29 replies to this topic

[sheiko]

Hello people,

In the process of commissioning a nitrogen supply line (see attached isometric), i am facing the following problem:

When opening the pressure control valve (PV 112 25) to allow nitrogen to flow from the pressure regulators (PCV 112 17/19), the pressure indicated on the manometers PG 112 55/56 becomes erratic (pressure varies from min to max scale intempestively) and the pilot-operated pressure safety valves (PSV 112 11/12) start discharging (even chaterring). The whole system starts then to make a strange noise as if the PV and PCV were "pumping" (yeah i know its kind of vague description but my auditive memory is not that good...).

I believe the problem does not come from the PCV because when the PV is closed, the PG indicate the good value: 50 mbarg (the PCV setpoint. See process data below). In this case the PSV remain closed.

My opinion is that there is not enough distance (straight) and too much elbows between the PCV and rhe PSV, what would promote turbulence and, thus, cause the pressure to become erratic (i have not performed hydraulic calcs yet to make thiigs clear). Any other ideas? Have you ever experienced such a situation?

I was also wondering if you had seen such a configuration (pressure regulator + pressure control valve) before (i am asking because i was not part of the Engineering team of the project who decided such an arrangement...)?

Process data:
Fluid is Nitrogen (MW = 28)
Normal Flowrate is 15 Nm3/h (0°C et 1 barg) and operating Temperature is 15°C
Design pressure upstream PCV is 15 barg
PCV reduce pressure from 10 barg to 50 mbarg (CV = 2.2 in US units)
PSV set pressure is 80 mbarg (sized for PCV failure scenario: 335 Nm3/h, inlet line pressure drop < 40 mbard and total backpressure is < 7 mbard)
PV inlet pressure = 50 mbarg
PV outlet pressure = 12 mbarg

Thank you in advance for your help.

Edited by sheiko, 01 October 2010 - 02:15 PM.

[Zauberberg]

Sheiko,

Interesting post. However I keep failing to read through the attached Isometrics - simply can't grasp the piping/valve arrangement since the picture I got is quite blurred. Can you make a quick hand sketch?

Regards,

[sheiko]

Attached is a sketch. Hope it is clearer now.

Edited by sheiko, 02 October 2010 - 06:04 AM.

[Art Montemayor]

Sheiko:

Not only is the drawing bad in reproducing, but the Pressure Control Valves (PCVs) are not even shown – just the inlet and outlet flanges. Additionally, there is excessive convolutions of the piping layout – as if it is located out in a very cramped space – maybe an offshore platform(?). Some line sizes are shown – but are difficult to pickup. Basically, you are regulating nitrogen flow from 10 barg down to 12 mbarg (0.012 barg). You are reducing the pressure almost a thousand-fold. Guess what????? YOU ARE CREATING CHOKED FLOW.

All the convolutions and the choked flow mean the pipe size is critical in order to suspect vibration and movement – as well as the piping anchors. I have seen such a configuration (pressure regulator + pressure control valve) before – many, many times. But not configured or designed like what you show. When you are dealing with gases you are constantly going to come up to a choked condition, so if you are having to reduce the source pressure you have to do it in stages – usually with a special control valve (a “regulator”) or several in series.

You haven’t stated exactly what you are trying to do – you just told us you have a flow and it goes from left to right. You need to give us ALL the facts and story. For example,

• What, exactly is this nitrogen doing and how does it flow – continuously, or sporadically?
• Do you really require 335 Nm3/h of N2 to flow, when the design flow is for only 15 Nm3/h? (and by the way, “Normal” conditions in Europe are measured at 1 atm and 0 oC – not at 1 barg; is this a NEW normal condition?)
• Where is the N2 going?
• A lot of the design could be circumvented by using fixed, resistance orifices to cascade the pressure down.
• Why is the piping so complex?
• Why are there 2 PCVs in parallel, when only one could do the same job?
• If 2 PCVs can be justified, then why not 2 PVs also?

I can’t criticize the design without knowing the scope of work and all the basic data; however, I believe I can question it. If if it is what I suspect it is, then the design could be a lot simpler and more fail-proof. With sonic flow being provoked, you will generate some noise and movement if the piping is not properly designed and is intermittant.

[sheiko]

Dear Art thanks.

To answer your questions:

1/ Basic scope:
The nitrogen is used to inert and isolate the interwall of a LNG storage tank. It is a continuous process.
The project is a revamp.

2/ PCV configuration
I don't know the "why" of this configuration in parallel.
As for the complexity of the system, i still don't know but i am investigating. Having been on site, i believe it is exagerated. No need that much convolutions.

3/ Choked condition
Thanks for remind me such an important phenomenon.

4/ Normal conditions
You are right (1 atm).

Please let me know if you need more info to understand my issues.

Thanks

Edited by sheiko, 02 October 2010 - 06:06 AM.

[Zauberberg]

If I got your statement correctly (and thanks for uploading the P&ID), there is no stable pressure control once when you open the PV-112-25. And subsequently, the inflow of Nitrogen through PCV-112-17/19 is such that it causes piping downstream of these two regulator valves to overpressure?

Well, in a properly designed systems whatever comes in - must go out. I would try to open PV-112-25 slowly in manual mode, and watch how the system responds. I can imagine four scenarios from this point:

1. Controller improperly tuned, i.e. does not react fast and accurate enough when in AUTO mode, based on Nitrogen flow demand changes;

2. PV/PCV valves undersized or oversized, resulting in poor control. You will notice this by capturing valve % opening when in automatic mode - there shouldn't be drastic changes in % opening during operation;

3. Huge variations in the Nitrogen flow demand, in the network downstream of the PV valve. This would be equivalent to improper controller tuning or inadequately sized PV valve, manifested by frequent and large changes of the PV valve % opening.

4. The surge volume between PCV's and the PV is very small, resulting that - whenever there is a small change in Nitrogen demand, the interaction (and the speed of interaction) between these two valves (PV/PCV's) is such that it cannot provide smooth and continuous control.

Can you give us further comments on this?

[sheiko]

Zauber thanks,

Item 1: What i have described above is what is happening with PV in AUTO (local pneumatic controller) and MANUAL modes.

Item 4: How do you check for the required surge volume?

Thanks again

Edited by sheiko, 02 October 2010 - 06:52 PM.

[Zauberberg]

It's difficult to give some general guidelines about surge volume, because it depends on the expected (out)flow. To picture this issue, imagine PV being just 1m away from PCV's - that would almost certainly cause relief valves to pop-open frequently, anytime when Nitrogen flow demands increases - you agree? In such case, simply there is no capacity in the system to accommodate even for minor flow changes.

If with PV in MAN mode and constant % opening, you still have problems with relief valves popping, I would say - the PV is not a problem for sure. Then look at PCV behavior once when you start bleeding the Nitrogen from the space in between the valves (PV/PCV).

It might be a controller tuning issue, it might be undersized/oversized valves (how much they are actually open in the field?), it might be surge volume issue. Do a few tests, analyze the system thoroughly and I bet we'll find the answer.

Is there anything else that we haven't considered so far? Hope to hear Art's opinion further in this thread.

Cheers,

[Zauberberg]

Not sure about the purpose of PV, but I believe two PCV's would serve the purpose - opening and closing based on actual Nitrogen demand in the network, while maintaining the system pressure.

You will see similar arrangement on majority of tanks blanketed with Nitrogen, also in the Cold Box N2 purge system (discussed recently) - a pressure regulator, a needle valve (for maintaining constant flow rate), and an optional flowmeter. See attached scheme.

In any case, let's not expand the subject till we get further feedback from your side.

Attached Files

•  Cold Box N2 Purge.pdf   61.39KB   44 downloads

[sheiko]

Thanks for the explanations and advice.

I indeed need to make a few tests and analyze them before being able to nail the problem.

FYI:
The PV vendor data sheet indicates:
min flow = 5 Nm3/h and opening = 16%
normal flow = 15 Nm3/h and opening = 26%
min flow = 200 Nm3/h and opening = 77%
rated CV = 70.8 (US units)

Regards

Edited by sheiko, 01 October 2010 - 02:53 PM.

[fallah]

min flow = 200 Nm3/h and opening = 77%

Minor Correction:

max flow = 200 Nm3/h and opening = 77%

[sheiko]

Do you really require 335 Nm3/h of N2 to flow, when the design flow is for only 15 Nm3/h? (and by the way, "Normal" conditions in Europe are measured at 1 atm and 0 oC – not at 1 barg; is this a NEW normal condition?)

Art,
335 Nm3/h is the relieving flowrate through the PSV (PCV failure scenario).
Max. operating flowrate through the PCV is 200 Nm3/h (vs normal flowrate = 15 Nm3/h)

Not sure about the purpose of PV, but I believe two PCV's would serve the purpose - opening and closing based on actual Nitrogen demand in the network, while maintaining the system pressure.

Yes indeed.

Minor Correction:
max flow = 200 Nm3/h and opening = 77%

Thanks Fallah.

Edited by sheiko, 02 October 2010 - 06:08 AM.

[Qalander (Chem)]

Dear Hello/Good Morning, Just a wild thought;what is the Nitrogen source temperature and state?

If it is liquid and in cryogenic temperatures,then few of the above indicated problems might be generated
from intermittent expansion/refrigeration effect;in case system not dry enough/duly purged from any moisture
internal cooling,condensation and freezing of water vapors could have contributed to the issue.

Hope this adds another point to ponder and check-out.

[sheiko]

Thanks for the input Qualander,

The gaseous nitrogen is flowing at about 15°C and is dry. We will however check for humidity presence in the tubing of the pilot/impulse lines.

Edited by sheiko, 16 September 2010 - 08:24 AM.

[Qalander (Chem)]

Thanks;I wish success in your problem resolution 'the soonest possible'

[sheiko]

Just to give you some feedback on my problem...

Finally, the tubing check (for pilot-operated pressure safety valve and pilot-operated pressure regulator) did not reveal any trace of humidity that could have been the cause of the misoperation of the insruments.
From my research (see attached file), i have found that increasing the spring rate of the PCV could reduce their sensitivity and thus, improve the stability of the system...I am thinking of inviting the PCV vendor on site to make a global check of their instruments.
If the manufacturer proves that the PCV are correct, then i will invite PSV manufacturer the same purpose. Or maybe replacing the actual PSV by conventionnal ones will help... We shall see.
Hope these will solve the issue, otherwise it will mean that there were a (or several) design error(s).

Attached Files

•  CVR402SelfOperatingPressureRegulators.pdf   188.64KB   32 downloads

Edited by sheiko, 18 September 2010 - 07:01 PM.

[ankur2061]

Sheiko,

I am entering late into this discussion but when I saw what your PCV is doing (reducing the pressure from 10 barg to 50 mbarg), it sure is expected to do a hell of a job.

The pressure drop of this magnitude requires multi-stage pressure reduction using a mult-trim control valve. You simply cannot use a conventional control valve for this kind of service. Large pressure drops in ordinary (single-trim) control valves cannot be handled. A single-trim control valve will cause cavitation leading to excessive noise and vibration.

Have you checked this aspect of the pressure control valve specification. IMHO, this needs to be a multi-trim pressure control valve for good operation.

Hope this helps.

Regards,
Ankur.

[sheiko]

A single-trim control valve will cause cavitation leading to excessive noise and vibration.

Ankur,
How cavitation can occur in gas service?
In my case, PCV are pilot-operated pressure-reducing valves (two-stages) not control valves.
Thanks.

Edited by sheiko, 01 October 2010 - 02:54 PM.

[ankur2061]

Sheiko,

My apologies. Cavitation in control valves is used for liquids causing bubbles to form & then collapse. However, noise & vibration is still observed in high pressure drop control valves. Control valves with large pressure drops are classified as "severe service control valves". You will find a lot of information if you google "severe service control valves".

Regards,
Ankur.

[sheiko]

Basically, you are regulating nitrogen flow from 10 barg down to 12 mbarg (0.012 barg). You are reducing the pressure almost a thousand-fold. Guess what????? YOU ARE CREATING CHOKED FLOW.
...
With sonic flow being provoked, you will generate some noise and movement if the piping is not properly designed and is intermittant.

I am entering late into this discussion but when I saw what your PCV is doing (reducing the pressure from 10 barg to 50 mbarg), it sure is expected to do a hell of a job.
...
Control valves with large pressure drops are classified as "severe service control valves".

Dears,

I was wondering if having choked flow through a pressure regulator (PCV) is unusual? I'm asking because the calculation note from the PCV vendor indicates a critical pressure drop below the actual pressure drop (9.95 bard), what confirms that the flow is indeed choked at some point in the valve.

I am also wondering if that sonic conditions (346 m/s for nitrogen) will be maintained in the downstream piping?
- if not, why/how?
- if the flow is indeed choked through the valve, then shall the downstream piping be sized for choked flow?

The instrument lead engineer of the project deny the fact that choked flow could be the issue. He believes that, because we have two-stage PCVs, we are correct. But i can't understand how two stages can be enough to mitigate the decrease from 10000 mbarg to 50 mbarg without creating choked flow, unless we can't consider that the flow through the PCV is similar to that through an orifice or a nozzle...

Could you please help clarify my doubts?

Edited by sheiko, 02 October 2010 - 06:47 PM.

[Qalander (Chem)]

Dear

I assume your instrument lead is suggesting

• In a way two-step throttling action that may work and
• in all fairness throttled flow just below the physical restriction may or may not be 100% choked,I understand
• due to immediate downstream open space availability.

Please do correct me if wrong!

[katmar]

Sheiko,

I cannot access the isometric and P&ID drawings that you mentioned, so I am missing some of the crucial details. What is the diameter and length of the pipe held at 50 mbarg between the two control valves? I believe this is the most critical element of the problem - I am agreeing with Zauberberg's comment regarding the surge volume between the valves.

What do we know for sure? We know that the PCV (ie upstream controller) controls well when the downstream PV is closed. So we know that it is a flow related problem. The PCV is doing its job, despite the huge ratio between the inlet and outlet pressures. This sort of ratio is common for gas regulators and I believe we can rely on the manufacturer as having got that part correct.

Did you try Zauberberg's recommendation (14 Sept 05:15pm) of starting with the PV closed and the intermediate pressure steady at 50 mbarg and then gradually opening the PV under manual operation? If you are able to get the downstream pressure up to 12 mbarg while the intermediate pressure remains steady then you know it is a control problem (which I suspect it is). Please tell us what happens when you try Zauberberg's experiment. This behavior, together with the knowledge of the intermediate pipe section's diameter and length, should go a long way to solving the problem.

What I suspect, is that the intermediate pipe section has been designed for a normal low pressure gas velocity of 15 - 25 m/s. If this is so, and your pipe length between the valves is only 2 m or so, then your residence time in the 50 mbarg section is only of the order of 0,1 second. A small pressure excursion will happen in a fraction of this and this is not enough time for the self acting pressure regulators to respond. Plus, you have the added problem of the upstream and downstream valves acting against each other. What I mean by this is that the system wants the flow rate to remain constant, but if the flow rate rises and the intermediate pressure rises too then the upstream valve will tend to close to lower the pressure. BUT because the intermediate pressure has risen the downstream valve will also CLOSE, thus working against the upstream valve's action. If you are unlucky with the natural frequency of the interaction between the controllers and the piping system then it will happen that you will pop the PSVs.

If you could increase the diameter of the intermediate pipe section to give a residence time of 5 or 6 seconds you will give your controllers a much better chance of doing their job. If you already have this amount of residence time then it has to be a badly tuned controller on the downstream PV.

Another advantage of having a larger diameter pipe between the valves is that the pressure drop will be negligibly small and the pressure that the downstream PV "sees" will be virtually independent of the flow rate (assuming the upstream PCV is doing its job).

A side issue - in 1982 IUPAC revised their definition of Normal Conditions from 0°C and 1 atm abs to 0°C and 1 bar abs. Also the term "Standard" conditions seems to be preferred to "Normal" these days. I find this adds to the confusion because "Standard" conditions in Europe used be be taken as 15°C and 1 atm abs (760 mmHg). The right thing to do is to always define your normal conditions (as Sheiko did here, except that he said barg instead of bara)

[sheiko]

Did you try Zauberberg's recommendation (14 Sept 05:15pm) of starting with the PV closed and the intermediate pressure steady at 50 mbarg and then gradually opening the PV under manual operation? If you are able to get the downstream pressure up to 12 mbarg while the intermediate pressure remains steady then you know it is a control problem (which I suspect it is). Please tell us what happens when you try Zauberberg's experiment. This behavior, together with the knowledge of the intermediate pipe section's diameter and length, should go a long way to solving the problem.
...
Plus, you have the added problem of the upstream and downstream valves acting against each other. What I mean by this is that the system wants the flow rate to remain constant, but if the flow rate rises and the intermediate pressure rises too then the upstream valve will tend to close to lower the pressure. BUT because the intermediate pressure has risen the downstream valve will also CLOSE, thus working against the upstream valve's action.

Katmar many thanks.

Yes we tried what Zauberberg had recommended by slowly opening the PV in Manual mode. But once the PV is, let's say, 5% open, then the PSV (between PCV and PV) start discharging.

We also have tried another way, that is let the PV wide open, and start opening the manual valves upstream the PCV. We made the same observation: at the very beginning of the opening (about 5%), the PSV start discharging.

BTW we have pilot-operated PCV and pilot-operated PSV. Line size is 2". Line length between PCV and PSV is about 1 m, and length between PSV and PV is about the same (1 m). Min. straigth length between the items is 5D.

As for the interaction between the PCV and the PV, i understand that the valves tend to work against each other, but i don't catch how both the intermediate pressure and flowrate can rise at the same time, in your example?

Regards

Edited by sheiko, 02 October 2010 - 06:40 PM.

[Zauberberg]

Having 2 meters between 2 control valves (PCV and PV in your sketch) doesn't seem to be a viable scheme. They simply cannot respond fast enough while avoiding to act against each other.

1st case: you start opening PV on manual, slowly. The flow downstream increases. Pressure in between the valves decreases. PCV starts to open and Nitrogen flows in and raises the pressure. PSV lifts off while the PCV is in manual. I believe this would not have happened if the surge volume between two control valves is large enough to compensate for such fluctuations.

In my opinion, this is the evidence that the system cannot operate in a self-controlled way even when minor flow changes are involved. If you can't operate in manual, than it is not a Control problem - it is a process problem. The difference between set points for PCV and PSV is only 30mbar. For the piping volume in between the valves = 4 liters, the volume of Nitrogen required to lift the PSV is only 0.12 additional liters (please re-check my calculations as I have just come back from Night shift). That certainly doesn't sound promising for control.

Another possibility is that a shock wave (due to critical flow) is created as soon as the PCV is just slightly open, which would lift off the PSV in any case. But it's not very likely especially if you have a reputed vendor of pressure regulators. If this was too complicated to design and operate, I think that blanketing of storage tanks with Nitrogen would turn into a nightmare across the planet.

[katmar]

I agree that if the problem occurs in manual mode then it is more of an intrinsic flow problem than a control problem. Zauberberg's calculation of only an extra 0.12 litre required to lift the PSV puts the situation in perspective. BTW, I agree with the calculation of the extra 0.12 litre since 4 x (1080/1050) = 4.12. To put it even more in perspective, consider that the residence time is close to 1 second and the required flow variation is only 3% (i.e. 0.12 / 4 = 0.03). A variation of 3% on flow for 1 second is probably the sort of control variability that could be expected from a self acting regulator.

BUT, the problem occurs in manual mode as well so it is definitely more than a control problem. Previously I suspected that the residence time was about 0.1 second but in fact it is close to 1.0 second. With this in mind you probably need more than the 5 or 6 seconds I recommended previously. 2 m of 200 NB pipe would give you about 15 seconds.

As Zauberberg has also noted - the nitrogen blanketing of tanks is a universal practice and in general it is not a problematic operation. This makes me suspect there is something else strange in your setup - especially seeing that it occurs in manual mode. Are there any other restrictions, valves or fittings in the line between the PSV and PV? Hopefully nothing like a spring operated non-return valve?

I don't catch how both the intermediate pressure and flow rate can rise at the same time, in your example?

What I had in mind was the situation where the intermediate pressure has gone low, and the PSV responds by opening. This increases the flow rate and the intermediate pressure starts to increase at the same time - i.e. the desired outcome of the control action. There will be a brief period of overshoot after the set point is reached where the pressure is still rising and now the PSV starts to close. But the PV also sees the higher intermediate pressure and also starts to close.