I am analyzing a jacketed line that needs to withstand seismic loads. Along with the "normal" cases, I am running two seismic cases, one for a uniform load of GX=0.33/GY=0.2 and the other GZ=0.33/GY=0.2. Additionally, since the building the piping is in will rack, I am pre-displacing support points by the amount of the building shift (upper levels move relative to lower levels by about 4"). While the pipe passes code compliance for the normal sustained and expansion load cases, the predisplacements are causing the analysis to fail code compliance for the occasional (seismic) load cases. The project has decided that during the seismic event the pipe can leak or rupture but cannot "fall down", and the routing is pretty much fixed for process reasons (it's sensitive to the developed length). Any idea as to how I reconcile the code with this directive?

I may have additional information to consider. You do not mention to what code you must comply. I am working with a static seismic analysis that is to IBC 2003, ASCE 7 2002. In "NEHRP Recommended Provisions: Design Examples", Aug 2006, load combinations are discussed, and the allowable stress approach of B31 and the strength design approach of Provisions are discussed. The resultant moment (Mb)is to be induced by the design force Fp/1.4, where Fp is as defined by the Provisions (which is the same as the ASCE 7 force). And the resultant moment (Mc)is to be induced by the design relative displacement Dp/1.4Rp, where Dp is as defined by the Provisions (and the same as ASCE 7 drift).

I have not found the explaination for this Dp/1.4Rp term. And I am working to understand this, but regret I have no more to offer at this time.

The Provisions are available as a free download from the FEMA website, if you need to obtain a copy.

I am also interested in Anindya's thread about equipment loading requirements as pertains to operability, leak tightness, and position retention under IBC & ASCE 7, what are your project's requirements?

From what I understand from your question, you have failure in Seismic anchor Movement ( SAM) and not by inertial loading. In fact studies have shown that most seismic failures take place due to SAM rather than due to Inertial loading.B31.3 addresses the interial loading as primary ( hence 1.33Sh requirement) and the SAM stresses as secondary ( EXPANSION STRESS RANGE)

If you are working on B31.3 rules and you find that in the EXPANSION STRESS RANGE you have the failure, look for provisions in SEC III Subsection NB, NC and ND. Also you need to know if it is OBE level seismic loading or SSE. For SSE level loading, you have provision in Appendix F also.

There are some requrements in SEC III where SAM range has to be considered. In that case your results will mostly show problem in stress. Regarding this scenario, you have to keep the two things in mind: No. of cycles that the system can see and what is the factor of safety based on Markl's work for lower no. of cycles. Typically for lower no. of cycles the F.O.S is between 5-6. Now , coming to no. of cyles, a conservative estimate can be taken as 100 cyles( basis; 20 seconds strong earthquake at 5Hz)per earthquake.Definitely for 100 cycles the F.O.S is very high and it is of no justification to use the range for SAM.

Regards

Thanks very much for your replies. The applicable code is B31.3. Based on Anindya's comment above, I believe I have my load cases defined incorrectly. I had been considering the predisplacements in the load cases that result in the calculations for occasional loads, meaning I was considering it as a load similar to weight, pressure, etc. I should be considering it similar to an expansion load, much as you would include a displacement at a turbine or boiler nozzle? My current (incorrect?) load case definitions (D1 are the displacements due to story drift applied to the support points; U1,2 is the G force in x and z directions respectively) are shown below. Any suggestions on what they should be? Should I create a separate one W+T1+P1+D1 and then subtract W+P1 to get the expansion stress?

1 W+T1+P1 OPE
2 W+T1+P1+U1 OPE
3 W+D1+T1+P1+U2 OPE
4 W+P1 SUS
5 L2-L1 OCC
6 L3-L1 OCC
7 L4+L5 OCC
8 L4+L6 OCC
9 L1-L4 EXP

I will not write the load cases as I want you to think on it independently, Besides CAESAR II manuals provide lot of tip on it. I will however give you some suggestions:

1)For non linear systems extract the pure occasional loads like U1, U2, U3 ( where by these terms I am referring to seismic inertial loadings in X, Y and Z directions respectively) use cases like (W+P+T+U)-(W+P+T)

2)Use load cases to involve both MAX. and MIN. temp. to get these U's.

3) Find Max U's in the three orthogonal directions.

4) Combine them by SRSS.( You can use 100/100/100 ie giving equal weightage to all three directions or you can use them as "factored"). There is an interesting discussion on this subject in the book titled 3Dimensional Static and Dynamic analysis of structures by E.L.Wilson.

5) Combine the Seismic value just obtained with OPERATING ( this involves min., max. actual operating etc) as well as SUSTAINED ( for SUS+OCC stress check). This will give you the Max. restraint loadings. Use+ and - sign when combinations to simulate the dynamic behavior of the load.

6) Use the same with Displacements also.Only thing: In case of displacements do not check the SUS+OCC ( displacements).

7) The above will change if you are using SEC III rules .

Hope this helps.

It does - -thanks A.S.