30 replies to this topic

[sheiko]

Hello experts,

I am a french process engineer working for an EPC company and i am facing a little "problem" during a revamp project:

The system under study is: vessel + centrifugal pump + control valve + column

The scope is that the flowrate through the system will increase and, thus, the existing pump will be replaced by a new one of the same type and i have to assess the possibility to keep or not the control valve downstream of the new pump.

Could you please let me know the proper methodology to perform that knowing that the only data i have about the control valve is its model (masoneilan camflex II), size and rated CV (at 100% opening)?

[djack77494]

sheiko,
My response to your "little problem" is quite simple. Your control valve must be sized so that it can control your system. I generally like to see a maximum expected flowrate through the valve happen when the control valve is less than 85% open. [This max expected flow is something you could reasonably expect and that might persist for at least (say) several hours.] If you don't have some "room to maneuver", then you are not in control. Similarly, a minimum expected flowrate should happen with the valve at least 15% open. Considering changes in the hydraulic losses through your piping, the control valve must be able to force the pump to move up and down along the pump curve so that the desired range of flowrates is achievable. Since both pumps and control valves have wide ranges in behaviour of flow versus pressure/head, it is difficult to much generalize. Generalizations are appropriate during conceptual design, but must be refined as one proceeds towards detailed engineering. You appear to be near the point where a "final design" may be selected.

[JoeWong]

Establish the control valve requirement i.e. maximum / normal / minimum flow and correspondence pressure drop. Ensure the pressure drop across control valve has sufficient control margin i.e. 25% of total pressure drop from pump to column. With these value, establish the CV for each conditions. If the control valve opening at required CV are in the range of 85%-90% (you may need the characteristic curve for this purpose...check with vendor), i think it is fine...

Apart from control valve suitability, other checks like cavitation, noise, material suitability, etc. You may needs expert or vendor advice.

[Jiten_process]

hi

As far as control valve is concern, it seems you are not replacing the control valve. And to check the suitabiilty for existing control valve for the matter mentioned by djack and joewong you need to have control valve datasheet and characteristic curve.

First, see trim characteristic, equal percentage or linear..(most cases it happens to be equal percentage only). once you know this, take the rangebility figure for globe valve (generally assumed to be 50), take the rated CV and find out percentage opening for the desired new flows by finding required CVs.

you have to use Delta P values based on new pump's characteristic curve to find out required CV values.

Result should be in accordance to what hav already mentioned by djack and joewong. However it is advisable to consult vendor as said by joewong.

good luck...

[fallah]

There is a question: Is the set point of Control Valve changed during revamp's engineering study?

[Zauberberg]

Pump performance curve is independent of system layout, you should already have it with you - from the pump vendor.

[fallah]

Your piping system also could be affected and therefore should be resized. I think, at the first, you should finalize this issue.

[sheiko]

Thank you

This issue has been finalized

[fallah]

[/quote]

Thank you

This issue has been finalized
[/quote]
You can size Control Valve for new situation by having:
-Normal/Max./Min. Flow
-Correspondance DP for each flow (having size and isometrics of piping)
-Assuming percent of DP (say 25%) dedicated to Control Valve and estimating DP for it in each flow
-Model of Valve and using relevant software

And finally compare new control valve with existing one and taking final decision.

[djack77494]

sheiko,
Don't be afraid to push your boundaries a bit with the design. Revamps are not the same as grass roots designs, and they shouldn't be treated that way. For a grass roots design, I might not want my control valve opening to exceed 85% under maximum continuous operating conditions. For a revamp, though, you need to be real practical. If the valve can do the job but your calcs say it would be 95% open, don't just recommend replacement. Time to give the situation some real serious thought to evaluate how likely it is that you would ever want to use the full available Cv of the valve. What are the consequences of running into a control valve limitation from time to time? If the situation is acceptable, make the best recommendation of all - no change.

[Jiten_process]

Dear Joe,

If i well understand your methodology, knowing the flowrate and assuming the existing control valve pressure drop at a given condition (by taking 25% of total pressure drop from pump to column) allows me to calculate the required CV. Then, knowing the characteristic curve i can see in a catalog or with vendor if the opening is in the good range. Is my understanding correct?

No, you cant do this since you have changed your pump with new one (pump model and curve is changed...).

anyway, your matter is finalised let me just address your questions.

rangeability is defined as the ratio of maximum controllable flow to minimum controllable flow, it depends on trim characteristics and type of valve you selected. Mind well, rangeability is NOT the ratio of max. flow to min. flow (it is called as turn down...). Generally for globe valve 50 figure is sensible to take up. It will be used to find out the %opening of your control valve for defined flow & CV values (rated & calculated).

Pump performance curve you need to have with you by applying priliminary DP values for control valve and subsequently checking it with real one. This is how it is normally done in practice.

hope it clarifies...
good luck...

[smalawi]

Hi,

Can you let us know what was finally done ?

For me, its simple ! download the exsisting valve cv curve or better the free vendor valve program (yes they have one and its free and you need it) establish the min / norm / max and back calcualte the Dp from CV @ max valve open conditions, normally at 85% (check valve type)

generally, check the valve operation with vendor software at all possible flows and see that the % open does not fall too low or too high (15 to 85 %) is typically ok for good control.

check how much Dp you have and how much you require for good control and cross check with your new pump curve (from pump vendor). If you need 1 bar and pump gives only 0.5 bar at max flow you need new valve or new trim.

Just make sure your new pump shut-off head is within pressure design limits or you will need a RV to protect the line/equipment.

cheers,

SM

[sheiko]

Establish the control valve requirement i.e. maximum / normal / minimum flow and correspondence pressure drop. Ensure the pressure drop across control valve has sufficient control margin i.e. 25% of total pressure drop from pump to column. With these value, establish the CV for each conditions. If the control valve opening at required CV are in the range of 85%-90% (you may need the characteristic curve for this purpose...check with vendor), i think it is fine...

Apart from control valve suitability, other checks like cavitation, noise, material suitability, etc. You may needs expert or vendor advice.

About control margins:

• In a pumped circuit, the pressure drop allocated
to the control valve should be equal to 33% of the dynamic
losses in the system at the rated flow, or 15 psi,
whichever is greater.
• The pressure drop allocated to a control valve in the
suction or discharge line of a centrifugal compressor
should be 5% of the absolute suction pressure, or 50% of
the dynamic losses of the system, whichever is greater.
• In a system where static pressure moves liquid from
one pressure vessel to another, the pressure drop allocated
to the valve should be 10% of the lower-terminal vessel
pressure, or 50% of the system’s dynamic losses, whichever
is greater.
• Pressure drops in valves in steam lines to turbines, reboilers
and process vessels should be 10% of the design
absolute pressure of the steam system, or 5 psi, whichever
is greater.

Source: Ease control valve selection, CEP magazine, Nov. 2002

[djack77494]

You quote so called "rules of thumb". That's fine if you are operating out of ignorance - if you have no other information and you're trying to estimate how your final design will be done starting with a blank sheet of paper. But you have a largely existing system, it seems. So you aren't forced to use blind/when there's nothing else "rules of thumb". I have repeatedly seen these "rules" put on a pedestal and treated as if they were akin to standards. Obviously they are not.

In your case, if you've already finalized your piping, then you essentially have your system curve. Or at least, all but the control valve. Make a preliminary pump selection. That should be fairly easy since most sites have prefered vendors and the vendors usually will happily supply selection software. Now you have your pump curve. Select a control valve that can force the intersection of the system and pump curves to vary over the range of flowrates that you'd like to achieve. Confirm that the system is still controllable at both extremes. Job done. It's that simple and the end result is a solution for your system and not some general purpose approximation.

[fallah]

Confirm that the system is still controllable at both extremes.

How? A brief explanation would be appreciated.

[sheiko]

Hi,

Can you let us know what was finally done ?

cheers,

SM

I finally used the methodology described in the attached article, which is basically in accordance with what the contributors of this post said.

Thanks to all.

Attached Files

•  Pump_Control_valve_Interaction.pdf   239.93KB   107 downloads

[sheiko]

One last question please. I am almost sure it is true but i would like a confirmation:

Does the available pressure drop across a control valve, located downstream a centrifugal pump, equal: Head developped by the pump minus Head required by the system?

[djack77494]

Confirm that the system is still controllable at both extremes.

How? A brief explanation would be appreciated.

Consider maximum and minimum expected flowrates, including any over/under design you wish to incorporate for off-design operation.

Assemble a "worst case" scenario for both of these cases. {Let me offer an example to assist in defining this. For maximum flow case, consider the pump's suction pressure to be at its minimum; for minimum flow case, consider the pump's suction pressure to be at its maximum. These situations are the hardest to handle.}

Calculate the pump's discharge pressure for both scenarios using the scenario's suction pressure and the pump curve. If fluid density changes are possible, then use the worst cases in a manner similar to pump suction pressure discussion.

Calculate hydraulic losses for both scenarios. In keeping with the general approach being outlined here, if variations in fluid viscosities are expected, use the worst cases.

Use the worst cases for destination pressure and any static head that must be overcome.

When these scenarios are completed, you will have a worst maximum flow case and a worst minimum flow case. Calculate (by difference) what pressure drop the control valve takes in each case.

Knowing the control valve size and characteristics, confirm that the control valve POSITION for both scenarios is not overly close to the maximim or minimum. If the control valve is adequately away from the full open and full closed positions, then it should perform adequately. If not, you will need to replace the control valve (i.e. because your system would NOT be controllable).

One last question please. I am almost sure it is true but i would like a confirmation:

Does the available pressure drop across a control valve, located downstream a centrifugal pump, equal: Head developped by the pump minus Head required by the system?

sheiko,
In simple cases, the answer is "yes". I hesitate because the situation is more complicated when there is a superimposed pressure on the suction and/or destination location. Perhaps a better way of describing things would be to say that the discharge pressure of the pump (which is calculated for a set of conditions) minus the hydraulic losses (as a differential pressure) minus the destination pressure is equal to the control valve pressure differential. Be careful when discussing pressures - always indicated whether you mean gauge, absolute, OR differential pressure.

[sheiko]

Thank you,

I was considering differential pressure: in bard

Could you please give an example of superimposed pressure at suction or discharge of pump, because i have difficulties in understanding what you mean...?

[djack77494]

Could you please give an example of superimposed pressure at suction or discharge of pump, because i have difficulties in understanding what you mean...?

If you are pumping from (say) an open pond or pit to a second open pond or pit, then the superimposed suction pressure = superimposed discharge pressure = atmospheric pressure (or zero gauge pressure). Conclusion: Don't worry about superimposed pressure. Instead, if you start at an open pit and pump into a drum operating at 10 barg, then your superimposed destination pressure will be 10 barg; suction will be 0 barg, and difference will be 10 - 0 = 10 bard. So let's generate an example. Say that at some maximum flowrate your hydraulic losses are equivalent to 100m and that the drum's liquid level is 20m above the level at the suction pit. Also, let's imagine that you've decided you want your control valve to have a loss equal to 80m at these conditions. The pump must generate a head that is equal to the static head difference (20m) plus hydraulic losses (100m + 80m) plus the difference in superimposed pressures of 10 bard. (The 10 bard must be converted to its equivalent head in meters, which I'll leave for you to do.)
Hope this helps,
Doug

[sheiko]

Could you please give an example of superimposed pressure at suction or discharge of pump, because i have difficulties in understanding what you mean...?

If you are pumping from (say) an open pond or pit to a second open pond or pit, then the superimposed suction pressure = superimposed discharge pressure = atmospheric pressure (or zero gauge pressure). Conclusion: Don't worry about superimposed pressure. Instead, if you start at an open pit and pump into a drum operating at 10 barg, then your superimposed destination pressure will be 10 barg; suction will be 0 barg, and difference will be 10 - 0 = 10 bard. So let's generate an example. Say that at some maximum flowrate your hydraulic losses are equivalent to 100m and that the drum's liquid level is 20m above the level at the suction pit. Also, let's imagine that you've decided you want your control valve to have a loss equal to 80m at these conditions. The pump must generate a head that is equal to the static head difference (20m) plus hydraulic losses (100m + 80m) plus the difference in superimposed pressures of 10 bard. (The 10 bard must be converted to its equivalent head in meters, which I'll leave for you to do.)
Hope this helps,
Doug

Thanks

However, considering that, the head developped by the pump and the head required by the system both include the pressure head terms of source and destination drums, i thought the effect of superimposed pressure would cancel when differentiating the heads. No? I hope my description is clear enough

[djack77494]

However, considering that, the head developped by the pump and the head required by the system both include the pressure head terms of source and destination drums, i thought the effect of superimposed pressure would cancel when differentiating the heads. No? I hope my description is clear enough

sheiko,
This may just be a problem with semantics (terminology). When I consider the head developed by the pump, I am speaking only of the mechanical (or hydraulic) abilities of this piece of rotating mechanical equipment. It has nothing to do with any superimposed pressures. When I saw your statement about the head required by the system, I took that to mean pressure differences which were normally or commonly expressed as head (i.e. linear dimension). I see that you are using the term differently and that you are "pulling in" the superimposed pressure terms. Let us not belabor what to call the various terms. My question to you is, "Do you understand and accept the example I previously generated?" If your answer is "yes", then we are in agreement. If not, please provide additional information about what you dispute or don't understand.

[fallah]

Thanks
However, considering that, the head developped by the pump and the head required by the system both include the pressure head terms of source and destination drums, i thought the effect of superimposed pressure would cancel when differentiating the heads. No? I hope my description is clear enough

In a stable pumping system:
Developed head by the pump=(superimposed discharge pressure-superimposed suction pressure)+(velocity head in discharge line-velocity head in suction line)+(static head of discharge line-static head of suction line)+(hydraulic loss in suction line+hydraulic loss in discharge line)

[sheiko]

Thanks
However, considering that, the head developped by the pump and the head required by the system both include the pressure head terms of source and destination drums, i thought the effect of superimposed pressure would cancel when differentiating the heads. No? I hope my description is clear enough

In a stable pumping system:
Developed head by the pump=(superimposed discharge pressure-superimposed suction pressure)+(velocity head in discharge line-velocity head in suction line)+(static head of discharge line-static head of suction line)+(hydraulic loss in suction line+hydraulic loss in discharge line)

1/ What do you mean by "stable pumping system"?

2/ I rather think what you describe is the system head

[fallah]

Thanks
However, considering that, the head developped by the pump and the head required by the system both include the pressure head terms of source and destination drums, i thought the effect of superimposed pressure would cancel when differentiating the heads. No? I hope my description is clear enough

In a stable pumping system:
Developed head by the pump=(superimposed discharge pressure-superimposed suction pressure)+(velocity head in discharge line-velocity head in suction line)+(static head of discharge line-static head of suction line)+(hydraulic loss in suction line+hydraulic loss in discharge line)

1/ What do you mean by "stable pumping system"?

2/ I rather think what you describe is the system head

1-"stable pumping system"?is a system that could be considered dinamically stable without any considerable change in its main characteristics such as pump RPM,flowrate,pressure heads,....
2-What i described is an equation relates developed head by the pump to system head....