Series connection of FCV and PRV

[Kadongo]

Hi All,

Many Thanks for your help,

I need to know what are the considrations to be noted in case of a series connection of flow control valves group and PRV group as per the follow pic:

I know that those valves shall fight each other but I am not sure about this? and what is the considreations in tems of hydraulic sizing?

Many thanks

 

[LittleInch]

Why two valves?

Just use one and feed the position controller with two signals and then use a low selector block to drive the valve.

one input / PIC controller will provide a position of the valve base don keeping the pressure at 20bar, the other will control the flow base don whatever figure you put in it.

The valve will then take the lowest position in a seamless manner.

I've done this with one control valve having four different inputs including a manual control position to be able to start it gently. Worked like a dream.

Is this gas or liquid?

But why would the valves "fight" each other? You're controlling on two different parameters. Probably need to do some tuning of the response time to avoid wide fluctuations, that's in the PLC control loop

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Kadongo]

Hi Littleinch

Thanks for your reply.

The valve must be a Plunger Valve.

The service medium is Water

The PRV will down the pressure from 50 to 20 bars and in this case we have no control over flow.

After that we provide a flow control valve on the same line to provide the required flow.

They will fight as the PRV will provide specific flow to control pressure and the FCV will control flow to provide the required flow value specially that the distance between them is small. "That's why they will fight, I understand that at the end PRV is flow control valve also"

The flow is variated from max to min value.

 

[1503-44]

2 control valves will always fight to the death, if they are "within range". The meaning of "they are within range" is, if changing the PCV's pressure setting affects the flow at the FCV, or changing the FCV's flow setting changes the pressure at the PCV. Obviously that creates an infinite feedback loop, at least until one valve or the other reaches a point where it has no effect on the other. That's often at full open, or full closed, or, depending on relative gain settings, some other spot of no use, where it is in effect dead, at least as far as its effect on the other valve is concerned.

The valves will work, if out of range of the other, perhaps at a far enough "hydraulic distance" where the effect of one on the other is very small. Small meaning that whatever effect on its control variable is tiny enough to escape detection completely by the sensor, or by dulling its position change signal at the valve's PIC by setting its gain very low. 2 valves immediately next to each other are not hydraulically isolated, pressure at the PCV immediately affects flow of the FCV and vice versa, so you can be assured that they will fight each other in an infinite feedback loop. That feedback loop is from the hydraulic characteristics of the system; that pressure at one valve affects flow on the other, and v/v. The PCV's influence on flow at the FCV, or the FCV's influence on pressure at the PCV are hydraulically locked together. You would have to tune at least one of the valves out at its PIC in order for the other to work. Otherwise one or the other, or both, will move to their full travel positions, as I described before.

The good news is that you never need 2 control valves within range of each other. The reason that there are 2 valves there is because someone does not understand that FCVs will yield a certain pressure at a given flow rate setting and that PCVs yield a certain flow at a given set pressure. So either can be used to control the other variable, if you know what affect the valve will have on that other variable. You see that both an FCV and a PCV are basically just a valve. They both have a certain change in flow rate at a certain change in pressure at some position. If it is a FCV, I want to set the position, obtain the flow rate change I want and basically accept whatever corresponding pressure change it gives me. If it is a PCV, I want to set the position, obtain the pressure change I want and basically accept whatever corresponding flow change it gives me. These effects can be seen on the valve's curve that usually give a Cv, or metric Kv value for all positions. You take the Cv value, apply your flow rate and get the pressure change. You then see how that pressure change affects your system. Or you could alternatively solve for flow rate with the Cv coefficient by applying the valve's pressure drop and getting the resulting flow. You then have a look at your system to see how that flow rate affects the pressures. Knowing how flows and pressures change in relation to each other at any valve, an FCV, or a PCV, will allow you to determine the flow resulting from a PCV setting, or the pressure resulting from a FCV setting. With that knowledge, you can actually control both, with either. The only difference being what signal you choose to use to effect control, pressure, or flow. If you use the flow signal, you have an FCV. If you use the pressure signal, you have an FVC.

Note that valve fighting is not limited to PCV against FCV. Two FCVs and tow PCVs can go at it in the same manner.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[mk3223]

They will fight as the PRV will provide specific flow to control pressure and the FCV will control flow to ……

Should the PRV be used to provide a fixed pressure at downstream instead of the fixed flow prior to the FCV?

IMO, if the PRV outlet pressure dropped due to the more flow through FCV, the PRV may open more to maintain the constant pressure which may add more flow to FCV as required. However, the response time of the system pressure and flow rates may be considered.

 

[LittleInch]

A lot here depends on the operating mode and rate of change. SO long as the FCV is fairly constant or slow acting (generally slower than the PCV, you won't get "fighting". Only if the two are similar ramp speeds / time to respond or lots of fast changes to flow or pressure could the two conflict.

Your PCV controls to 20 bar. That's its function in life and it will control to 20 bar within I guess a fairly large flow range.

The FCV controls to a set flow ?

So long as that set flow is within the min to max range of the PCV, i.e. somewhere between 20-70% open, then you shouldn't have an issue here.

It's all about controlling the speed of response and making sure one is a lot slower than the other. Even self powered actuators have controls within them to modify speed of response, for exactly these reasons.

But why can't your FCV work with an inlet of 50 bar? Buy a different valve if you have to but then you only have one valve.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TugboatEng]

This is a fairly standard configuration. I see it used for soft engagement of clutches. There is a better method but this works as well.

 

[1503-44]

A downstream pressure regulator can be set to 20b. It's flow will then be dependent on the upstream pressure. If that upstream pressure increases, the PRV will close a bit to hold 20b and flow through the regulator will tend to decrease. I say "tend to decrease", because the upstream pressure went up. Now the differential pressure dP across the valve has increased, so the PRV flow will go to whatever flow corresponds to that new dP. The upstream system plays a part here. The PRV inlet pressure just went up. That might slow the flow upstream, but it might not. If the upstream segment of the system holds the same flow rate with the higher pressure at the PRV's inlet, then the new pressure at the PRV inlet will not change any further. If the upstream flow rate decreases, due to the higher PRV inlet pressure, then flow through the PRV changes to match that upstream flow rate. If that results in a 20 bar downstream pressure, it's done. If not, the regulator opens or closes again to match 20b discharge and this cycle PRV-upstream cycle starts again. Whenever it is cycling, it is adjusting its flow rate.

The flow rate coming from the PRV goes into the FCV. IF the flow is higher than the FCV setting, the FCV closes a bit. If it is lower, the FCV opens a bit. The flow out of the valve is now at the FCV setting. If that flow matches the flow through the PRV, it's done. If flow is different than the PRV than its not over, and the pressure at its inlet will rise or fall so that the FCV flow rate maintains its setting. It does that by changing the pressure drop across its inlet and outlet. That can change the pressure at the PRV outlet. If it changes, a new cycle of the PRV begins once more. If there is no change, then everything holds as is, for the moment, because now we have to see what is going on downstream. There is a downstream cycle just as there was an upstream cycle.

The downstream cycle begins at the FCV's outlet. If the FCV adjusted its position to control how, it now has a different outlet pressure and we have to see what happens downstream as the result of that pressure and the FCV's flow. If the downstream segment holds, its done. If not, the pressure there will start to change. If that causes the pressure in the downstream segment to change, it may cause a new flow at the FCV outlet, which will cause the FCV to cycle. The FCV will want to increase or decrease its differential pressure to hold its flow rate setting. If the differential pressure changes, the FCV cycle is initiated once more. As such, any mismatch will cause a cpotential cascade of actions throughout the system. If it finds a solution where all pressures result in the 20b at whatever the FCV setting is, the cascade stops and things hold stable until the next upset happens. But usually it never finds a stable position and all the cycles are triggered, which results in valves pegging in the wrong positions, or a continuously cascade, which results in somewhat stable system, but with a lot of valve operator maintenance due to the continuous movements. That condition should be avoided. It can be stopped by reducing the sensitivity and subsequent response of one or the other valve settings, but then you have effectively said, "I really don't need that valve".

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[1503-44]

Tug, if the downstream accepts any flow rate without causing pressure change to propagate pressure beyond what the PRV can adjust its position to and still result in 20b at the outlet, its good.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

Mr 44 - Your first para doesn't hold water to me.

The Pressure regulating valve just regulates pressure. Flow is controlled by something else. If that flow stays the same then the flow through the PRV will stay the same, but the pressure difference will increase. The actual velocity upstream the valve will reduce, but mass flow through the valve is not dependant on upstream pressure, but is set by the downstream components.

You are correct that each valve can start to cycle up and down, hence some mpening needs to be inbuilt and if the frequency of the two valve movements gets too close they can start to feedback on each other until one is full open as the other is nearly closed then vice versa.

So you need to set one valve, probably the FCV valve, to be fast acting and the PRV to be much slower to damp out any oscillations.

But only having one FCV able to handle 50 bar incoming would be much much better. IMHO.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

Suspect this PRV is 300lb piping class , while the downstream piping and the FCV is 150lb ?
If so you need a PSV in between to handle any transient control issues at the PRV, which is quite likely. A cheap non pilot operated PRV will make matters worse. To minimise frequency of triggering this PSV, install a gas charged surge vessel or dampener upstream of the FCV to "soften" the transient. PSV should be sized in accordance with API rules.
This arrangement is not ideal. As suggested, switch out to a fully rated FCV good for 50bar, and select an appropriate Cv to get good turndown.

 

[1503-44]

This paragraph?

A downstream pressure regulator can be set to 20b. It's flow will then be dependent on the upstream pressure. If that upstream pressure increases, the PRV will close a bit to hold 20b and flow through the regulator will tend to decrease. I say "tend to decrease", because the upstream pressure went up. Now the differential pressure dP across the valve has increased, so the PRV flow will go to whatever flow corresponds to that new dP. The upstream system plays a part here. The PRV inlet pressure just went up. That might slow the flow upstream, but it might not. If the upstream segment of the system holds the same flow rate with the higher pressure at the PRV's inlet, then the new pressure at the PRV inlet will not change any further. If the upstream flow rate decreases, due to the higher PRV inlet pressure, then flow through the PRV changes to match that upstream flow rate. If that results in a 20 bar downstream pressure, it's done. If not, the regulator opens or closes again to match 20b discharge and this cycle PRV-upstream cycle starts again. Whenever it is cycling, it is adjusting its flow rate.

Please note that I didn't say "flow is controlled" by the PRV. I said it (flow) is dependent on the upstream pressure. Flow through the regulator is solely dependent on the pressure differential across it, just as it is for every flow element (something having flow). It does not matter what it is. Even a FCV. A FCV has a Flow rate setting, but its positioner controls its %Open to allow the differential pressure to change and drive a flow through it. If the flow equals its setting, position changing stops. In the case of a PRV, The same thing happens, but instead of stopping its positioning action at a certain flow rate setting, it stops when (in this case) outlet pressure is sensed at 20bar. The flow through the PRV will then correspond to whatever flow across its %Open area that its differential pressure at the moment will drive. Only its dP, not the upstream pressure, and its %Open, determined by where it should be to get a 20b output determine what its flow will be. No matter what the upstream pressure is at the time, The PRV does not care. It works to hold the 20b downstream, upstream pressure be damned, and the resulting dP drives whatever flow across it.

OK, say the PRV changed its flow rate by virtue of its need to hold the 20b outlet pressure. Assuming its a steady state flow in the system, holding that flow rate at the PRV will try to force all other flow elements in the system to do whatever they need to to feed the ZPRV that flow rate.
If upstream pressure needs to increase, it will look for more pressure. Maybe an oil well can supply it. So we look at the Pressure vs flow curve of the well. If the particular flow required in the upstream piping, now equal to the flow through the PRV can be supplied at some pressure, the well moves to that point on the curve, and it's done. If it can't reach the new flow, it meets what it can and delivers a flow that it can at whatever higher pressure it can reach. So now that flow rate is heading to the PRV. When it gets to the PRV, the PRV senses there is a flow rate change through it when (if) the 20B rises, or falls in response to that well flow rate's effect on the inlet pressure of the PRV and the new differential pressure across it. It changes position to hold 20b. That changes the differential pressure across it again along with determining a new flow through the PRN. That new flow rate now heads upstream to the well to see where that hits the well P-Q Curve. Now we look at the downstream half of the system.

The same thing is happening there, the new flow rates through the PRV are having similar effects on the downstream half of the system. The inlet pressure on the downstream half is 20b. The downstream system tries to make its outlet pressure to correspond to whatever it should be to pass the PRV's flow rate through its part. If the downstream outlet is OK with that P and Q, its done. But let's say that the pressure rose (or dropped if you want) so some controller there, a human being, goes out and nearly closes the outlet, that increases pressure there, which slows the flow there. The pressure starts increasing with flow rate dropping it in that half of the system. These effects continue moving upstream until they reach the 20b sensor and that changes the %Open of the PRV such that outlet pressure is held at 20B. Of course that changes the flow through the valve. Now we have to look at the upstream half to see what that does to the oil well pressure and flow rate. Etc. Etc. The cycling continues until some stable solution is found that satisfies all dP, 20B and the total system pressure drop that results in all flow elements reaching the same Q, OR it continues cycling forever. The gains in all the PIDs have to be set within ranges that allow the valves to position properly and long enough to reach stability. If they are not adjusted properly to allow that, cycling may continue forever, even though a stable solution exists. Its just all valves won't reach their stable positions at the same time. Some might need more gain, others less. At times it can be a bit tricky to get them all right. One reason to keep the numbers of PIDs and valves and other things that cause glow and pressure to vary as few as you can.

Did that make sense? I can make a simulation in Stoner that will clearly show all of these effects, if you need additional proof. I can put in a VFD pump, tank or an oil well to feed it and a tank or injection well or a constant pressure or constant flow Node, or an entire other system at its outlet. Or all. As long as there is at least one source and one sink, I can sim it.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[georgeverghese]

Whether one configuration will work better than another will depend on what is downstream of this FCV, and what is upstream of the PRV. A blinkered view is all we've got for now.

 

[1503-44]

Specify both system inlet and outlet behaviors from these possibilities
Select one from each line.
P:....... Up.........NC.......Dn........|
Q:....... Up.........NC.......Dn........|


--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

Mr 44 - You know me so I'm not arguing for the sake of it and we're probably talking about the same thing.

The key is how you interpret "dependant"

For a constant flow which is set by the FCV or some other downstream equipment, then as the inlet pressure rises or falls, the flow through the valve should be constant as long as the valve is quick enough to close or open to maintain the downstream pressure.

Now if the pressure change is rapid then there might be some very small increase or decrease in flow, but given that this is a liquid, the speed of change of pressure to a change in flow will be quite quick. So the pressure and flow might vary by one or two percent or even up to 5% if the valve isn't fast acting. But in essence the flowrate will be fixed = or - a small amount as the valve adjusts its opening to cope with changes in the differential pressure.

I got a bit lost in the rather lengthy post above, but pretty sure we're talking the same thing, just slightly different words no?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

I thought so, but your first paragraph ..made me wonder. In any case, I wrote out the whole process, not because I wanted to lecture you, more so that I 1.) I just wanted to write the process down, so I could have it to refer to later on (I archived it) and to make it easy to spot where we might potentially disagree. I think I am working in a slow motion time frame, where you are in fast forward. Sorry it looks like a lecture. That was not intended. I was thinking of doing a UTube video on this at the time and it turned into a script.

The Pressure regulating valve just regulates pressure. Agreed. .

Flow is controlled by something else. I thought you were talking about the PRN's instaneous flow at any time, which varies the milisecond that it changes position. I see today where you could be talking about the system flow when it eventually stabilizes again. So, yes. In that case, Agreed. At least up to this point..

If that flow stays the same then the flow through the PRV will stay the same, but the pressure difference will increase. We were OK up to there. However you said the dP is increasing, in which case flow does not remain the same. Q across the PRV increases with every increasing increment of dP and the 20b begins climbing too. That continues until sensed and the PRV's position is corrected to hold 20b. At that instant the flow through the PRV is not what it is at the FCV. It is probably a bit higher.

In my time frame .. this is how I was thinking, so consequently I continued to have doubts.

The actual velocity upstream the valve will reduce, At this instant, Upstream flow into the PRV has increased, because the dP increased. The velocity just upstream is a bit higher. OK, could be the time scale thing.

but mass flow through the valve is not dependant on upstream pressure, If incompressible flow, mass flow just increased with the velocity into the PRV. That bit is not affected by time scale. Long term, we have to know how the system inlet pressure and flow will respond. Without a P-Q Curve, its anybody's guess.

but is set by the downstream components. At this instant, the FCV has not yet corrected the flow through the PRV. It may take some time to do that. The flow rate is not yet controlled, per say, although that will happen as deltaT progresses. Another time scale thing, right?

In any case I can still delete all of that if you want. I was really just trying to get the whole process written down to refer to later. I can copy it now to my notes, or am happy to leave it for future reference by others that are not so acquainted with the nitty gritty bits of the process. (if everyone agrees that what I did write is the correct process.) What do you think? Leave it here, or delete it and make a UT Video?

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[GBTorpenhow]

the PRN's instaneous flow at any time, which varies the milisecond that it changes position

Sounds like you are overthinking this. A regulator or any other control valve can modulate as needed to change DP without changing flow. Flow through a regulator is dependent only on downstream demand. In a distribution system the system offtake or consumption sets the demand and may vary, here the demand is constant, set by a FCV.

What exactly is the purpose/function of this arrangement, OP? Is the regulator just for pressure class rating reasons? Are these design or operating pressures? 50 and 20 bar look suspiciously like CL300 and CL150 pressure limits.

 

[1503-44]

Would you agree that is how it works, if we were talking about a 36" control valve with a slowed down gas over hydraulic 6ft long actuator that is going to travel half its length?

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[GBTorpenhow]

Are you asking if a slow regulator is still a regulator?

 

[1503-44]

I'm curious to know why you think I'm overthinking this. Maybe I'll understand better if you can explain to me how a regulator or any other control valve can modulate as needed to change DP without changing flow. IMO, that is not hydraulically possible. The hydraulic equations says this is how it is. Look at their Cv values vs %Open and the equation describing flow and pressure drop across all types of valves and other flow elements, those where dP = (Q/Cv * SG^0.5)^2 Pressure drop is dependent on flow. Flow on dP. A similar equation as a f(Q,dP) can be made for each and every flow element in the system. Pumps need a pump curve. Supplies and sinks need a Q-dP curve, as those items do not necessarily obey dP = (Q/Cv * SG^0.5)^2.

And here.. you say, "Flow through a regulator is dependent only on downstream demand. In a distribution system the system offtake or consumption sets the demand and may vary, here the demand is constant, set by a FCV." You are making the same argument that I thought LittleInch was making, which I already responded to above. You are referring to some ultimate steady state target flow rate , through out the entire system, without recognising that regulator and control valves are not steady state devices. They actually do not function as control devices at all in a perfectly steady state flow condition, other than to provide whatever pressure drop they are contributing to the system in their motionless state. A hand valve could do the same. These things only go into action (changing their %Open) when their control variable setting is breached, ie the system was upset and not in steady state. Then they start moving actuators, changing %Open which changes both their flow and dP in an effort to get the system back into a condition that returns their control variables to their set points. Then they go static when (or rather if) they reach steady state flow again... until the next time an upset to that steady state occurs. Fcv, PCV PRV only function when the steady state runs off the rails. They try to return it to a steady state corresponding to that which simultaneously solves all the individual device dP-Q equations to equal flow in = flow out at every device and the sum of all dP equals the system's total pressure drop and that all control devices are at their set points. If it finds a solution, it's steady state. If not, then it continues indefinitely looking for it.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."