Control Valve Blocked In Gas Service

[LeonInst]

Hi. I'm new here, so my first question concerns a problem we've had with a control valve. First, a few details:

Valve is used in a spillback system on a Refinery Crude Unit. The valve is on a line from the discharge vessel (we'll call it vessel 3, operating pressure 6-7kg/cm) of a compressor back to the suction vessel (Vessel 2, operating pressure approx 1kg/cm). The valve is controlled from a pressure transmitter (via Honeywell TDC & 4-20mA o/p, then i/p) on another vessel ('vessel 1') which is linked to vessel 2 with a common line.

When the pressure drops too low on vessel 1 (and equally low on the vessel 2) the valve should open to recirculate the gas and thus increase the pressure in vessels 1&2.

The product is C3/C4s. Uncertain of the temperature its at, but it makes the pipework warm to touch, so obviously fairly hot (will post temp when I've checked).

The valve is a 3" Blakeborough BV(500 or 900) globe valve, =% plug, air fail closed, air on top of actuator to open (spring forces valve up to close).

The issue with the valve is that operations have been claiming that the valve will not control and believe there is no flow through the line. The valve is awkward to check as it is in split-range control with a second valve which opens to flare if vessel 1 pressure rises too high, so operations are uncomfortable to allow tests without full written procedures etc. However I am certain that the valve stem is moving, so I thought it was possible the stem was snapped inside or valve plug jammed.

Thermography was carried out on the valve and put the temperature of the body upstream at 27 deg C and downstream at 11 deg C, although I suspect the upstream body temp is higher than that.

Process engineers then had both gate block valves and the control valve body x-rayed. This shows that the block valves are both clear. It also more importantly shows that the valve plug is not seated in the valve, so I believe that it is physically open.

When the bypass valve was opened I have been told (although as I did not witness this I do not know how far open the valve was) the downstream temperature increase.

Has anyone got any suggestions as to what could be the issue? I suspect that the difference in temperature is due to the Joule-Thomson effect, and that opening the bypass valve (which I believe is a gate valve?) the D.P. was reduced so the temperature changed stopped.

However I cannot figure out why the valve is not having any control. Is it possible the gas is cooling to the point of icing up around the plug, reducing flow?!? Or could the wrong plug type be causing this? Any suggestions?

Removing the valve for inspection sounds obvious, but it is positioned awkwardly with little access to use a crane. I'm thinking about trying to block it in and drop the bottom flange off the valve to check inside, but again this will require a mountain of paperwork to do.

 

[hacksaw]

no control means that the valve is not accomplishing what it takes to operate the plant.

could be sizing, i.e. flow requirements greater than the valve was sized for, etc.

split ranging the flare valve with the spillback valve is a bit worrying. While it should not be a problem, I'd feel better if it had its own controller and setpoint.

 

[LeonInst]

By no control I mean the valve position seems to have no effect on the pressure in the vessel. The pressure in Vessel 1 is currently just relying on the pressure upstream to stay high enough at present. The process engineers continue to insist that the valve cannot be opening because there appears to be 'no control'. I've checked the line personally in the field and there are no block valves closed to prevent it functioning.

I'm wondering whether to fit pressure gauges to each of the vent/drain valves each side of the valve to check, and if the D.P. across the valve changes depending on valve position then I will be able to rule out the valve being closed.

Vessel 1, 2 and 3 all have PSVs and I believe Vessel 1 also has a high Pressure shutdown on it, but I agree different loops would have been better. Unfortunately I was not at the Refinery when it was installed. The other drawback is that at present, both I/Ps have been wired in series, so I cannot even break the current loop or the flare valve will be unusable. Its poor design, but I don't think its whats causing the valve not to control.

I would have thought that the valve would still have some effect if it was undersized, unless the original design group got it drastically wrong (not impossible...)

 

[JimCasey]

I don't think you have a broken stem because I believe you would get full flow most of the time.

I scrolled thru the Blakeborough catalog.

Not every manufacturer shows everything in the catalog, but the way I read it the 500 and 900 only work in the pull-up-to-open mode. The 800 is the top-and-bottom-guided, invertible body that <might> be push down to open. So if you are sending the valve an increasing air signal to open, you may be closing it. Obviously then your control would be "challenged". Just when you are calling for a lot of flow, you get bupkus. The "quick" fix would be reversing the positioner action, but the proper repair would be replacing the actuator with a spring-to-push-down design.

If it is an 800, there are top and bottom flanges on the body. Push-down-to-open would likely then be accompanied with at least inverted lettering on the casting. You can change an invertible-body valve from direct to reverse by flipping the body and screwing the valve stem into the other end of the valve plug. You'll need gaskets, and whatever locking device they use on the stem-to-plug connection. ( I have seen pins used)

It's something to check for.....

 

[LeonInst]

Sorry, I've just looked at the valve catalogue and I'm pretty sure it is a Blakeborough BV803. It is top and bottom guided (one suggestion was that there may be debris in the bottom guide preventing it from moving down) and it has a 'v'-port plug.

You're right in saying that if the stem was to have snapped the valve would have fallen wide open and allowed full flow.

The positioner (Fisher 3560 I think, so a force balance positioner using a flapper/nozzle and bellows arrangement) is set to reverse action, so with the output from the local current-to-pressure transducer giving an output of 3-9psi (split-range), the positioner has been set to give fully open at 3 psi input (high output from positioner), and fully closed at 9 psi input (low output from positioner).

The spring range on the valve is stated to be 6-18psi, and when I checked the valve on Thursday (I haven't been in work today, this is being written from home) the input to the positioner was 6 psi, and the positioner output was 11 to 12 psi. The valve stem appeared to be about 50% travel according to the plate (not a reliable guide, but its an indication of sorts).

So I think the valve is working, but something is not quite right in the system, or the valve is wrongly specced for the service...

 

[JimCasey]

When one reverses a positioner's action, it is also necessary to reverse the pneumatic piping between the positioner outlet and the actuator. Otherwise you'll just get snap-acting on-off response. I generally caution against reverse acting positioners (in this sense: zero input, full output) because the failsafe mode is not consistent. Lose air, and the valve goes one way, but lose signal and the valve goes full travel the other. One of those positions is likely not safe for an un-controlled valve.

Split-ranging is hard enough for Joe Average maintenance tech to deal with, but split ranging AND reverse action is just evidence that somebody is outsmarting himself.

 

[itdepends]

Let me see if I've got this straight.

Vessel 2-->compressor-->vessel 3

Vessel 3-->vessel 2 via spill back line with this (faulty?) valve in it

Vessel2--> connected to Vessel 1 via common line?

So your trying to increase the pressure in vessel 2 by recirculating the compressor discharge back to it's suction?

That's like trying to increase the pressure in a suction tank by using the discharge pump to pressurise it up- it dosn't work. All that happens is that the flow in the spill back line is dictated by the pressure drop between the pump (or in your case compressor) discharge through the spill back line and into the suction tank/vessel.

The pressure in vessel 2 is set by the pressure drop in the downstream equipment (i.e. the pressure in vessel 3).

Perhaps I've misinterpreted your description- in which case a diagram (better than my poor single liners) may help.

Cheers