29 replies to this topic

[Qalander (Chem)]

Dear Sheiko, Again I have one of those wild thoughts triggered from reading the katmar's post..

Please check and visually inspect whole inside of the concerned pipeline segment if possible(if not done already)or pigging with similar dia cleaner pig for end-to-end clearance without any restrictions.
Sometime it happens totally unexpectedly and usual diagnostics and analysis appeared to fail.

• Since under pumping tank head was good full,
• not a single intermediate pipeline valve was throttled/shut but while commissioning the multistage centrifugal vertical crude pump,
• strange phenomenon of seal oil pressure drop and reduction of load(instead of increasing E/Motor amperes/flow rate on discharge valve opening) was observed.

I analyzed it as a result of NPSHa drop(for unknown apparent reasons); insisted&got obstruction confirmed through opening suction (bucket) strainer as we unexpectedly found fairly long wooden plank inside obstructing flow path.

Make sure this may be the case as indicated by katmar in above post.

[sheiko]

BTW, I agree with the calculation of the extra 0.12 litre since 4 x (1080/1050) = 4.12.
...
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.
...
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?

Dear Katmar,

1/ I presume you assumed a closed system between PCV and PV for the volume increase calculation. Actually, i would expect the volume of nitrogen to decrease as the pressure increases from 1050 to 1080 mbarg. Am i wrong?

2/ And how did you come up with a residence time of 1s (instead of 0.1s)? i thought you had calculated the residence time by dividing the length between PCV and PV (about 2 m) by the LP gas design velocity (about 20 m/s as you mentionned in your first post), resulting to a residence time of 0.1s...

3/ Lastly, how the flow can be choked in the PCV and not in the downstream piping? Is it because of the two stages inside the PCV? Or because of the diameter increase from the vena contracta inside the PCV to the downstream line size?

To answer your questions, there is no check valve in the circuit from the PCV to the PV, mostly elbows and tees.
As Qualander suggested, we have already blown the piping with nitrogen to clean and dry it.

And just to fix terminology: PCV = pressure regulator, PSV = pressure safety valve and PV = pressure control valve (all 3 in serie in the system under study).

Best regards and thanks all again for your willingness to help on your free (and/or money) time

Edited by sheiko, 03 October 2010 - 06:25 AM.

[katmar]

1. Sorry - I was sloppy with my terminology. The actual volume of the pipe is fixed, and indeed the specific volume of the nitrogen will decrease as the pressure increases. I meant that 0.12 Normal litres would be added to the +/- 4 Normal litres already in the system.

2. 15 Nm³/h in a 50 NB pipe gives a velocity close to 2 m/s. Since the length of pipe is about 2 metre the residence time is approximately 1 second. Previously I assumed 20 m/s over a distance of 2 m giving 0.1 second.

3. The flow will only be choked in the gap between the plug and the seat of the pressure regulator. The velocity in the 50 NB pipe is nowhere near choked. I do not think that choked flow is the root cause of the problem, although having choked flow across the PCV may be making its control accuracy worse, or allowing some slightly fluctuating flow.


A question going back to an earlier post - when you left the PV wide open and gradually opened the manual valve upstream of the PCV, was the flow still going through the PCV? When I first read this I took it to be a bypass around the PCV, but on closer reading I now take it that the flow still went via the PCV.

The fact that you were able to lift the PSV with the PV wide open and the manual valve upstream of the PCV at only 5% open suggests that the PCV is very much over-sized. You also mentioned that the PSV was sized for 335 Nm³/h (being PCV max flow), against a design PCV flow of 15 Nm³/h. You are asking the PCV to control at only 5% of its capacity and that may be the problem. The variation is 3% of 5% or only 0.15% of the full flow capability of the PCV. This is probably too high an expectation of a self acting regulator. Before modifying any piping I would try to get a lower range PCV, or fit an orifice upstream of the PCV to bring it into its comfortable range.

[Zauberberg]

Having the system responding in such a drastic way at only 5% valve opening surely means that something is incorrect in the relationship between the available surge volume capacity and valve size/controller tuning. It is a similar case to e.g. having extremely large controller gain (Kc) in a system with very small volume. High fluctuations will certainly occur in such system.

As suggested in the first troubleshooting post, % valve opening should show you whether you have any of the valves (PCV/PV) undersized or oversized. As Harvey also says, and what is also clear from your updated post, at 5% you are very likely to have inappropriate valve size for given surge volume capacity.

[sheiko]

Thank you all for your help. I believe (hope) i have enough inputs to be abble to find a correct solution and I have to say that i am really impressed by the quality of your arguments and explanations.
Anyway, i'll let you know what was found.

Cheers.