3 replies to this topic

[KeJoSa2010]

Hello,

I have a question about analyzing a blocked flow situation when sizing a PSV protecting 2 and 3 phase separators. Is it common practise to assume that all of the outlets are completely blocked? Under what conditions can I take credit for one or more outlet that may not be completely blocked?

Thanks!

[fallah]

Hello,

I have a question about analyzing a blocked flow situation when sizing a PSV protecting 2 and 3 phase separators. Is it common practise to assume that all of the outlets are completely blocked?...No... Under what conditions can I take credit for one or more outlet that may not be completely blocked?...Some conditions such as CSO or LO valves on outlet line...

Thanks!

[KeJoSa2010]

Hi Fallah,

Thanks for your response. So, if I have a 3-phase separator and there are control valves on each outlet (Vapour, HC Liquid and Aqueous) then according to API 521 section 4.2.2 it is not double Jeopardy to assume the absence of a beneficial instrumentation response....My question is...would this be a situation (3-phase separator above) in which I can assume all outlets are blocked or, should I assume only one outlet is blocked and take credit for the outlet flow of the other two streams? There are no cso or lo valves on any of the outlet lines.

Thanks!

[paulhorth]

KeJoSa,

You need to think carefully about the different causes of overpressure and different modes of malfunction. API RP14C adopts the principle that there should be two independent means of protection provided for any identified malfunction, allowing for two independent failures.  So you have (1) malfunction, (2) first protection failure (3) then the relief valve as second level of protection.

Let us consider two causes of overpressure - gas blowby from an upstream vessel, and blocked outlet from the vessel under consideration.

(1) Gas Blowby

In this situation, you have already had possibly two failures: (1) failure of normal level control in the upstream vessel, (either instrument or valve) and (2) failure of the low level trip which should be present and acting on an independent valve. So it is reasonable in these circumstances to assume hat the separator outlets are all at their normal openings and not also failed. The relief case then becomes the gas blowby flowrate in, minus the gas flow out at the relieving pressure. Of course if you don't have a low level trip upstream then you have only one failure, not two, so you should not take credit for all being OK on the second vessel.

(2) blocked outlet

In this situation you have the normal flow into the vessel but one or more outlets are blocked. This can happen either by maloperation ( a manual valve closed) or a failure in a control valve. The protection systems on the separator should include a hign level trip and a high pressure trip.

Case (a)  If the liquid oultet is open and the gas is closed,  the high level trip will not activate but the high pressure trip should do. If this trip fails (one instrument failure) the relief case is then the total vapour inflow . So this relief case arises following one malfunction plus one protective level failure, as laid down in API RP 14C.

Case (   b ) if the liquid outlet is closed but the gas is open -, then the high level trip should act. If it fails then the high pressure trip should also protect you but not fast enough. The relief case would then become the total inflow (vapour plus liquid) minus the two-phase outlet capacity of the vapour outlet line. But this is likely in turn to become rapidly filled so it is better practice in my view to take the relief load as the total inflow.

In my experience, in the case of an inlet separator on a platform or receiving from a pipeline, it is the usual practice of oil companies to take the relief load as the total inflow ( a two-phase relief load) and not mess about trying to take credit for other outlets. If this relief load is excessive you will need to provide some kind of multiple instrument protection (HIPPS or its level equivalent).

 

Paul

Edited by paulhorth, 29 January 2013 - 04:01 PM.