the results from the spectrum are usually either used at ultimate strength design or they are scaled down before using working stress.

Any suggestions on how I can get a definite clarification on this?

Thanks.

Marco Italia

Javgarloz,

My understanding on this...
First you need to classify whether you are going to treat SAM as causing fatigue (secondary stress) or SAM causing primary stress failure. Then you can develop your load cases...

The codes are not very specific on the usage of SAM. The interpretation 20-45 states - "shall externally imposed displacement strains, excluding single cycle events, be considered in conjunction with thermal displacement
strains? " and the answer is Yes.
What is this "excluding single cycle events"....???
How can you count cycles during an earthquake event....and how many cycles exists...??
There is a possibility that the SAM is a single cycle and be treated in primary stress ---- occasional stress.
Hence treatment of SAM in B31E, is to use SAM along with inertia stress and check with allowables (equivalent to that from ASME Sec.III Sub.NC-upset condition). The allowable are not 1.33 Sh, but higher.
PD/4t + 0.75i(Msus + Mseismic)/Z < min(2.4Sh;1.5Sy;60ksi)

If SAM is treated as occasional, then combination with thermal is not required. i.e (T1+D1+D4+D5+D6 < SA)

ASME Sec.III codes treat SAM as causing fatigue failure and the allowable are....."3SA"
Hence do not mix codes...either use B31.3 and B31E
Or use ASME Sec.III

Mentors do advice....

Regards,
Suraj

Never take ASME Sec. III NC stress calculation, if the NC rules ( fabrication, dimension rules, testing..) not redeemed !

Mentors,

Yes, I do understand we should not consider ASME Sec.III NC rules for occasional stress evaluation of B31.3 piping...but B31E has come up with higher allowables...
I was expecting a further discussion on this subject....on the understanding of SAM...

Thanks & Regards,

Dear Mr. Diehl,

I have a question on the equation given in B31E/ 3.4 Design by Analysis and I kindly ask you to give a personal opinion.

More or less, my question is continuing the old question of Marco posted here.

In B31E equation, M_seismic is the elastically calculated resultant moment amplitude due to seismic load, including inertia and relative anchor motion.
The seismic loading input is to be specified by the engineering design in accordance with the applicable standard (such as ASCE 7) or site-specific seismic.

The question is related to the form of equation that – I think- is based on Allowable Stress Design approach.
If true, I would remark that within ASCE 7 "2.4 Combining Nominal Loads using Allowable Stress Design" there is a factor of 0.7=1/1.4 applied to E= earthquake load. If I understand well, this multiplication by a factor of 0.7 (or dividing by 1.4) is based on the fact that ASCE 7 "E" load is counted for strength level, so the practice is to correct it when count it for ASD method. So ASCE 7 apply a method correction when working with ASD.
Is applying ASME B31E a similar correction?

My doubt is coming up from other references as FEMA 451 chapter 13.4 "analysis of piping systems" which is based also on ASD, where it appears that this correction would be done by dividing Fp (The seismic design force applied in the horizontal direction) by 1.4 and considering the design relative seismic displacement as Dp/1.4Rp where Dp is the relative seismic displacement.

For sure, I have no intention to mix the requirements of FEMA 450/451 with ASME B31E.
However, I can understand the logic of FEMA but I have no idea on how B31E equation has been developed; I just would think that in B31E there is an ASD method (is it?) with a high allowable- that maybe already includes a correction.

Would you please to give your opinion on how B31E equation should be applied ?
Is the B31E intention to apply the seismic loads calculated under ASCE 7 as they are given there- and nothing more?

Thank you.

My best regards.

Timely discussion. I was just today arguing with my team about combining SUS+U+D for code compliance. My position is that the piping maximum acceleration will be roughly concurrent with peak displacement for a restrained piping system so SUS + U + D is the correct combo for code. This matches up with Itchy's post I think. Restraints would be OPER + U + D.

Just last week I finally had a loud enough argument with my structural guys to get clarification that their "ASD" design = real world force reactions. Per Mariog's last note, this would work out to 0.7 x ASCE acceleration (Fp/WP as used in ASCE 7-05 EQ 13.3-1). My structural people are still non-committal on reducing their SAM displacements which is blowing out my current piping analysis.

With regards to mitalia2009's question about ASCE spectrum - I'm taking the R value reduction noted in the CAESAR help file (~3.5 R value per ASCE tables in chapter 12) which brought the spectrum response down around the ASCE equation 13.3-1 accelerations. Would be interested to hear the Integraph guru's discuss that a bit since those chpt 12 tables are for building structures as opposed to the Rp reductions in chapter 13.

If I was going to use an LRFD/Ultimate/MCE design earthquake similar to recent Sendai magnitude, I would take the additional stress allowances in B31E 3.4(vaguely similar to Section III Class D piping in my limited understanding) to approach/allow some plastic deformation which is beyond 31.1/B31.3 occasional allowables (1.5Sy >> 1.2x or 1.33x Sa respectively).

SUS+U+D

Is this what B31.3 say?

Regards,

P.S. Maybe it is the time to wake up the comitee as nobody care what they say.

B31E FOREWORD "Seismic design of critical piping systems is often required by Building Codes or by regulation, or it may be voluntarily instituted for loss prevention and worker and public safety.

While seismic loads are mentioned in the various sections of the ASME B31 Pressure Piping Code, and allowable stresses are provided for occasional loads, there has been a need to provide more explicit and structured guidance for seismic design of new piping systems, as well as retrofit of existing systems. In order to respond to this need, this Standard was prepared by the ASME B31 Mechanical Design Technical Committee"

PS. You may address a letter to the Committee just to say them you have not such needs...

This is not what I wanted to say. But a note that address this may be (or should be) inserted in the code. Maybe they have already done in the 2010 edition, so apologize if they done.

I admit that I can be 100 % wrong but it interesting to read also Mr. Peng statement about this issue.

In his book, Mr. Peng presents a lot of sound engineering details not covered by a standard as B31.3.
It is hard to say that applying them when Contract is asking only for B31.3 means you are "right" and the rest of people may be "wrong". Eventually you take your responsibility, they take their responsibility.

Or, reading the following in B31.1 NONMANDATORY APPENDIX II, what it is your feeling about "wrong" or "right" approach?

"Two types of seismic bending moments occur. One type is due to inertia effects and the other type is due to seismic motions of pipe anchors and other attachments.
As will be shown later, the moments due to inertia effects must be considered in eq. (16), para. 104.8, in the kSh category.
Moments due to seismic motions of the attachments may be combined with thermal expansion stress and considered in eq. (17), para. 104.8 in the SA category. For this reason, it may sometimes be justified for the designer to consider the moments separately; otherwise both sets of moments would have to be included in the kSh category."

Maybe you would understand that sometimes you are allowed to be "right" otherwise you have to be "wrong"?

For sure looking in different codes you can see different approaches. For example API 650 mentions in Table E-8—Design Displacements for Piping Attachments values under the column "ASD Design Displacement", so at least it is clear they expect you'll input the values in an ASD method. They mention in text "Unless otherwise calculated, piping systems shall provide for the minimum displacements in Table E-8 at working stress levels (with the 33% increase for seismic loads) in the piping, supports and tank connection"

I'm afraid the discussion may continue with no "right" conclusion.

Returning now to B31E formula, I would think that without additional clarifications, that formula must be considered in a conservative way and that means to consider seismic loads as they are defined by ASCE 7 (including the displacements calculated for ASCE loads). Knowing no details about the formula am I wrong or right?

- B31.3 do not mention anything about the fact that displacements can be primary
- B31.1 recognize that displacements are secondary, but say that if you have dificulties in segregate this you have to include them with the rest of primary (occasional)- is their choice so i will not comment.

I can not comment the B31.E as it is not mandatory as per B31.3 and I had not studied yet in depth but seems to refer to structural integrity which is reasonable.

In any case we are not mixing codes.

In any case, yes, the code is not a cook book but seems to be for many a Witch book, which is not good. In this case I will be not the first but I will say that is WRONG that is obscure.

Sorry for my late reply...

ASCE 7 addresses the *0.7 or /1.4 issue in at least two places - Setions 12 & 13 - generally, apply these factors when using Allowable Stress Design (ASD). As far as I know - if pulling data from ASCE 7 to apply to pipe stress analysis, use 70% of teh ASCE 7 load.

That will give you the appropriate moment for stress.

But...

Not the right structural response for displacement and equipment load.

There is an update to B31E. Paragraph 3.1 now defines the a-p and R-p values for (welded steel) pipe. These values will give a larger displacement and load while also increasing the moments for stress. But these increased stresses are also used with the larger allowables defined for piping in B31E. So, if using B31E allowable stresses, you must use the B31E parameters.

I believe these disconnects (strength design vs stress design and stress vs. structural response) are teh reason why B31.3 does not reference B31E directly.

It's great to know there is now B31Ea giving more info about applying ASCE approach.

My special thanks for your answer and my best regards.