Dears,

In one specific location, ASME B 31.3 piping is designed for 150 miles/hr wind speed with Exp Catagory D (wind pr loading 43-73 psf from 0 ft to 330 ft height) as per AISC-7. For UBC Zone-3 (ZPA=0.3g, static accn. 0.21 g),the lateral loading of the piping for this piping specification, even with water-filled condition, is not more than 43 psf. Global effect of mass of piping on structures is being considered in structural seismic analysis.

I heard that in earlier days, people from east coast were designing piping for wind & west for seismic.

Can we avoid seismic analysis, in this situation, considering piping to be strong enough against seismic loading in the axial direction of pipe, as wind analysis does not load piping axially ?


regards,

sam

I think that you shouldn’t be designing for more than one contingency at a time. In addition, I think that when seismic movements are happening the wind tends to cease… but I am not 100% sure about this statement.

Hello Sam,

The B31 Piping Codes do not require considering seismic events concurrently with wind. Look at B31.3, Paragraph 301.5 for dynamic loadings.

Just my opinion, so you should always look at the specific Code and Paragraph.

Regards, John

Dears,

You have misunderstood my query. I have nowhere written that I was contemplating seismic loads concurrently with wind.

My query was - for a high wind area of wind speed 150 miles per hour, after designing piping for such wind loading, whether separate static seismic analysis resulting in much lesser lateral horizontal load will add much value to the piping analysis.

This may be needed, so far as code confomance is concerned. But, I want to know about the experience of our seniors. This is the reason that I wrote that,"I heard that in earlier days, people from east coast were designing piping for wind & west for seismic".

I repeat, nowhere we contemplate wind & seismic events to occur simultaneousely in piping design.

regards,

sam

It depends... assuming one is in a rather active seismic zone lets say South California and have a basic wind speed of 150MPH.... I could see where large heavily insulated vapor filled line would be more influenced by wind then seismic...

Conversely the same line instead if it was liquid filled would be more influenced seismically especially if its high in a building...

.........and, in cases like this where the loadings are not additive, the analyst can design the systems to the loading combination that best "envelopes" the highest creditable operating loading. Occasional loads are a separate consideration with their own higher maximum allowable stress limit.

Maybe a more interesting discussion might evolve around the question: "are wind loads always occasional loadings?". Designing for the 50 year or 100 year "wind" (hurricane, tornado, or whatever) as an "occasional" load (with higher maximum allowable stress) seems to be OK but what about at the Coastal Aruba Refinery where there is a CONSTANT 25 to 45 mile per hour wind from East to West (did I remember to say all the time, every day). Is this a loading that is additive for normal design combinations (e.g., sustained weight plus sustained internal pressure plus sustained wind - with the stress limit being 1.0 x Sh)? Good thing that Aruba has not seen a hurricane direct hit in 200 years (but some ocean swells have hit as the huricane goes by to the North).

And from that point of view, what if the location was in a seismically active zone? Then we would have Sustained loadings (W+P+Wind) additive to the seismic event and compared to the higher maximum allowable stress limit (perhaps 1.33 x Sh). The Code may not provide guidance for every design challenge but the ENGINEER is still responsible (to the owner) to think about all this fun stuff.

Regards, John.

Thanks to both of you, John C. Luf & John Breen for your opinions with insights, which help us to learn from you. Now comes the problem of load combination in Caesar-II for static seismic analysis with U1 (0.21g in X) U2(0.21g in Z)
& U3(0.14 g in Y, Vertical).

I want to combine, taking cue from an earlier entry by Kalpesh in this discussion forum, as follows:

L1->W+D1+T1+P1 OPE
L2->W+P1 SUS --------->CODE COMPLIANCE
L3->W+D1+T1+P1+U1 OPE
L4->W+D1+T1+P1-U1 OPE
L5->W+D1+T1+P1+U2 OPE
L6->W+D1+T1+P1-U2 OPE
L7->W+D1+T1+P1+U3 OPE
L8->W+D1+T1+P1-U3 OPE
L9->L1-L2 EXP -------->CODE COMPLIANCE
L10->L3-L1 OCC(ALGEBRIC)
L11->L4-L1 OCC(-"-)
L12->L5-L1 OCC(-"-)
L13->L6-L1 OCC(-"-)
L14->L7-L1 OCC(-"-)
L15->L8-L1 OCC(-"-)
L16->L10+L12+L14 OCC(SRSS)
L17->L11+L13+L15 OCC(SRSS)
L18->L2+L16 OCC(SUS+OCC)-->CODE COMPLIANCE
L19->L2+L17 OCC(SUS+OCC)-->CODE COMPLIANCE
L20->L1+L16 OPE(OPE+OCC)
L21->L1+L17 OPE(OPE+OCC)
L22->L20,L21 OPE (MAX)

Will this do ?

Now, if some one wants to use with friction coeff 0.3, the following seismic loading,

+/- 0.21g X, +/- 0.14g Y(Vert), +/- 0.063g Z
+/- 0.063g X, +/- 0.14g Y(Vert.), +/- 0.21g Z
+/- 0.15g X, +/- 0.14g Y(Vert.), +/- 0.15g Z,

what will be the load combination ?

regards,

sam

You can model friction into the model and then in the load case editor you can turn it "off" by multiplying by a zero. Doing this lets you run both friction and non-friction with the same model see the Users Manual 6-21 to 6-23.

Load cases would be the same or similar to the above setup. Also read up with the online help on the "MAX" command and make sure its what you wanted.

You seem to have the right idea as far as case combinations is concerned. I assume algebraics and scalars are correct in the above

Thanks to you, Sir, John C. Luf. I have corrected the load case combinations, where I had erred due to straight lifting from other's work online and keeping it incomplete, although I had given credit for it. But, still I accept the error, as we, the Engineers remain responsible to the Owner anyway, as John Breen had pointed.

Dr. A.P.J. Abdul Kalam, President of India, the Missile Scientist says that, "A candle loses nothing by lighting another candle.This noble act will promote creativity and a socially conscious society". Due to the people like you, people at Coade, John C. Luf & John Breen, USA is Great & remains as the Leader in the World, as you spend your precious time helping others invisibly, nurturing, rewarding & if required, by scolding the engineers younger than you. Wish You All Excellent Health & A Great Future Ahead.

regards,

sam

-------------------------------------------------

the corrected Load cases:

L1->W+D1+T1+P1 OPE
L2->W+P1 SUS --------->CODE COMPLIANCE
L3->W+D1+T1+P1+U1 OPE
L4->W+D1+T1+P1-U1 OPE
L5->W+D1+T1+P1+U2 OPE
L6->W+D1+T1+P1-U2 OPE
L7->W+D1+T1+P1+U3 OPE
L8->W+D1+T1+P1-U3 OPE
L9->L1-L2 EXP -------->CODE COMPLIANCE (ALGEBRIC)
L10->L3-L1 OCC(-"-)
L11->L4-L1 OCC(-"-)
L12->L5-L1 OCC(-"-)
L13->L6-L1 OCC(-"-)
L14->L7-L1 OCC(-"-)
L15->L8-L1 OCC(-"-)
L16->L10+L12+L14 OCC(SRSS)
L17->L11+L13+L15 OCC(SRSS)
L18->L2+L16 OCC(SUS+OCC)-->CODE COMPLIANCE(SCALAR)
L19->L2+L17 OCC(SUS+OCC)-->CODE COMPLIANCE(-"-)
L20->L1+L16 OPE(OPE+OCC)(ABSOLUTE)
L21->L1+L17 OPE(OPE+OCC)(-"-)
L22->L2+L16 OPE(OPE+OCC)(-"-)
L23->L2+L17 OPE(OPE+OCC)(-"-)
L24->L20,L21,L22,L23 OPE (MAX)

Dear Mr.John C. Luf,
I quote your last post
"You can model friction into the model and then in the load case editor you can turn it 'off' by multiplying by a zero. Doing this lets you run both friction and non-friction with the same model...."

Analsis of a model with friction and non-friction gives different stresses & support loads. We have to ensure the model is safe for both the case and supports are to be designed for highest load among the two cases. Please correct me if I am wrong.
Regards,

You are correct. One should look upon friction in the following ways...

1)There is no single constant for static Mu tests show that it is a value which varies tremendously.

2)Static Mu quickly becomes a sliding Mu once the line breaks free. Sliding Mu values are significantly lower than static Mu.

3)Friction should be viewed and used in a model only as negative influence in other words friction should be looked upon as either increasing stresses or reaction loads. It should never be used to decrease stresses or reaction loads.

Years ago Mu was not something which analysis programs could deal with. Work was done successfully without it in the model by realizing that its effects were there, and designing accordingly for it. By the same token persons of less competence had designs that were adversely affected by Mu because they simply overlooked it!

As such when friction is significant its effects can be easily reviewed in the same model. Computers, piping codes, and other tools do not negate the value of an experienced competent designer executing the work.

The list of 24 load cases appears exhaustive here but I am afraid that people will assume that this is the final word on the matter. I have a few comments...

LC3-8: This approach implies the model is nonlinear, that a positive load is different from the negative load. If the job is linear, then these +'s & -'s are unnecessary as the seismic components could be analyzed separately, without the operating loads included. Also, if the nonlinear nature of the system supports consists only of "small" gaps, then these gaps can be removed from the analysis making the system linear. There are a few words to this effect in the referenced document below. One more point... assuming the systemn is nonlinear, then analysing Operating plus U1, U2, U3 separately will not necessarily produce the proper "state" of the system that is hit with some combination of U1, U2, & U3 together (as in operating+U1+U2). I am not saying more analysis is required, I'm just saying that what you see here is not everything.

LC10-15: This is a good way to extract the occasional stresses from the various "states" of the piping system under operating and seismic loads. Again, this approach assumes nonlinear supports in the system.

LC16-17: Another approach (rather than SRSS of all three U's) is the 100/40/40 (or 100/30/30)method. Here you would use 100% of each component with 40% of the other two and then pick the worst case. Other than that, I am not sure if you can support the claim to isolate the +U's from the -U's this way.

LC20-23: These look like the load cases to determine the loads on supports and overall deflections. But why the Absoulte summation? Cases L1 & L2 are signed and the others are not. I would suggest adding four more cases: L1-L16, L1-L17, L2-L16 & L2-L17. All of these should then be Algebraic summation.

LC24: If you agree with what I just said then this list of load cases should also include the new ones too. Is this for the maximum restraint loads and systems deflections? If so, then I suggest the summation method be SignMax rather than Max. I then would include another case with the same load list but this time using SignMin "summation".

Sections 4.4.9 and 4.4.10 of "Guidelines for Seismic Evaluation and Design of Petrochemical Facilities" (published by ASCE) cover the combination of directions of seismic load. For horizontal loads, the SRSS method and the 100%/30% & 30%/100% method are reviewed. The vertical component may or may not be 2/3 the horizontal before it, too, is summed with the horizontals.

Another source is George Antaki's document with American Lifelines Alliance. See Seismic Design and Retrofit of Piping Systems . It includes an early draft of B31S - the general piping document for seismic analysis.

Evaluation Guide for the Seismic Operability of Active Mechanical Equipment

Many lifeline system mechanical components perform critical functions during and after earthquakes (e.g., a pump that must start to provide fire suppression water or a valve that must close to isolate a toxic or flammable spill). The earthquake performance of these components is assessed by various entities using a combination of analysis, shake-table testing, operational experience, and informed judgment.

To provide a single information source, the ALA developed a guide that encompasses the available performance data for six classes of mechanical components: valves, valve operators, pumps, compressors, fans, and packaged air handling units. For each component class, the guide identifies seismic failure modes and the primary contributors to each failure mode. Checklists are included to facilitate the evaluation of components for new and existing applications in both commercial and industrial facilities.

http://www.americanlifelinesalliance.org/pdf/October-ALA-Eqt-SD_110404.pdf

My initial query was not regarding lifeline pipings.

My query was - for a high wind area of wind speed 150 miles per hour, after designing piping for such wind loading, whether separate static seismic analysis resulting in much lesser lateral horizontal load will add much value to the piping analysis. Restraints provided for wind loading can be adequate enough for seismic loading.

Mr. John C. Luf's first post on 1st June addressed the problem qualitatively in the following lines:

"I could see where large heavily insulated vapor filled line would be more influenced by wind than seismic...

Conversely the same line instead if it was liquid filled would be more influenced seismically especially if it's high in a building...."

But, practical benefits can result from quantification of the above idea. Then only it will add value to the bottom line of the firms we serve.

Just like a stress critical line list makes the number of lines in a plant reduced for proper attention of stress analysts, can we derive a rule-based exclusion list where wind loading can make seismic analysis redundant ? There must be a safe & smart way out there for our search!


regards,

sam
---------------------------------------------
Seismic analysis is excluded generally for the following:

1) gas piping of diameter lesser than 1 inch ID.
2) all other piping except containing hazardous fluids less than 2.5 inch ID.