Custody Transfer Meter Skid Back Pressure Control Valve Issues 2

[Tom Connor]

In 40years of experience I've never considered (or seen) a CTM back pressure control valve used to mitigate a surge event. My question, is there anyone out there that have or would consider using a CTM back pressure control valve to mitigate s surge. I've real issues with this and question the validity of doing so.

Tom Connor, PE
Operational Fluid Mechanics

 

[LittleInch]

Rather unusual I would agree.

Where is the surge coming from? Upstream or downstream?

A control valve in these circumstances is presumably because otherwise your meter skid has too low a pressure?

Why else is the control valve there?

What product?

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

 

[Tom Connor]

The surge is coming from downstream, the product is crude oil, the client is proposing to put limit stops on the back pressure control valve so it does not fully close, but closes ahead of the surge wave thereby sending a rarefaction wave downstream reducing the surge impact. My position is that a meter skid is not designed to mitigate a surge and a proper surge relief system needs to be installed.

Tom Connor, PE
Operational Fluid Mechanics

 

[LittleInch]

I think you're correct. The only thing that would really set off a reflection wave is a non return valve. I don't understand why the valve would fully close anyway, but we can't see or understand your system operation.

Meters, especially CTM, really don't like going backwards, even slightly.

I agree - fix the surge issue, don't mess with a CTM.

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

 

[Tom Connor]

The pressure wave is coming from a closing butterfly valve and the only valve faster is the CTM backpressure valve, the limit stops are to prevent the CTM valve from completely closing and setting off a secondary pressure wave. The CTM back pressure valve would be commanded to close by automation on the movement of the BTF valve (without a command) and limited by a mechanical stop at a point where the secondary pressure wave would not increase above the MASP of the system which is ANSI 150. So far I've refused to sign off on this. Client is pushing the meter skid designers (already delivered but not yet in operation) to approve this approach. Should they accept it and assume responsibility for operability of the CTM then I'm likely obligated to accept the manufacturers recommendation. Professionally I would never design such a system to mitigate a surge.

Tom Connor, PE
Operational Fluid Mechanics

 

[georgeverghese]

If you ask me, I would agree with the client position on this - I've done the same / similar before for emergency shutdown valves. In some cases, a tamper proof travel stop can be used, or in other cases, a tamper proof closing speed controller. I'd rather spend some time to ensure this travel stop cannot be tampered with and that operating procedures clearly state that a PtW is required for any work on this backpressure valve.
Not clear why you have this backpressure valve on this CTM when the butterfly valve is downstream of the CTM - it would make more sense,from a min backpressure point of view, if the BV were to be upstream of the CTM.

 

[Tom Connor]

The BTF is 200-800 feet away in a tank manifold that feeds several tanks. There is an MOV at the tank that could close. Meter skids require a back pressure valve to prevent flashing in the meter. The BTF is for cutting off a manifold run to specific tanks and not designed for holding back pressure on the meter skid.

Tom Connor, PE
Operational Fluid Mechanics

 

[1503-44]

"Professionally I would never design such a system to mitigate a surge."

It appears that you already know what to do.

Since when has been permissible not to have a check valve immediately downstream of a custody tfx meter? Makes no sense to meter custody if you allow any backflow. Thats also most likely why the downstream surge is reaching the meters.


I also question the need for the backpressure control valve to act so fast. Slow down the manifold valves and eliminate surges backing into the meters. If you cant do that, build your surge control system. I dont think your meter designers will go for a system as designed. Whats in it for them to blow up their own meters. Hopefully ther are smart guys and will shut down the nonsense before it goes on for too much longer.

 

[Tom Connor]

The meter skid does have a check valve immediately on the skid at the exit of the skid. The surge is too fast (3500 ft/sec) thru the system and doesn't close the check valve (which is good otherwise the pressure would jump from 35-40 psi to about 500 psi. Client wants a mechanical stop on the CTM PCV, closing the valve (by automation) on any downstream MOV movement (5 sec delay). The CTM PCV valve is faster that any downstream MOV. However not completely closed as it will also create a rather large surge uptream, the answer is an upstream PCV at the ANSI 600//150 spec break that closes faster than the CTM PCV and raises the pressure in the 600# class system which can take it up to 1628 psi or a 150# class surge relief system (client is against this because of the cost) immediately upstream of the CTM system.

Tom Connor, PE
Operational Fluid Mechanics

 

[LittleInch]

These actions all sound like sticking plasters on a gaping wound.

The issue is surge pressure from closing the downstream valves too fast / not being able to slow down the flow.

These lines must be running at a fair velocity to generate 500 psi from a normal 30-40??

I think you are correct - the control valve on the meter shouldn't be trying to do this separate function and I still don't really understand why it needs to close super fast on downstream closure in the first place. It's function is to maintain pressure in the metering skid. Why is it closing?

I'm not so sure your check valve won't function in some way or other in the surge scenario. Maybe not the initial wave but certainly subsequent ones.

How is this being analysed?

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

 

[georgeverghese]

The fog clears a little on this. Looks like the root defect here is insufficent mechanical design pressure - the entire system should have been 1628psi right up to the MOV at each tank, including the CTM skid. The backpressure CV is being forced by control logic to close to prevent a buildup in pressure downstream of the PCV, at the time when the tank MOV begins to close. In normal operation, the backpressure CV holds a min pressure at the CTM to keep the cude oil in single phase liquid condition.
So, putting a min stop on the backpressure control valve will now cause a new problem - overpressure on the downstream manifold. I see why you're asking for this unwieldy quick acting surge relief (which really should be blocked in full flow relief), or this even quicker acting PCV upstream of the CTM to prevent overpressure( whose control response action integrity may be dubious).
Its getting a little scary now - too late (and painful) to go back to the drawing board and start anew on this?
A simpler solution, half baked as it may be, would be to interlock the source pump speed with this closing of any tank MOV - 5seconds after the MOV starts to close, drop its speed down temporarily, so that its developed dead head pressure is within the CTM skid / tank manifold mechanical design pressure limit?

 

[1503-44]

The answer to surge is seldom two control valves. They both fight themselves rather than control surge pressures. I have seen 100 psig go to 1750 when marine tanker closed their valves too quickly when departing the loading bouy in the face of a severe winds and thunderstorm. Fortunately the surge system at the onshore meter station (where I was) worked and filled the surge tank, but not before the marine tankers check valve broke and manifold burst spilling 5000 bbls over the deck and into the bay.

What kind of facility is this and what scenario is driving the need for the valves closing so quickly? I cant see, for example, why a valve closure at a typical onshore tank farm needs to be so fast as to create such surges. Basic pipeline control is always, at least partially, open a second conduit before slowly closing the first. Complete shut in of any kind is going to run up those pressure gages one way or another. No surge system often leads to ANSI 900 everywhere.

If you have to cave in, just be sure to get your client's signiture documenting his override of any of your recommendations. Write one up for his consideration ensuring that all responsibility for the bad design based on cost savings lands squarely on him.

 

[1503-44]

If your pump is close enough to be able to control pressure quickly enough at the meters (probably not a pipeline terminal) that just might work.

 

[Tom Connor]

This system is a multitude of storage facilities (tanks feed by 3rd parties anywhere from a few thousand feet to a mile away. Each facility has an incoming pig receiver and most have a PCV at the receiver just inside the facility fence and (in most cases with a few exceptions) also serves as the spec break 600#/150# with the sole purpose to limit the pressure into the facility under 2 conditions; a) when delivering to tanks the PCV holds the discharge pressure to 100 psig or b) if delivering to the suction side of facility pumps the pressure control would allow up to 150 psig on the discharge side of the PCV. CTM PCV's will reduce the pressure to what is required to enter a full or empty tank.

MOV's downstream of the CTM system are BTF valves with a closing speed of 5sec per diameter (12" valve speed = 60 seconds). The tank MOV's are mostly 30" gate valves (30" valve speed = 150 seconds). Distance from CTM to manifold varies from 500-1000ft and distance from manifold to tanks 200-1000 ft depending on the facility. Most of all the velocity way exceed our recommendations so were stuck with 12-14 ft./sec inside the facility which creates a major mitigation exercise with wave speeds around 3500-3700 ft./sec. Automation can intercede the movement of valves and somewhat control other upstream valve closures. We do not agree with mechanical limit stops for a multitude of reasons one being the thrust loads that are going to be realized when the CTM PCV closes attempting to mitigate the surge and the other is CTM's are not designed nor intended to function as a surge mitigation tool. I'm even wondering if DOT or the Texas RR commission may have restrictions related to this issue.

On most all of the incoming lines the client does have a communication link with the 3rd party but does not have authority to shut them down or control equipment it's strictly left to them and that's part of the reason the incoming lines have a spec break PCV at the facility fence so the 3rd party connection can be terminated by the client. We're proposing this valve should be made faster, is always controlling and not wide open and should be the 1st line of defense when necessary or surge mitigation controlled by either surge relief valves or bladder tanks (if the volume release is low.



Tom Connor, PE
Operational Fluid Mechanics

 

[hacksaw]

Don't confuse your accoustic wave velocity with fluid surge velocity.

 

[LittleInch]

Sounds to me like the incoming PCV class 600/150 should start to ramp down and aim to be closed before the tank MOvs to gradually slow the crude oil down. Sixty seconds would cover most eventualities and at least a control valve will operate in a linear fashion.

Interesting to note that a butterfly valve is surprisingly steep Cv drop in the first 1/3 of the closure so you would expect the end part to more gradually slow down the flow.

That's some velocity you have there so far from surprised you're getting surge issues.

what transient software are you using for this? Is it properly tuned and using the right Cv versus opening scenarios or are these pressures real ones you are actually experiencing?

Anyway I still agree that using the CTM CV is not the way to go.


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

 

[1503-44]

Pressure waves do travel at acoustic velocity of the fluid although total pressure buildup at any given spot does not typically happen on the first reflection, each reflection builds up pressure. Typically the first wave might reach 1/2 of total, the 2nd wave adds another 1/4 or so, the 3rd adds an 1/8 of the total surge pressure. After the 4 wave, pressure build is pretty much finished. In short lengths of pipe, like this station has, and with that accoustic velocity, which is pretty high (faster than water is, if I recall correctly). Crude should be around 2500 to 2750 fps, again if I recall correctly, falling to 22-2300 when pipe modulus is taken into consideration). you might want to check into the velocity calculation. They might be using the wrong coefficient, but in any case with those short lengths of pipe, whatever velocity you use won't make much difference. Pressure builds will be quick no matter what. 5-6 fps is a better range for tank farms. Too late to do much about that.

I think that PCVs should never be designed to actally close, if they actually close then they certainly are not regulating pressure. Normally I size CVs to operate between 30% to 80% open. That keeps you out of the smaller %open ranges that can have exaggurated effects on rapid velocity and pressure changes when the valves are nearing full closure. With Cvs near closing, upstream perssures often rise and negate the effect flatter Cv-%Open slopes, especially with surges going on anyway.

Pump control might be effective, but how do you control those 3rd party pumpers?
Maybe the meter mfg is your best hope to solve this quickly.

TXRRC, ASME B31.3, 4 and 8 or CFR 49 Part 192 give any requirements in this regard, outside their general limitation of surge overpressures not exceeding 110% of MAOP and (I think) the number of hours per year that they can be within the MAOP-110% MAOP range. I prefer to design systems to be actively reducing pressure whenever it goes over MAOP, rather than to try to keep operation during overpressures limited to a certain number of hours/year.

 

[Tom Connor]

The computer program is DNVL's Synergy Pipeline Simulator, alias Stoner Pipeline simulator, alias Drem Pipeline Simulator, alias SSI Intercomp TCON. I've been using it since 1974 and a Senior Specialist using this technology, At one time the program was the only DOT approved software for pipeline dynamics. Yes it tuned. I agree velocities should be 5-8, client ignored recommendations and velocity is 12-14 ft/sec. Bulk modulus is approximately 158000, surge velocity is 3700 ft/sec and declines over time. The program accounts for the change in wave speed according to the dynamics and wave speed is not fixed.

Tom Connor, PE
Operational Fluid Mechanics

 

[georgeverghese]

If you must retain the client proposed surge mitigation concept , suggestions would be
a) Slow down the closing speed of the CTM backpressure valve as it goes down to the min stop position, in order to minimise surge at the CTM skid and upstream kit - this also then gives more response time for the upstream 600#/150# PCV to keep its setpoint.
b)Cannot imagine why you need operating pressures ranging up to 600lb piping class upstream of the receiver side PCV for this tank filling operation; this must be stabilised crude oil with a low RVP; - as you say, this control parameter is within your client's control, but they should advise you what they plan to do here to reduce risk.
c)Slow down the tank MOV only for the last 30-40% of the closing stroke - this should reduce surge pressure further.

 

[LittleInch]

I realise this post has morphed into a discussion on surge, but I'm seriously struggling to see where you're getting high pressures at those closing speeds. On the reflection wave coming back to the original tank valve surely it just mainly passes through into the tank??

In these sorts of situations for me it's only the snap shut ESD / hose disconnect slam shuts that usually causes an issue.

Closing your control valve faster than that must make it worse no?

How is the loading into a tank supposed to slow down and stop in "normal" operation?

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