Greetings,
We are working on a piping system that included 2”NPS Fisher globe valves with pneumatic diaphragm actuator (17” diameter). The piping system has few bends and it is welded on to a central header.
The issue is that the actuation force we are coming with is close to 5000 lbf and we are trying to do static analysis with DLF = 2. Applying 10000 lbf at the globe valves fails the piping system on both sides of the valve and produces more than a 0.25” displacement.
We are thinking of securing the globe valves to the structural steel of the unit.
Please advise if you think the fundamentals of applying the force at the valve as a vector are correct or if you have dealt with this issue before how did it get resolved.
Regards,
K_J

Hi K_J,
can you show us where actually you had applied the force? If I got you correct, this force is due to the accelaration values. Correct me if i am wrong. I had deal with this situation before in one of our project. Please show us (image) how the arrengement will be made in Caesar-II.

Regards,
RK.

PK,

Thank you for looking into my question. Attached is the image of the jumper. The force is applied to the globe valve center. The second valve that shows in the image is an isolation valve.

Regards,
K_J

The usual case would be that the actuator is attached directly to the valve top and thus, whatever the closing force applied, the piping does not experience the load.

Yours is a wierd case indeed. Time to change the design ?

MoverZ,

Wouldn't the globe valve experience shock load from the actuator slamming the stem close? The valve body would transfer the force to the piping and I think that is what is causing the displacements we see. The senior engineer mentioned that he has seen the globe valves "shaking" every time they close. The piping might be too flexible(2" NPS).
The question is if it is fundamentally correct to load the globe valve in the CEASAR model with 10000 lbf (assuming DLF of 2)in static case. The valve closing time is close to 2 sec => the static analysis.

I appreciate your input on the matter.

Regards,
K_J

A closing time of 2 sec for a 2" valve is not fast and would not normally be a problem. I'm not clear where your 5000 to 10000 lb load originates. I would caution, don't confuse an imposed force from the diaphragm with mass. The moving MASS would be the valve spindle etc., and yes stopping that would have an external effect, albeit small in your case. However, the valve body and stem / actuator connection would resist the FORCE and there would be no external effect from it.

Hi,

The 5000 to 10000 lbf are coming from the diameter (16-17") of the diaphragm and the 50 psi supply pressure to move the actuator. The globe valve is normally closed so in case of complete loss of pressure the potential energy stored in the actuator springs will close the valve with 10000 lbf in less than a second. It also possible in different application to look for faster closing time of the same valve.
The common sense answer is that the force that transfers into the valve body from the actuator is based on the stem mass and acceleration as you mentioned.

The issue is that I am advised by my senior to apply the 10000 lbf actuator spring force as a vector force in CEASAR and the piping fails all around the valve. Now we are re-designing the structural steel to tie in all the globe valves and in my opinion this might not be necessary.

I need to open few books and educate myself a little more on the subject before I bring my questions up the chain.

I appreciate your input and experienced advise on the matter.

Cheers,
K_J

I agree with Moverz.
Maybe was a faulted valve. If this is a real problem, vendor should worn about, but until now I never heard.

Regards,

I second MoverZ opinion


I'll try to detail; unfortunately I haven't accurate data for your case, however I would try extrapolate form data I have...

So let's say the spring diaphragm actuator start moving at a pressure of approx. 3..6 psig and this correspond to a pretension in spring of aprox 300 daN.. 600 daN. The actual values are not so important for this discussion; I guess more important is to understand that in "no pressure" case there are two internal forces of this magnitude, transmitted to both seat and bonnet of your globe valve.

I think also that the the spring force value in the maximum compression is not important for your evaluation.
Under normal circumstances, the movement dynamics of stem/ valve closure is such that the "viscous damping" is more important than inertial effects, so the movement is overdamped.
So, having no important inertial effects, in the end of closure process will be again two forces of the same magnitude as 300 daN..600 daN.

I would add that -in normal service- the load to seat is not dynamically applied because in the end of stroke there are hydrodynamic forces (not only the pressure forces) that contribute to this behavior.

My opinion, of course.

best regards

K J,

Quote ...
The issue is that I am advised by my senior to apply the 10000 lbf actuator spring force as a vector force in CEASAR and the piping fails all around the valve. Now we are re-designing the structural steel to tie in all the globe valves and in my opinion this might not be necessary.

I agree with you here, but I think it is your senior guy who needs to open a few books.

Thank you all for the replies and the advise!

Regards,
K_J

If the globe valves are observed shaking on closure, this is more likely to be a flow induced vibration.

If you really wish to consider dynamic effects there are two elements:
a. fluid forces - i.e. changes in velocity and direction of fluids - I suspect that if you quantify these they will be very small

b. mass dynamic effects - the moving elements of the valve - this would be the spindle and globe and part of the diagphragm.

You would need to make an estimate of the valve stroke - for example a velocity vs time graph, calculate the accelerations and then the forces that this generates. Again I think you will find that even if you take a conservative approach (very short stroke time, ignoring any fluid cushioning effects, including DLF of 2), the small amount of mass will mean the forces are very small.

My approach would be to assume these forces were small enough to ignore for a typical pipe stress analysis, but knowing that control valves can be subject to flow induced vibration, I would ensure a sensible pipe support arrangement in and around the valve to hold the train firmly.

Ross,

I second your conclusion.

Referring to the theory I have few remarks on your explanation- maybe with no benefit for the majority of this forum.

In case of complete loss of air pressure, the spring force will move the spindle.
You say that the way to estimate what is happening is to calculate the acceleration from velocity vs time graph, to estimate mass of moving parts and then estimating forces (I guess based on Newton law F=ma)

I think you should add point c) internal friction forces.

Presuming the MFR- Fisher data would be available, I think the correct way is to consider the spring force (assumed linear with stroke), friction forces, fluid hydrodynamic forces, mass of moving parts and to write down a second degree differential equation based on Newton law...

Presuming the friction forces are important and their behavior is of "viscous" type, and presuming also that the rest of forces are small (as they are in control valves), it results that the inertial term is small and the second degree differential equation is reduced with a good approximation to a "first degree differential equation". The same is happening in theory of harmonic oscillator with viscous damping- when the movement is "overdamped".
So the spring force may be important, but the inertial effects are small because friction plays an important role.

To resume this explanation, the exact math of moving dynamics would be very complicated (and in practice it is not necessary to develop such academic works), the active force is the spring force (it would be considered as "big") and the main effect of this force is movement with small acceleration because frictional effects.

Under normal circumstances, the forces involved remain as internal forces in "valve+actuator" system and the field operator will remain "secured" in front of this theory...
The case of "flow induced vibration" is a very special case where flow has capability to transfer fluid momentum/ energy to piping in an "unexpected" way and a math model has nothing to do with the above explanation.

Best regards.

PS. Based on "similitude" of flow pattern in PSV and in Angle-Style Control Valve, I expect in few years I'll be obliged to consider in Control Valves the same "reaction forces" as in PSVs...
K_J, would you suggest this to your supervisor?