Dear All,

I just want to see your opinion about this subject, i have followed many discussion about this topic, in Eng-Tips.com, Cheresources.com, and offcourse COADE CAESAR Forum, and still i am agree wit Mariog on the thread of "PSV Reaction Force" when he said:



This is also inline with LC Peng opinion on his book "Pipe Stress Engineering" about this subject.

But recently i have seen a method that quite conservative, and with this method i think when we are doing detailed engineering with static analysis, we only need to apply the reaction force on vertical and horizontal direction in PSV Body (as seen in figure below) as long as the good support/restraint applied in the system (also as seen in figure below), with stopper in approriate location after discherge valve, it will minimize the effect of reaction load downstream.

I want to hear opinion about this, please..

Regards

D

I think that before anything else the question to be answer is:

Does "Pop up Condition" exist or not?

Regards,

At page 12 of the pdf there are some things about this:

http://www.parcol.com/docs/3-I_gb.pdf

Yes danb,
I'm also quiet often hear about that question (does "Pop Up Condition" exist or not)and have read also this reference from PArcol,that is why there is a stopper and rest support (with 10 mm gap) that intended to avoid transmitted force and heavy stress to PSV Body, and also downstream to the header (Large Area) where jet force ussualy will applied.
About Pop up condition, yes it happened, altough only in a few millisecond, that is why in App.2 non mandatory ASME B31.3 mentioned in Para.II-2.3.2, paraghrap two:


Ussualy the most accurate for Closed Discharge System is to do Time history analysis (as also stated in Para.II-3.5.2), but this will be time consuming.
Quasi static analysis as proposed by E.C. Goodling and also LC Peng, to calculate each load in each length leg as:

F= F1*L/c*t0
F1 = 129 W/1000 * SQRT(k.T/(k-1)M)....(KN)
L= Length of each leg
c or v = is fluid velocity or near sonic velocity (get this from process dept.)
t0= time valve opening (get from vendor or assume 4ms)

this force applied by multiplying it with DLF to each elbow or obstruction.
But the question with this method will be: until how far this load should be applied to be near to accurate or conservative result?
This will depend the judgement and experience of each engineer.

So because of the complexity of that method,we ussualy will provide good support and restraint system based on experiences that has been working for many years. With propered support, you will only need to apply the load as figure above unless more complex system and stringent system need another approaches.

By the way thanks for the comment danb.

Regards
D

Please note also some other things.

At the moment of "pop up", the line downstream PSV is yet cold.

Support upstream PSV should be a hold down support.

In some cases an u-bolt may not be appropiate.

In most of cases the elevation of the two trunnions will not be the same.

Regards,

You are absolutely right danb...thanks

See also Pipemill. Interesting is that Pipemill apply also a DLF to the vertical force while others don't.

And this is from EN13480-3, Appendix A

"A.2.4.1 General
The discharge of a safety valve will produce a reaction load on the piping to which it is connected. The initial rapid opening of the valve produces a dynamic component to the force which can be significant.
The effect should be treated as a localised event producing point loading at the nozzle connecting the valve to the piping, and should be incorporated into the design of the piping and the supporting arrangements."

Regards,

Just note that "F2" Pipemill formula considers vertical component as W^2RT/(pA)instead the "correct" form W^2RT*/(p*A). I've explained in other post where is the "fluid mechanics" error and without this correction "F2" has no theoretical support.

I would add that "correct" means here the API "free jet" reaction force formula, and I do not interpret this as an evidence there is a vertical force which must be calculated with that formula.
Of course, this is my opinion, perfectly in line with the title "your opinion on this sticky subject"...

Best regards.

Mario, I think you are refer to Gas in closed system calculation (PSV) , because for Liquid Closed system (PRV) it using different formula due to different nature of gas and liquid.

Related to this i still found some Engineering company using the formula of Gas in closed system when they are calculated Liquid in closed system. Offcourse the reult will become so vey conservative.

As an example, i have checked a pipe stress calculation for a liquid in closed system.
The system is 6 " Pipe Sch.40, with PRV 2"x3" 150#
Design Pressure upstream PRV = 16.4 Bar
Design Pressure Downstream PRV = 10 Bar
Setting Pressure = 17.40 Bara
Relief Temperature = 214 C = 487.15 K
Max. Disch. Rate/Valve = 4752 Kg/h (1.32 Kg/s)
Calculated or Required Orrifice Area = 506 mm^2
Actual or Selected Orrifice Area - 563 mm^2 (Type H)

Using Pipemill:
F1 = 50.72 N
so that F4 = 101.43 N

Using Excell SPreadsheet (that used by Engineering Company, because of they apply Gas in closed system formula:

F1 = 0.7 KN
So that the force in verical and horizontal of PRV:
F2 = 0.4 KN
F3 = 1.3 KN

As you can see it is too conservative and lead to costly design and stiff system.

Maybe they intrepret the line in Para II-2.3.2 of ASME B31.1 Non Mandatory App. II, (that mentioned :"The large steady state force will act only at the point of discharge, and the magnitude of this force may be determined as described for open discharge systems") without considering the different between gas and liquid.

CAESAR II also give a seperate calculation for gas and liquid as you can found in Dynamic input for Relief Load analysis. There are Relief Load Syntesis for Gas and Liquid.

This is also one reason i open again this discussion, because there are still many missconception about this topic.
Many International world Scale Project using that calculation for gas to be applied to liquid.

And for Liquid Closed system, i prefer not to input the force in CAESAR II input, unless it is significant. Just use proper support as above in my first post (with Hold down guide on upstream as danb said).

regards
dee

Well, Dee... Maybe is a PSV for tube rupture case. Even if it is on liquid line, it need to be sized for gas.

Regards,

Dear Dee,

I failed to understand the discussion is on liquid PSVs.

Of course my comments were focused on "API formula" which is one for gas applications.
Yes, that gas formula gives larger forces than for liquid because considers the "critical" gas speed.
Under normal circumstances there is no need to mix API formula for liquid application. However, for flashing liquid cases...


I would add something....In fact also in your scan paper appears the pop-up force of W^2RT/(ps*A) with ps- gas set pressure. What is interesting with this formula (active now all around the world) is the fact the set pressure is placed in denominator and that means the smaller is the set pressure, the larger is the pop-up force.
I would imagine that the author liked Parmenides' (and Aristotle's) famous dictum "horror vacui"...indeed "Nature abhors a vacuum" was a driving principle in science and (maybe) we would claim such pop-up force formula is the true "reaction" force measure in terms the PSV disk detests the low set pressure...

About the liquid PSV formula: it is good that there is no "horror vacui", it's just a moderate crash with "Bernoulli"...
But wait, Bernoulli for transient "pop-up" is the same as Bernoulli for "steady-state"? Well, here is another dilemma... which is better, "horror" or "vacuum"?

And just to finish my discussion on "opinions".... returning Dan's question "Does "Pop up Condition" exist or not?" it seems that he found the answer. I did it, too: "pop-up condition" exists, except cases when PSV is not opening!

About the magnitude of the pop-up force....It is said Galileo had accepted the weightlessness of air as a simple truth ... so, humble in thoughts, I must question myself who am I to understand such simple truth as the magnitude of pop-up force in PSVs?

My best regards.

Dear Dee,

In the scan paper you’ve attached (that one valid for compressible fluid in PSV closed systems), there is a note valid for "established conditions": "Only F2 is effective, as other forces are balanced".

I would like to discuss this conclusion; the goal is just to invite you all to think about rather than to say this approach is wrong.

It seems that they placed (in each piping elbow) two forces estimated by API formula – which is based on reaction force of "free jet" having critical speed. In addition they placed at PSV outlet the same force.
As a result it appears that forces are balanced, except vertical in PSV.
Why vertical force is unbalanced? I think the logic is we cannot assume the same "free jet" force upstream the PSV.

For "pop-up" condition (the other figure in the paper scanned) I guess the PSV is considered as an "enlarged bend" and probably the logic is that forces are imbalanced because the flow is initially established in this "bend" and not in the rest of piping system. As a result it appears that only two forces are in system and they are acting in PSV- one horizontal and one vertical.

I have few remarks based on some concepts of fluid mechanics.

- In a steady state flow through a regular piping system having only elbows, fluid exerts forces acting in elbows and these forces are "balanced" giving some tension in pipes. Under normal circumstances the dynamic components are negligible and these forces are pressure forces, these forces are balanced and give the pressure longitudinal stress (commonly considered as PD/(4t)).
If the dynamic components cannot be neglected, Fluid Mechanics (by "Momentum equation") gives us the expression for forces acting in 90 degree elbows.
At a first approach, we can consider that the fluid has the same pressure and same velocity for all elbows, so all forces are balanced in such system.
At a second (more detailed) analysis, one would observe that these forces are not identical because it is a loss of pressure in system and some changes in velocity, but in this case the "momentum equation" must be completed with supplementary friction and/or inertial forces, so again the forces appear as balanced.
The conclusion of this fact: we don’t need the "free jet" reaction force assumption in each elbow in order to get the balance of forces in such system. The forces are balanced anyway, because the physical reality; however, for compressible fluids, it would be difficult to accurately calculate "by hand" their magnitude.

- A good question: it is the "90 degree bend" fluid model an appropriate model for flow through PSV? My answer is NO, because we are not able to properly calculate the action of fluid in a PSV!

Here I need to explain what "momentum equation" is. It is the second Netwon’s law. Why we use it? Mathematically, one advantage of using "Momentum equation" for flow action in bends (and not only there) is the simplicity of method; we need only to draw a control surface such that it crosses uniform flow zones where the velocity and pressure are essentially constant. In back, it can be proved that that momentum equation gives us indirectly a surface integral of pressure, ie. the fluid force on that surface.
Without "momentum equation" tool, we need to carry out the integration p*dA (pressure over the surface), but in practice this would be a "tragedy".

The common form of "Momentum equation" neglects the friction. This is not so evident in some fluid mechanics books; in others they said "friction is neglected" or "shear is a topic for later".
But this is evident if we remember that "Momentum equation" is the second Netwon’s law and we have experience with this law applied in Mechanics; there are cases in which friction can be neglected, there are cases where friction drives the problem!


In my understanding, in a PSV, fluid friction cannot be neglected because PSV is constructed to destroy fluid energy and momentum by friction.

Can we evaluate what is imbalanced in a steady state system with PSV? Yes, after separating the effects of fluid shear forces, we need to carry out the integration p*dA fluid inside the PSV!.
It is not a matter to be solved by a stress engineer.
You have to consider results of CFD models in order to understand something.

- About considering the vertical imbalanced force in PSV…. Reducing to a stupid simplicity what it is inside the conventional PSV: there is a nozzle, a jet of fluid strikes a plate (the disc) and a spring provides resistance to plate movement.
There is the temptation to be focused on the momentum transmitted to plate and to the fact spring exerts up a force to PSV bonnet.
However -remaining into the same stupid simplicity...do not forget that there is a jet that exits the nozzle and there is also a "reaction force" transmitted to "nozzle". PSV nozzle and PSV bonnet are mechanically assembled together, so the "imbalance" should be proved before to be axiomatic declared!

My opinion; only for pop-up condition we can talk about an imbalanced force applied to disc and the consequence is the disc movement. See API 520 for a good description of "pop-up"- I attach a sketch with three figures from API.

But is there an imbalanced vertical force transmitted to PSV body? In my understanding, the answer is rather no...
Of course the right answer must consider that- inside the PSV- things are much more complicated than I described and- again- the CFD models can give the answer.

The last question: are CFD models reporting imbalanced forces?
In my knowledge, no, but I would like to say "yes" if you can give me a serious reference.
I prefer to quote the "abstract" of one reference I have:
"If the safety valve outlet piping is connected to a blowdown system or, in case of blowing off into the atmosphere, are equipped with a T-piece at the outlet, the stationary reaction forces are compensated completely. The transient opening process, however, develops flow reaction forces which culminate in peaks of short duration. […] Special importance is attributed to the short duration of the effect of the reaction forces which seems to have only a negligible impact on the supporting steel structure."

Best regards.

Hi Dee,

Providing line line-stop really helpfull to reduce stress due to PSV pop up. My recent project however client not allowed to have restrained close to the PSV (right before PSV outlet). Hence, in this case we should figure out case by case basis in which part the stress occured.

Force in vertical from the inlet of PSV normally can be negligible and only PSV reaction force from outlet of PSV that is having significant impact for calculating PSV case. DLF for coservative result can consider 2 times.

just my two cents
-wan-

Dee's earlier post

"Quasi static analysis as proposed by E.C. Goodling and also LC Peng, to calculate each load in each length leg as:

F= F1*L/c*t0
F1 = 129 W/1000 * SQRT(k.T/(k-1)M)....(KN)
L= Length of each leg
c or v = is fluid velocity or near sonic velocity (get this from process dept.)
t0= time valve opening (get from vendor or assume 4ms)

this force applied by multiplying it with DLF to each elbow or obstruction.
But the question with this method will be: until how far this load should be applied to be near to accurate or conservative result?
This will depend the judgement and experience of each engineer."
is Justified!
Then, why are we not using it, still, on other legs of piping?

reg,
sam

Dan's approach of L/(c*t0) forL less than c*t0 and 1 for L equal and greater than c*to is inline with steam hammer load calculation in old MEN issue and referred in CAESAR II time history webinar.

But, how many of us put psv thrust load in psv inlet and outlet legs and design enough restraints! Knowing something as right and not following the same is unethical forengineers

We have to believe that everyone is proceeding under the premise of appropriate caution and conservatism based on their understanding on how closed system relief piping works, and therefore it is not really an "ethical" problem, but rather disagreement of whose methodology and understanding is most mathematically sound vs economically viable to produce, which will inevitably vary based on application.

I understand,but ...

The US House Energy and Commerce Committee held a hearing on April 6, 2011 billed as “The U.S. Government Response to the Nuclear Power Plant Incident in Japan”.

On that issue,Martin J. Virgilio, Deputy Executive Director, Reactor and Preparedness Programs, once had testified earlier and had provided a very polite explanation of the fact that "the NRC, like the rest of the nuclear industry, encourages questions and doubts. It is part of our training and culture, inculcated to all practitioners from the earliest days of the nuclear industry. We are supposed to have a questioning attitude, to avoid complacency and to express a bit of humility".

Being from nuke background, we, technical auditors, are baptized in the name of "questioning attitude" and invoke Dr. Michael Corrandini who testified at the same hearing referred above,

"Engineers are taught to trust no one’s math other than their own – and then only if they have second and third checked it".

USA is safe in the hand of USNRC because of the ethics of engineers!

Closed discharge PSVs are part of safety systems where we need to improve our methodology, if found insufficient wrt mathematically sound procedure.

reg,
sambhu

Sam,

Indeed, here the problem is that the knowledge is still rudimentary. Maybe in 10 or 20 years will be easy to have in seconds a FEA fluid simulation of PSV opening and to activate a time history analysis. At that time will be a matter of Truth, Knowledge and Ethics.

Today we imagine that the PSV opening generates a plane wave- something like a piston in a tube. We cannot describe in math how the PSV's disk moving on vertical generates a plane wave in horizontal (and fact is some other models imagine there a jet instead a piston effect) and this failure of knowledge forces us to look around desperately for a math expression just to generate an input. This math expression is the API formula- F1 force- that describes a "free jet".
We are not able to say why this expression is linked with wave amplitude or to describe the wave propagation in terms of gas-dynamics. We just pretend we know the amplitude, shape, speed, etc of the wave based on F1 force, opening time, speed of sound, etc. Here is nothing related to Truth, Knowledge and Ethics, it is rather a dummy procedure that may give an acceptable design validated by experience. The problem here is that one can say that just applying F1 horizontally in PSV gives also an acceptable design validated by experience and I cannot argue based on Knowledge and Ethics...

Dear Mariog,

Knowledge, in this regard, is avaiable in Intergraph MEN Jun 94 pages 8-13 regarding fluid transient wave load experienced due to pressure transients. Steam hammer loads discussed here is similar to loads faced by PSV relieving disussed in pages 401-403 of Peng's book on Pipe Stress Engineering.

reg,
sam

Does this vertical force at PSV F2 in spreadsheet attached DEE is an unbalanced vertical force at this point.If yes in which direction it should act?
Most of the references do not talk at all about this vertical force at PSV.

Dear sam,

I'm sorry, for me is not really a similitude.

Just to start: I would observe that Peng book does not mention any similitude and I would assume that Mr Peng has the knowledge to see it, in case it is so evident.

My remarks.

First, Peng book discussion is about closing a valve. One difference is that- for Peng model- the fluid medium (steam) is already there, in the pipe line. When the PSV is opening, the gas is released in the line, which is more or less empty in practice and "empty" in theory. Is that theory dependent or independent on the presence of fluid medium?

Second, Mr. Peng presents a model based on physical considerations- formula (12.56), page 403, expression of "total pressure rise". The total surge force (12.57) is a consequence of the pressure rise.
For PSV case it seems we are unable to evaluate the pressure rise. But we pretend we know the force because we have in hands an API document mentioning a force. We are not focused on the physics beyond that formula, we are not really interested which is the meaning of that formula, we use it and find the equivalent "pressure rise" of the wave. Is more important the formula or the physics phenomenon related to?

And just to maintain the discussion on the coordinates you've set-up and trying to conclude, do you think is the above "knowledge" applied inside an ethical system of engineering?

Dear M Waheed,

There are two questions.

First: is there a vertical force?. My opinion: there is a hydrodynamic one applied to the PSV disk, but probably the net vertical force transmitted to the piping can be neglected. I cannot prove my opinion. But here I posted a sketch and a question: which is the net force transmuted to that "device" (which is not a PSV)?

Second: is the F2 formula based on a physical model? My opinion: No. It is rather a result of the work of a person that imagined he/she has some knowledge in fluid mechanics. I can prove my opinion and I did it in a former post.

Hello Stressers,

This is indeed a very fruitful discussion. Although this is not so related in this topic, I would like to share also what I have read in a (Famous FEA software in the field of piping and pressure vessel) article regarding acoustical frequency (due to wave speed and acoustical length) coupling to mechanical natural frequency causes resonant vibration and ultimately fatigue. The example given is for PSV piping system (although can happen also in some other high energy systems) in which fatigue failure occurred in an elbow. Additional restraints were added in the system to de-couple the systems acoustical/mechanical frequencies and therefore avoid vibration.

Maybe, in the design stage we can use the EIG guideline regarding FIV to filter if there is high turbulent energy in the piping system and maybe take a closer look in the restraint system of the piping system. Although, I have read in one article that EIG guideline for some case is sometimes conservative as they have proven a one system successfully running but not satisfying the EIG guideline.

Any other opinion is greatly appreciated.

Cheers!!!

Re: M Waheed.

To answer the question about imbalanced forces. Remember, when someone says "relief loads," they can mean:
1. The relief valve (or rupture disc or rupture pin, etc) activation force, where the surface that is normally sealed becomes unsealed. (t=0)
2. The resulting pressure wave as the discharge piping goes from P(operating) to P(relieving). (t=0 to some small number)
3. Steady state conditions. (t=∞)

At step 3, most forces have balanced themselves out. The exception would be for any tail pipe. Even at the relief valve. There is, as Mario indicates, a hydraulic imbalance as a result of having P1, A1, v1, and mdot on one side of the device and a P2, A2, v2, and (same) mdot on the other side of the device. I will note that this same force exists at every reducer and every elbow and especially reducing elbow, but we have a habit of ignoring those, because we assume they're minimal or otherwise cancel themselves out with other bends (though technically, this places piping into tension).

Direction of force of fluid onto pipe is always opposite the direction of flow.

Re: Mariog

Liquid relief loads and liquid hammer loads are directly analogous. Both result in a nearly instantaneous exchange of energy between pipe and liquid.

Imagine a pump, long pipe, and open valve at the end. Close the valve. Water hammer, right?

Imagine a pump, long pipe, and two valves at the end, 90° from the pipe, 180° from each other. One's closed, one's open. Open one simultaneously as the other one closes simultaneously. There is no hammer load, despite the fact that you closed a valve.

Imagine a pump, long pipe, and one closed valve at the end. Assume the pump's on recirculation or is capable of withstanding 0 flow but still supply the pressure. Open the valve at the end, and potential energy is released, and you have the same hammer loads, except in the opposite direction of opening said valve.

Replacing liquid with gas, notionally nothing changes. The loads are still there, but the energy is swept up into the gas's ability to compress, and the pressure waves are vastly mitigated.

re: Borzki

I agree that it could be theoretically done... Finding the longitudinal wall natural frequency assuming an infinitely long run of pipe, and compare with the frequency of a water hammer frequency, similar to pipe configuration's natural frequency. But I would think acoustical fatigue due to relief / hammer loads is a subset problem within a subset problem, for which to truly design, we'd have to consider a host of real world conditions that would be extremely arduous to fully consider. Imagine having to guarantee your pipe not only in new conditions, but corroded conditions... and everywhere in between. Now do the same thing for Tmax and Tmin... and everywhere in between. We typically don't do that last part now with CAESAR. We analyze at discrete temperatures. Who's to say there isn't an intermediate temperature that can't result in higher stresses due to interactions with non-linear supports?

It's a good thought experiment, though. One day we might get there.

Hi Michael,

Thanks for your opinion. That's true, it's hard to address all the possible scenarios that can exist in a piping system. Now I know why my senior stress engineer when I was still a junior stress engineer let us run a quick modal analysis to check if our piping system has enough guide and restraint. I think this is one of the reason to cater for those unknowns. But this check will only work fine for not so high temperature system because as we know thermal stress and dynamic stress has opposite solution. So it's not a mandatory requirement but rather a rule of thumb. That's why I have seen some high temperature systems experiencing vibration due to the fact that enough flexibility is required to qualify the system and dynamic stress is addressed as it happens in the field (since it's hard to address in design stage).


Any other opinion is greatly appreciated.

Cheers!!!!

Michael has rightly commented .. there is no difference between fast transient generated by steam turbine inlet psv closure and psv opening or closure on pipe legs bound by two elbows, so far as force time histories are concerned.
Anyone working with fluid transient software know it very well since long.

In recent past, we have seen spring cans toppling in intermediate legs of centrifugal compressor hot or cold bypass line similarly just after opening of the bypass valve as the first few legs were only axially strengthened, not the longest one in pipe rack.
Regarding ethics,we engineers, earn to live a life worth living while saving people who depend on us!
Unethical life is a dog's life!
Sun will rise again, however dark the night is! We have faith in Almighty and ourselves!

Reg,
Sam
Refer the link to know the load .. tail pipe has both loads, while intermediate ones only wave load.

https://www.google.co.in/url?sa=t&source=web&rct=j&url=https://cdn2.hubspot.net/hubfs/232514/Fauske-2017/PDF/EvaluationofDynamicTailPipeLoadsDuringReliefValveDischargeWithRELAP5.pdf%3Ft%3D1505490261269&ved=2ahUKEwjM49uds_fYAhXFppQKHR9cDjs4ChAWMAJ6BAgTEAE&usg=AOvVaw0JDLzRmfbxmQRuMEpIzcLG

Thanks sam for sharing that information. Nowadays, it is becoming common to do fluid structure interaction multi-physics simulation, to see how the system will behave under certain conditions. Although, it takes a lot of patience to study and understand the theory behind this.

I have seen some consultants really setting up physical experiments to validate & back-up their software.

Cheers!!!

Micheal,
I guess that studying more would change your strong believes. It happened to me, I suppose this would happen also to you. Unfortunately this is not the place where I can help you in this regard.
Of course, noting wrong to remain to the same opinions.
And I repeat what I said previously: I can follow any specification imposing how to proceed considering PSV forces, even I often consider them rather absurd. The "wave" theory based on API force is not the worst but I remain to the opinion already posted.

Sam,
It is my opinion that abundance of words about ethics is the worst strategy one can have when try to convince others that the theory just discovered is the best. I do not question your good intentions however I do not second you. Otherwise my opinions about truth and knowledge are rather following Kant approach.

Mariog,

Do you recall the source of the information that changed your mind?

Thanks.

Sure, Michael. My PhD in fluid mechanics.