Problem regulating the Flow Control Valve at 0-2% open position

[moonshineJ]

There are two air-operated Flow Control Valves in the system allowing the controlled flow of materials. When one valve is open at 2%, another will be open at 98%, when one valve is open at 30%, another valve will be open at 70% and so on.

We noticed that the valves seem to be letting too much flow through even at the point they crack open. This continues from 0 to 2% of valve opening. Within this initial range flow drastically increases to as much as 8%. Pass the 2% mark flow rate increase becomes linear and remains that way.

The only variable that made a positive change in test performance was lowering inlet pressure. We ran the test with no inlet pressure and it was good, but that isn't how the system is designed to run. We checked the flow through the valves with a pressure gauge on the inlet and the outlet and provided that data to the Flowserve valve.

During site visit, we attempted to run the gradient with a ton of different variables. Even at 0.5 bar though it was still oscillating hard up until 7-8% and the same on the end of the gradient. We adjusted PID parameters, inlet pressures, flow rates, etc. We check valve position via the electrical signal and physically and everything seems to be correct.

What can potentially be wrong? Apparently he problem persisted for long time, but other known good units work fine throughout the entire range. Running the system without having head pressure is not an option; all systems designed to have an inlet pressure of .5-1 Bar

 

[tom baker]

1. What type of valve are you controlling?
2. Can a smaller valve be installed?
3. How is valve position measured?
You may have a valve issue, not a control issue

 

[don_resqcapt19]

I agree with Tom. I have worked on installations where a separate smaller sized valve had to be installed for control at the low end of the flow range.

 

[moonshineJ]

It's a Flowserve Kammer Small Flow series valve
It's rather not about modifying the system (I am sure this can be done!), but rather trying to figure out what would cause this poor-controlled flow at 0-2%. The valve manufacturer's tech reps saying that this was a pure luck, that valves were operating just fine in the past.
I would buy this, but there are other installations with the same valves which still work fine.
The valve position measured fully closed and fully open positions. The attachable mechanical dial would give the position of stem when valve is operating, so it can be correlated to electrical signal sending command to the valve.

 

[RumRunner]

Your inquiry is best addressed in a different forum specializing in INSTRUMENTATION.
There is a forum for Instrumentation Engineers and Technicians.

Members of this trade seem to be a rare breed in these times and place. You don't see many of them around.
I'm not faulting you for being in the wrong place, it's a good try though.
This forum is for installers. . .opinions run rampant and NOT short in active speculations.

These are precision instruments that deal with precision metering that could have serious consequences if not done properly. It involves calibration so precise and the equipment are very expensive.

Are you the same person who posted the same (verbatim) inquiry?

I noticed the same wording as the one you posted.
His name is Ron Andrew.
Library 8917
Here is his post:
Description of the problem:

There are two air-operated Flow Control Valves in the system allowing the controlled flow of materials. When one valve is open at 2%, another will be open at 98%, when one valve is open at 30%, another valve will be open at 70% and so on.
We noticed that the valves seem to be letting too much flow through even at the point they crack open. This continues from 0 to 2% of valve opening. Within this initial range flow drastically increases to as much as 8%. Pass the 2% mark flow rate increase becomes linear and remains that way.
The only variable that made a positive change in test performance was lowering inlet pressure. We ran the test with no inlet pressure and it was good, but that isn’t how the system is designed to run. We checked the flow through the valves with a pressure gauge on the inlet and the outlet and provided that data to the Flowserve valve.
During site visit, we attempted to run the gradient with a ton of different variables. Even at 0.5 bar though it was still oscillating hard up until 7-8% and the same on the end of the gradient. We adjusted PID parameters, inlet pressures, flow rates, etc. We check valve position via the electrical signal and physically and everything seems to be correct.
What can potentially be wrong? Apparently he problem persisted for long time, but other known good units work fine throughout the entire range. Running the system without having head pressure is not an option; all systems designed to have an inlet pressure of .5-1 Bar


Google INSTRUMENTATION FORUM.

 

[MrJLH]

Open the valves up and check the trim and cage. Is a bypass open? Is a bypass leaking through?

I would pull the datasheets and check to see what type of kit is in each valve, i.e. equal percent, linear, quick opening/

 

[synchro]

Are both valves behaving the same below a 2% opening? If both have the problem then that makes it less likely that there's excess stiction or hysteresis in the valve itself that's causing the issue, because then it would need to be to happening in both valves.

The higher slope of the flow response below the 2% opening could be increasing the control loop gain (i.e., process gain) and making the loop unstable. Depending on how the control loop is implemented, perhaps lowering the inlet pressure also lowers the loop gain and is therefore making the loop more stable.

Can the controller be programmed to compensate for the nonlinear response of the valve? In other words, to lower the gain in the region below a 2% opening to make the overall response more linear? If so that might improve loop stability and prevent the oscillation.

 

[Eddie702]

Sounds like the valves are over sized. Proof of this is the fact that you lowered the inlet pressure and the control operation improved

 

[moonshineJ]

Yep, I posted it there, too. I found a few forums around, but some are not as frequently attended by people. This one is one of my favorite ones for plenty of the old school folks having experience in various trades.

 

[moonshineJ]

Open the valves up and check the trim and cage. Is a bypass open? Is a bypass leaking through?

I would pull the datasheets and check to see what type of kit is in each valve, i.e. equal percent, linear, quick opening/

Unfortunately, I was not at the site; I rely on information from other people.
The trim on both valves was found to be in good condition; this is what the manufacturer's techs say. There is one *but* , in their words it's good for 10-90% and they don't guarantee a good operation outside that range. It looks you are right and replacing the trim kit is # 1 thing to do.

Synchro, I will look closer into what gain is capable to do for this particular model.

 

[paulengr]

Flow control valves have problems controlling turndowns that low. Even if they work you get turbulent flow. Any good valve tutorial talking about valve CV should explain it.

 

[Devin Hanes]

What are you using to measure flow?

 

[moonshineJ]

What are you using to measure flow?

It's a coriolis type flowmeter.

Flow control valves have problems controlling turndowns that low. Even if they work you get turbulent flow. Any good valve tutorial talking about valve CV should explain it.

I agree, it's rather a common issue. Nevertheless, whoever setup the system in past made it work even at low opening rate. As I said, the next step will be replacing the trim kit. We'll see how that goes.

 

[drcampbell]

You may benefit from valves with a non-linear relationship between valve position and flow characteristics.

https://insights.globalspec.com/article/10415/characterized-and-v-ball-valves-provide-improved-flow-control

 

[RumRunner]

It's a coriolis type flowmeter.

I agree, it's rather a common issue. Nevertheless, whoever setup the system in past made it work even at low opening rate. As I said, the next step will be replacing the trim kit. We'll see how that goes.

Moonshine

The flowmeter senses the volume of material that is going through the sensing tube. This is done by creating an artificial turbulence caused by the movement of the material passing through.
A series of tubes affected by the turbulence, . . hence, using this coriolis effect.
You probably know this already. . . this is just a rehash to elucidate this function in terms of using it to control the valve.

Since you are using this feedback signal (Coriolis Effect) for valve positioning--it would be worth mentioning that the variation of the valve position is dependent on the output of the material passing through the system.

It senses standstill (no-flow) to a high-flow as a result of the full-bore opening of the valve.
These are used in nuclear power plants, ships, and oil extraction and refineries.

Having said that--both the valve and the sensing device (Coriolis Effect) have to work in concert with one another. . . . and this requires (precise) calibration.

This requisite becomes more important when you are controlling this system remotely.
Using network for remote control also requires established algorithm to accomplish the desired outcome.

Compare the functions of the working units to this one that you're working on, and see the different settings.

Are you using MODBUS, BACNet or similar network protocol for PLC interoperability?

 

[Russs57]

Given that pretty much nothing is perfectly linear from 0-100%.......is it possible to select a pair of valves that will operate over something like 20-80%?

It isn't uncommon to have a two valve solution with one small valve bypassing a larger valve at low flow rates. Maybe with different spring rates. Sometimes piping in series works best, but usually for pressure, not volume control.

You make no mention of flow rates or the media being controlled.

If it worked in the past....maybe you have some seat/plug wear.....maybe spring has become fatigued?

 

[paulengr]

Valves have a maximum turndown ratio which depends on flow rates, pressures, and valve design (to some degree). The degree of difficulty depends on a valve rating called CV and the conditions. At some point the flow becomes turbulent and uncontrollable by any valve. When this happens flow is completely uncontrollable or you get wicked things like cavitation eroding the valve or the downstream piping.

The practical solutions are smaller valves and decreased pressure or flow drop. As mentioned it’s not unusual to have say a large and small valve work in tandem so the large valve opens and closes only at higher flow rates or some other similar combination. Generally you never want to try to control a valve outside the range of 20-80%.

 

[gadfly56]

Valves have a maximum turndown ratio which depends on flow rates, pressures, and valve design (to some degree). The degree of difficulty depends on a valve rating called CV and the conditions. At some point the flow becomes turbulent and uncontrollable by any valve. When this happens flow is completely uncontrollable or you get wicked things like cavitation eroding the valve or the downstream piping.

The practical solutions are smaller valves and decreased pressure or flow drop. As mentioned it’s not unusual to have say a large and small valve work in tandem so the large valve opens and closes only at higher flow rates or some other similar combination. Generally you never want to try to control a valve outside the range of 20-80%.

What you said. My original career was in chemical engineering. It is very rare to find a valve that has a better than 10:1 turndown ratio. One of the other problems is hysteresis. When a valve moves from 20% to 50%, it doesn't necessarily give the same flow as moving from 70% to 50%. This issue is exacerbated when dealing with valve movement at the extremes of the valve range.

 

[moonshineJ]

Are you using MODBUS, BACNet or similar network protocol for PLC interoperability?

Myspark, in this case it's Profibus. Thank you for the insight on coriolis flowmeter operation.

Generally you never want to try to control a valve outside the range of 20-80%.

This is the valid point. It's why cars have different gears, it would be extremely difficult to get all speed range using just one set. Assuming that the valves have been operating ok thru the entire range at beginning, I would do the trim kit replacement first, on each valve.
While it seems to be not a whole lot of issue for a single valve operation, the most problems happen when both valves work together to achieve gradient. This is when they start affecting each other. We will see if the trim kit replacement will work.

 

[Devin Hanes]

some things to check; valve flow direction, Coriolis tube orientation for the media involved, Coriolis drive gain and density reading(bouncing) to check if air is affecting measurement, some of the newer micro motion meters have a zero verification you can run, check cal factor is entered correctly for Coriolis.