If you have done any trim character

[4carats]

If you have done any trim characterization in control valves operating in parallel for high pressure steam letdown, please share your experience with 4carats. Thank you.

 

[4carats]

This is not the answer to my own question above. Our Company has a serious problem in Steam Letdown Valves. To capture the problem in more details: Three Steam Letdown Valves, A, B & C, are arragned in parallel, each has an upstream pressure of 900 psig, and an outlet pressure of 600 psig. The "A" valve is a small valve having a flow of 17 kg/sec maximum. The "B" & "C" valves are big valves having a flow of 68 kg/sec. Could someone tell me how to characterize the trim of the big valves so that a smooth transition can be made mechanically when control passes from the small valve to the big valves (either B or C) or vice versa? Thank you in advance.

 

[TD2K]

What is the trim characteristic of your big valves?

Unfortunately, in my experience, control is never good where the second large valve is starting to open or close and it can be worse depending on the type of trim you have. If you have to run right at the point where one valve is just barely open, it's never going to give you good control.

As soon as the controller starts to open (or close) the big valve, you are going to get a lot more flow relative to what an equivalent movement on the small valve from your controller has been giving you. That's hard for a controller to handle. The problem is even worse with an equal percentage trim as at that end of its curve, small movements in valve opening gives you very large changes in the valve's Cv.

Do you have one controller controlling all three valves? You might want to put a separate controller on the big valves and use different tuning parameters. On pressure control like this, lots of gain will give you good pressure control but at the expense of lots of valve movement, you'll need to come to a happy medium what is 'best' overall.

 

[4carats]

Thank you, TD2K.

Each of the three valves has their own controller. I advised my Company to change the trim of the big valves except I have no experience in how characterization of the trim is done. I know how much flow overlap is needed; I need to know how to translate the desired flow vs lift characteristic curve into the trim design in order to write a good specification such as what resolution should I specify, the definition of resolution, acceptance test etc. Could you help me?

 

[TD2K]

I think I may have mislead you with my comment on equal percentage trims, please ignore that for the moment, I need to think that over.

Back to one comment of yours:
"I need to know how to translate the desired flow vs lift characteristic curve into the trim design in order to write a good specification such as what resolution should I specify, the definition of resolution, acceptance test etc."

I don't understand what you are asking for here? "resolution I should specify", "definition of resolution" are all terms/phrases I'm not sure what you are trying to say.

With a steam let-down system, you obviously have a fairly wide range of flow rates that doesn't fit in one valve and as a result, you've gone to three of them. At the point where the second (and to a lessor extent) the third valve starts to open, control is going to be a problem. You can't do much about that any more than you can 'fix' poor valve control when its bouncing off the seat.

Hopefully, the valves have been specified considering the sites loads in mind such that you don't have to operate with the small valve is essentially wide open and the large valve is a couple of % open. If you do, one option would be to close the small valve and push all the load onto the big valve and get the opening up and into a better control range (your operators will quickly decide to simply put the small valve in manual and close it). Without understanding how your system is expected to work, it's difficult to be more specific.

 

[saxon]

4carats, If I'm reading your note correctly, what you have is multiple control va. set trying to maintain a constant steam pressure around 600 psi. This is a normal type of set up for wide ranging flow requirements from HI to Low flow requirements that are outside of the range of a single valve. The proper control scheme is known as "Split Ranging". As to your set up where each control valve has it's own controller is a little odd. You can set up the control scheme with one controller. Since you have three valves, you will need to split the output control signal into three bands. If the signal is a standard 4-20 ma. the signal will be broken into three control bands, 4-9.3ma, 9.3-14.6ma, 14.6-20ma. Valve A (small) will operate from 0%-100% open on the 4-9.3 signal. Valve B will operate 0%-100% open on the 9.3-14.6 signal, Valve C will operate 0%-100% open on the 14.6-20 signal.

As for the valve trim characteristics, equal percent. is normally used on fast processs like pressure control but in split ranging this can cause problems and a linear charc. works better.

For full and complete discussion on valve characteristics and application and the split rangeing control scheme, see the "Control Valve Handbook" by Fischer Controls. Also see Chem. Eng. Reprints "Practical Process Instrumentation & Control" by McGraw Hill.

Hope this helps.
saxon

 

[hacksaw]

Ditto the split ranging comment. You start opening the small valve first when it approaches end of travel start picking up the next valve and so on... the process happens in reverse as you ramp the valves closed.

some overlap is desirable so as to smooth out the overall flow characteristic.

your split ranging can be done in the dcs as an algorithm or with separate hard wired controllers by proper adjustment of the valve positioners.

 

[4carats]

Hi TD2K,

By resolution I mean for every unit change in stroke, what is the change in downstream transient pressure. I am talking about valves of size 12 inches and bigger. The hydraulic forces of the fluid (and enthalpy change) are so huge that the valve(s) are vibrating. What resolution can you specify when re-characterizing the trim? What criteria do you use as acceptance test?

4carats

 

[TD2K]

There's a lot of different issues coming up. Vibration of your valves is different from getting good control as you transistion from one valve to another valve from a control point. Is the vibration always a problem when the valves are in service or only at specific points such as when they are starting to open? What is happening to the downstream pressure at this time? Vibration is typically an indication of excessive velocities in your piping and/or inadequate mechanical support (by the way, there is NO enthalphy change across a control valve, it's an isenthalpic process).

You said you were writing a specification for this problem. Who is this being sent to? A consulting company? A valve company? Are you asking for a dynamic simulation of your system with these valves?
A dynamic simulation is about the only way I know how to quantify 'for every unit change in stroke, what is the change in downstream transient pressure'. Be cautious if you go this route and know exactly what you want an outside company to do. There are a LOT of variables to consider. If you dump a vague request on a consulting firm, they'll be more than happy to crunch through every case under the sun and hand you the bill for this.

 

[4carats]

TD2k
Thank you.
The positioner linkage broke several times due to vibrations. The lantern rings (stainless steel cylinders 2 inches tall X 1 inch diameter) within the valve --- one disappeared; one was worn down to 1/2" inches due to vibration. Vibrations affected control. Downstream desuperheater outlet temperature trips. Desuperheater outlet temperature loops have never been tuned and could not be tuned successfully. As big valves open, more steam flow, more heat, more spray water required --- enthaply does not change, total heat increases. Characteristics of small valve does not match with those of the big valves....DCS Control is the brain; control valves are the muscles. The brain can do it; the muscles cannot do it....

 

[saxon]

4carats, Now the facts begin to emerge out of the muddy waters. No wonder control has been lost, you can't control the process when the final control elements have failed. This failure has resulted in overloading the Desuperheater causing the Temp. Ind/Control loop to alarm out. Kinda of like a set of dominoes. Go back to the valves and get them repaired according to specification, check the program in the DCS to make sure that its is properly setup for split range control and that the proper Proportional/Integral/derivative algorithms have been used and are properly tuned to the process to give proper control. Then and only then, you can get around to properly tuning the Desuperheater Temp. control loop. As for the vibration problem if it is still there after getting the control loops straightened out and operating correctly, Pipe Stress/flexibility analysis will need to be redone, and additional supports and anchors installed as required.

Hope this helps.
saxon

 

[4carats]

Saxon,
Thank you.
The two big valves were upsized in 1984 for 850,000 lbm/hr of steam flow each. [The energy from one valve is big enough to supply electricity for a city of population of 50,000 people.] The upsize was done independent of the small valve, i.e. the small valve has been left unchanged. During commissioning (before I joined the Company), the big valves caused the piping (24" chrome) to shake; the safety valves lifted. Immediately, people placed mechanical stops on the two big valves to limit the flow back to their orginal flow of 400k lbm/hr each (and stayed there up to the present moment). The valves did not pass commissioning test; they failed immediately after newly installed. This is the 900#/600# Steam Letdown Stations, A, B, & C. Meanwhile similar things were done downstream in the 600#/50# Steam Letdown Stations (big valves B & C upsized, and immediately mechanical stops have to be installed similarly to limit their flows back to the original value of 400k lbm/hr each. Even the earth-quake arrester broke away from the anchoring pedestal.) Control logic have been improving for the last 15 years trying to make the control valves work. The logic works great at small flow in the absence of upsets; the muscles cannot handle the huge amount of energy when required suddenly. Should the control valve characteristic be the first step in solving this problem? Thanks in advance.