According to "Chinese Power station steam&water pipe stress analysis technical rules"(SDGJ6-90),We must consider the force&moment acting on the equipment when "relaxation cold state case" happen.The "relaxation cold state case" as following:
After pipe has run for ages(long time),then when we want to overhaul boiler or turbien,so the pipe's temperature should change from hot to cold.After pipe has been cooled,there should be force&moment acting on the equipment,but this force&moment is not equal to initial cold state(just before first running,it's CEASAR'S SUS case).Maybe force&moment in relaxation cold state case should be more than force&moment in SUS case,but how can we combine the relaxation cold state case to meet our requirement by CEASAR2?
As far as know,the GE company and HITACHI company are all considering this case,and there are some design department in USA(as A/E,EBASCO) consiedring this case too.But we do not know if they use CEASAR2.
Thanks for your attention!

Up,please help me,Thank you in advance!

confused:

sounds like shakedown or, yielding?
if you have yielded the pipe on start up, then it won't be where it was when cold. (possibly requiring cold pull on re-connection?)

Unless high cycle (doubtfull) i can only assume the problem is in forcing the pipework back to its cold possition?

Confused as well

Hi, Twing!

Your problem is quite interesting but meanwhile complex. For me, the actual reasons for those remanent equipment loads are still unclear.

My opinion is that Caesar II cannot handle this problem since it operates within linear-elastic range of material behaviour.

When we talk about non-linearities in piping stress problems solved with Ceasar II, we mean friction effects, gaps/clearances in supports, or spring supports devices exclusively. Material yielding/plasticization or any other kind of remanent strains (as creep/relaxation problems, for instance) are not commonly taken into account.

However, I remember there is an article in an older COADE Engineering News publication, where the creep stress problem is treated. Do a search on COADE site to find this article, maybe some ideas could occur.

Regards,

Dorin.
B31.3 /319.2.3 allows yielding or shakedown.
This is 'Allowed for' in the displacement stress range. Is this not what the cold relaxation case is intending to tackle?

Thanks for above firstly.
We use CEASAR2 only a short time,so we can not calculate force&moment in this case,and we can not explain the process that pipe change from hot to cold.
Before CEASAR2,we use pipe stress program name "GLIF" that from Pipetress2010,it include this case.
Best regards.

While i think about it, yielding and shakedown is an empirical phenomenom and it would be difficult for any program to predict. but in the instance of attaching to equipment, is it not the case that nozzle loadings are set so as NOT to allow yielding of the nozzle? therefore, the cold relaxation case is largely invalid?

Maybe the chaps at Coade have a more (correct) opinion.

From twing's message it seems that he is searching for 100% relaxation condition cold load at installation temperature i.e. Ra in 319.5.1(b) of ASME B 31.3:2002 Ed.

When I was working in S&L few years earlier, we were using their legacy piping stress software where such relaxed condition load could be available after specifying the value of C, the cold spring factor.

Assuming that no cold pull was used in piping, one can easily get this load by putting highest temperature as installation temperature & installation temperature as the minimum temperature to analyse for (consider minimum temperature E temperature young's modulus available in latest ver 5.00 of caesar II). For a simple system, it can be
Ra = Rsust - C* Rexpansion where
(C = 1 for 100% relaxation).

Or, the following L4 loadcase can give the 100% relaxed cold load:

1) [OPE] W + P1 + T1 + D1
2) [SUS] W + P1
3) [SUS] W + P1 (sustained code compliance case)
4) [OPE] W + P1 - T1 -D1 (100% relaxation cond)
5) [OPE] W + P1 + (0.333 * T1 + 0.333 * D1)*Em/Ea
(T1, D1 loads if computed with Ea)
6) [EXP] L1-L2 (expansion load code compliance case)

M/c Coade or other forum members can give you simpler loadcase set up to find out the above.

regards,
sam

Doesn't (sams) case only apply if the system has
yielded or if there has been cold spring applied during instalation?

if so, 2/3c is a worst case senario for shake down

or am i barking up the wrong tree.?????

Dear Superpiper,

Relaxation occurs irrespective of cold spring.

Except for some high temperature steam turbine stability checks in 500-800 MW unit size steam turbines & high temperature steam turbines in combined cycle power plants , we never bothered about 100% relaxed load consideration.

Everything depends on the source of steam turbine - if vendor like Siemens need the data, the same has to be furnished.

regards,

sam

Sam,
Creep/Yielding/Shakedown etc are/will be evident
i suppose,if the pipework is disconnected, but i have not (and probably never will) be assesing steam turbines.

Why is this check performed?

Dear Superpiper,

No vendor of high speed rotating m/c likes steam turbine desire high load from connected piping & perform stability check with these loads. As present day steam piping work in creep range, after some thermal cycles 100% relaxation can be attained during operation in later phase of operating life, when, too, the steam turbine stability has to be maintained.

Thus, due to high operating speed - 3000/3600+ RPM & high temperature of main/reheat steam piping in creep range may be the reason for 100% relaxation load computation.

regards,

sam

I don't think you can perform this analysis with any "linear elastic" pipe stress program. The big issue here is that you can't (accurately) define the model or the loading "after the fact".

In order to analyze this condition, you would have to employ software that considers non-linear material behavior, and allows for load stepping, such that the entire history of the system is maintained in the analysis.

So the first case would be the installed case. Then the next case, starting from the end of the installed case, would be the operating case. Then the next case, starting from the end of the operating case, may be some upset condition case. (Note, if an upset occurred and wasn't recorded, then you're toast, you can't get to the end.) Now, and only now, can you run your relaxation case, based on the end of the upset condition.

Basically you have to know the entire (stress) history of the system in addition to "where" it may have yielded.

But, many proprietary softwares like PIPSYS gives various relaxation load case; are these wrong as per code?

regards,
sam

Sam,

I can't answer your question since I'm not familiar with PIPSYS. I know similar types of analysis are performed with ANSYS using the steps I outlined above.

okay some reponses...

twing said... "As far as know,the GE company and HITACHI company are all considering this case,and there are some design department in USA(as A/E,EBASCO) consiedring this case too.But we do not know if they use CEASAR2.
Thanks for your attention!"

J C Luf Repsonse...
No this cold relaxed case is not considered, I just switched firms but my former company was a rather large firm specializing in Power and we did not attempt such analysis and yes we used CAESAR II. Frankly I seriously doubt that anybody would do this read further....
*************************************
Sam said..."But, many proprietary softwares like PIPSYS gives various relaxation load case; are these wrong as per code?"

J C Luf response....
Sam I do not know about PIPSYS but I do know B31.1 and B31.3 both codes use a simplified approach to displacement stresses as Richard Ay discussed. This shake down analysis is not a code requirement of these codes. I will also add that when I get back home (I'm in Puerto Rico right now) I will check on PIPSYS....
***************************************
As for the requirements of "Chinese Power station steam&water pipe stress analysis technical rules"(SDGJ6-90)" I also am ignorant of this standard.


Now having admitted my lack of knowledge I would say that such an analysis while difficult is possible in our current day but I am not sure of its benefit.

When a piping system is initially heated up its displacements may cause certain points in its geometery to experience localized yielding the beloved Mr. Markl found that after only a few cycles of heat up and cool down the yielding would be all done with.

The largest end reactions loads will occur on the first heat up subsequent cycles will all be reduced. The self springing that will be evident at shut down will have a reaction less than the first heat up cycle and its load is applied while the turbine is off line and cooled down so its own strength is greater at that moment. The bottom line is the hot analyis end reactions along with the initial installled cold reactions are the maximum load the trubine will experience (exclusive of seismic etc.)

So thats my 2 cents I don't read Chineese (unfortuanately) so I am sure I won't be able to read the cited document but I wlll go to the local Universities Library and see what it may contain.


Hasta La Vista Baby!

Thanks to Luf Sir for taking interest in PIPSYS in this holiday season. I admire You.

It is not that I said that Sargent & Lundy's PIPSYS had 100% relaxation condition, but I verified it before writing. It is an old linear analysis tool.

In one 2001 print-out cold & weight reaction is based on 100% relaxation of piping system & is equal to weight - relax(here 1.00) * thermal load and hot & weight reaction is weight + Eh/Ec * thermal load. For many equipment nozzles of steam turbine, such 100% relaxation loads were asked few years ago for steam turbine stability check at the end of life of the plant.

In a paid consultancy we can't bother about the appropriateness of the software used so long it is having a updated validation record. So I can't say the approach was correct or not, but it was not against B 31.3 code of that time at least, I believe.

regards,

sam

Thanks above.
"Chinese Power station steam&water pipe stress analysis technical rules" is coming from ASME B31.1,but I can not find informaion about the relaxation cold state case in ASME B31.1.But according to our rules,we should compare relaxation cold state case with initial cold state case,and select the bigger result(Force&moment)to match the requirement of equipement nozzle.
So,it's not about stress,it's only about the force&moment acting on the nozzles.There are some manufacturing company need these data to check their nozzles.So maybe it's only the equipment factory's requirement,and maybe they know how to calculate these.

It sounds to me like it's weight loads plus and minus thermal loads with hot modulus used for the added thermal loads and the cold modulus used for the subtracted thermal laods.

If it's initially relaxed, the entire thermal component adds in. Then if it relaxes in its hot position, only the weight remains and the thermal strain will now subtract as the system cools.

It should bound most loads without dealing with the complexity (and reality) that Rich points out above. Local overstrain might screw things up though.

Dears,

I am reopening this discussion thread due the following reason to get the considered opinions of our forum members:

For a piperack, we have provided horizontal loads due to anchors by pipe stress calculation. Here,
the thermal expansion loads have been provided with +/- sign with 100% relaxation of piping into consideration, both along the piperack axis & the lateral direction. Is it right to do like analysts do in PIPSYS software of S&L ? If not, what will happen after piping get relaxed to the pipe rack anchors ?

regards,

sam

Loads passed to other departments (structural) are usually asummed to be without sign

Horizontal loads on racks from anchors etc, should be treated as exceptional, with the particular spar being designed to suit. not the whole rack.

Dear T.J.N,

Thanks for your reply. But, the problem with us,engineers is that we swear by ASME B 31.3 code, not by our own opinions.

While talking about sign of horizontal loads, I have not talked about treating of the two sides of an anchor in piperack in separate analyses; there, if these thermal expansion forces from two separated analyses balance, let them be.

But, I have concern about self-springing or shake down. If you read clauses 319.5, 319.2.3 & Eqn 23 of clause 319.5.1 of ASME B 31.3, you can notice that the design of anchors & restraints in piping systems can't ignore effect of self-springing. As a conservative approach, instead of C1 in Eqn 23 of clause 319.5.1 of ASME B 31.3:2002, I have considered C1=1, considering racks can be safely designed for reversible axial loads, without much problem.

regards,
sam

Sam,

Paragraph 319.5.1 refers to simple "TWO-ANCHORS PIPING SYSTEMS WITHOUT INTERMEDIATE RESTRAINTS".

In addition, as paragraph 319.5.2 stipulates, "...for multianchor piping systems and for two-anchor systems with intermediate restraints", equations (22) and (23) from par. 319.5.1 are not applicable.

In my opinion, the reaction forces and momements developed on the piping supports and connected equipment nozzles should be conservatively quantified considering the OPERATING RESULTANT VALUES (i.e. Weight + Pressure + Thermal_Expansion + Imposed_Displacements + ...) that are computed with Cold Young Modulus based on linear-elastic material behaviour. Thermal_Expansion should be considered between Installation Temperature and Thermal Flexibility (Design or Maximum Operating, as Project specification requires) Temperature. No self-springing or cold-springing simplified formulas are recommended. In fact, most of the Piping Stress Project specifications I've worked with, do not recommend cold-spring as a reliable design solution.

Regarding the initial problems that have been brought in discussion by Twing, it should be noted that the initial topic refers to long-term operating boilers and/or turbine piping systems. It may happen that these systems are defined by complicated layout and multiple movement restraints, so that I don't believe that a linear-elastic piping behaviour analysis is able to provide reliable formulas for the reactions (forces&moments) assessment. We have to take into account that B 31 Code design formulas are based on the linear-elastic analysis of the piping systems. Definitely, the results are conservative from the reactions level point of view.

Therefore, I believe that the best approach for these problems (when the actual reactions magnitude is required/desired) is that described by Richard above (i.e. non-linear finite element analysis and system loading-unloading complete history consideration).

Regards,

Dear Dorin,

Thanks for your reply. On 5th Jan's entry, I wrote about the background of the problem: it is not related to cold spring, nor I have suggested % relaxation from Paragraph 319.5.1 of B313: 2002.

It is simply related to the fact that B 31.3 piping running on pipe rack has taken some anchors from anchor bays of pipe rack. Along the pipirack axis we have some axial loads resulting from thermal expansion load case. Civil wants to know about the sign or sense of the load vector. I want to have it as +/-, considering 100% relaxation of piping within the lifetime of plant. B 31.1 & B 31.3 codes are based on self-springing or shake down & this has nothing to do with cold spring. The expression of C1 in Eqn 23 of clause 319.5.1 of ASME B 31.3:2002 gives some approximation of % relaxation; but C1 can't be greater than 1, so I had considered it as 1 for anchor design. If we analyze piping with shakedown concept for stress & don't provide anchor/restraints accordingly, I consider the same as violation to code. Please think over it now & tell whether I did it right or wrong.

regards,

sam

Sam,

So far, I haven't work with PIPSYS, but I assume that similarly to Caesar II, it is a "linear-elastic" piping stress software.

Therefore, for multianchor piping systems or two-anchor systems with intermediate restraints, where additional nonlinear effects may accur (not necessarily local yielding, but friction, gaps/clearances in guides or axial-stops), the relaxation effect cannot be quantified.

Therefore, in my opinion, it is meaningless to talk about maximum reaction forces&moments for the original/installation conditions on the linear-elastic analysis basis. In fact, under linear-elastic analysis circumstances, if no initial anchor displacements are imposed at installation moment, and considering weight, pressure and thermal expansion as cyclic loads, the final "cold" reaction forces&moments developed on restraints after a full loading-unloading cycle should be ZERO.

That's why the common practice is to communicate to the structural dept. the OPERATING REACTION/LOADS developed by the simultaneous action of the external loads (weight, pressure, thermal expansion, possible imposed displacements etc.). Meanwhile, the "Cold" situation corresponding to the weight action exclusively is taken into account.

In fact, considering the self-springing coefficient C1 = 1, and estimating Ra = C1*R, you obtained the same OPERATING LOADS too.

However, since the dead-weight action is not a cyclic load, we may evaluate the cyclic component of the reaction loads (on linear-elastic basis), considering the following loading cases (as are defined for Caesar II programme):

L1 WNC (...+H...) (SUS)
L2 W + P1 + T1 (...+ D1 ....) (OPE)
L3 L2-L1 (OPE)

The reaction forces&moments identified for L3 case represent the maximum cyclic components of the support/anchor reactions and may be useful for structural fatigue evaluations (if required).

Now, regarding the reaction sign, it's true that in many situations, the reaction forces and moments are sumbitted to structural dept. in their absolute values.

Personally, I use to attach the sign to the support reaction values forwarded to the structural dept. The reason is that I do not have practice in structural calculations and I believe that more accurate evaluations may require the loads sign if necessarily. When the reaction (force or moment) corresponds to an actual movement restraint (vertical rest restriction, axial-stop or lateral-guide), I attach the sign as it has been established by the programme (meaning the action of the pipe upon the structure). When the reaction corresponds to the friction force (such as the typical case of the horizontal friction forces induced in the simple rest points), I attach the both "+" and "-" signs, so that structural calculations are to identify the most unfavourable load combination.

We may identify the origin of the reaction watching the displacements of each supporting point.

Regarding friction, as many other Forum members pointed out before, due to the stochastic character of this phenomenon, I use to perform both the friction and non-friction static analysis for the same system and to retain the most unfavourable results.


I hope these details clarified my personal point of view related to this subject.


All the best,

Dear Richard and David ,

The post has been discussed in detail and open my thoughts as well ;
Yes, Caesar ii is a " linear-elastic" piping stress analysis program and it added "creep " case for stress checking based on EN 13480 starting from Caesar ii 2017.
Also there are a lot of power plants built in the meantime;
thus, I still want to know
is there any improvement for Caesar ii to deal with various relaxation load case (fully relaxation or partially) after the pipes operated in creep regime and considering the force and moment act on the equipment when the relaxation cold state happen ?
If not, is this a only way to know the entire stress history of system in addition to where it may occurred through FEA to obtain the more accurate reslust as Richard's saying?
Or anybody would like to share your seeing or thoughts is appreciated also.

Thanks & Regards

CAESAR II has no means of accounting for creep relaxation in the analysis. It's all F=KX.