14 replies to this topic

[carly]

Hi,

May I know when we are talking about Overfilling scenario for liquid relief sizing, do we consider the drum pump to be running to bring the drum liquid level down? Relief load for PSV sizing would be maximum inflow - outflow (i.e. pump capacity)?

Thanks for answering my question.

[JoeWong]

Hi,

May I know when we are talking about Overfilling scenario for liquid relief sizing, do we consider the drum pump to be running to bring the drum liquid level down? Relief load for PSV sizing would be maximum inflow - outflow (i.e. pump capacity)?

Thanks for answering my question.

Credit on the outflow should not be taken for overfilling scenario.
Reason being the pump (outflow) could failed stop and with continue inflow into the tank.

[carly]

may i ask then what is the distinction between overfill and block outlet? i think i am confused.

[JoeWong]

may i ask then what is the distinction between overfill and block outlet? i think i am confused.

Lets take an example...
Liquid feeding a tank with proper control valve and control loop. And liquid is pump out from the tank feeding user with proper control valve and control loop.

i) Block discharge
The control valve on the pump discharge closed due to control loop failure results no flow.
This may apply to vent line.

ii) Overfill
The control valve on the feeding line failed (to FULL OPEN position) leading to liquid level increase. An overflow line may be a potential ultimate safeguarding measure for tank ovefill.

Hope this help.

[latexman]

I think the distinction can be somewhat semantic too. A vessel or tank that is not normally run full, can be overfilled. A vessel that is normally run full, cannot be overfilled, but can be blocked-in.

[carly]

thanks for your explanation, it helps!

[aju_1807]

Hi All,

Liquid overfilling with following condition is a credible scenario or is it a double jeopardy?

We have a process vessel with LCV on the liquid outlet line. Now, due to LIC failure the control valve goes to closed condition adn it will result in the overfilling of the process vessel. Normally, all process vessel dont have the residence time of more than 3-5 minutes whereas API521 calls for the operator action time of 20 minutes now in this case its always impossible to provide the operator response time of 20 minutes to avoid the overfilling scenario and vessel relief valve has to always size for the liquid overfilling case. Dont you think its a doubel jeopardy consdiering the LCV has failed and HH level in vessel is not tripping the inlet shutdown valve????

Particulaarly, when we have large liquid relief case we cant size KO drum for such a high liquid relief load...

I am really very confused. Appreciate expert opinion on this issue.

Regards,

Ajay

[fallah]

Dont you think its a doubel jeopardy consdiering the LCV has failed and HH level in vessel is not tripping the inlet shutdown valve????

Considering below fact from API 521:

"4.2.1 Double jeopardy
The causes of overpressure are considered to be unrelated if no process or mechanical or electrical linkages
exist among them, or if the length of time that elapses between possible successive occurrences of these
causes is sufficient to make their classification unrelated. The simultaneous occurrence of two or more
unrelated causes of overpressure (also known as double or multiple jeopardy) is not a basis for design."

and if there is one LT for LIC and separate one for HH level,i think it would be double jeopardy.

Edited by fallah, 08 September 2009 - 12:51 AM.

[aju_1807]

Dont you think its a doubel jeopardy consdiering the LCV has failed and HH level in vessel is not tripping the inlet shutdown valve????

Considering below fact from API 521:

"4.2.1 Double jeopardy
The causes of overpressure are considered to be unrelated if no process or mechanical or electrical linkages
exist among them, or if the length of time that elapses between possible successive occurrences of these
causes is sufficient to make their classification unrelated. The simultaneous occurrence of two or more
unrelated causes of overpressure (also known as double or multiple jeopardy) is not a basis for design."

and if there is one LT for LIC and separate one for HH level,i think it would be double jeopardy.

So in general, if vessel have LIC and separate LT for HH trip then we dont have to consider PSV sizing for liquid overfilling?? Actually,we normally dont take any credit of any instrumentation for the sizing of relief valve,isnt it?

Appreciate expert opinion on this issue.

Regards

Ajay

[fallah]

So in general, if vessel have LIC and separate LT for HH trip then we dont have to consider PSV sizing for liquid overfilling?? Actually,we normally dont take any credit of any instrumentation for the sizing of relief valve,isnt it?

In your case, no.What are the other scenarios?

Actually,any PSV should be sized based on worst case among all applicable scenarios.We don't know all scenarios in your system to be able to answer your question accurately.

Edited by fallah, 14 September 2009 - 03:55 AM.

[aju_1807]

So in general, if vessel have LIC and separate LT for HH trip then we dont have to consider PSV sizing for liquid overfilling?? Actually,we normally dont take any credit of any instrumentation for the sizing of relief valve,isnt it?

In your case, no.What are the other scenarios?

Actually,any PSV should be sized based on worst case among all applicable scenarios.We don't know all scenarios in your system to be able to answer your question accurately.

Hi,

Whatever is the sizing case for vessel but my worry is not sizing case but KO drum capability to handle this flow rate (huge)....

I have taken care of all other sizing cases but looking help on the credibility of the overfilling case for process vessel as the residebce time will be surely very less for it.

Regards,

Ajay

[aju_1807]

So in general, if vessel have LIC and separate LT for HH trip then we dont have to consider PSV sizing for liquid overfilling?? Actually,we normally dont take any credit of any instrumentation for the sizing of relief valve,isnt it?

In your case, no.What are the other scenarios?

Actually,any PSV should be sized based on worst case among all applicable scenarios.We don't know all scenarios in your system to be able to answer your question accurately.

Hi,

Whatever is the sizing case for vessel but my worry is not sizing case but KO drum capability to handle this flow rate (huge)....

I have taken care of all other sizing cases but looking help on the credibility of the overfilling case for process vessel as the residebce time will be surely very less for it.

Regards,

Ajay

Hi,

As I mentioned in my earlier post, the scenario is during shutdown event allSDv closed and inlet to slug catcher failed to close which will evntually result in overfilling of vessel. Now, there is following protetion for vessel:

1. High pressure alarm
2. High level alarm
3. High High trip (but in this case we considered the failure of inlet SDV failed to close)

In this scenario, basically we are relying on intrumentation only and as per API we should not take any credit of instrumentation. Now, my problem is if we size the KO drum for liquid relief scenario then it will be huge flow rate (~500 m3/h) and normally at first place we shgould try to avoid teh overfilling scenario,isnt it?

I am really clueless to handle this situattion. Normally, in most of the vessels we dont have much of the action time (above HH level, as API says operator action time of 10-30 minutes and action time mentioned in API is always impossible to provide in vessel.

Please provide me your vaulable feedback on this issue. I have raised this issue many times in forum but dint get any positive feedback.

Appreciate some expert view.

regards

Aju

[Lowflo]

aju,

This is a textbook application for using HIPS rather than a relief valve. Refer to API 521 Annex E for information on the use of HIPS.

[gvdlans]

Before Texas City Refinery accident, see Wikipedia page, I would not have considered an overfilling scenario if the operators would have > 20 minutes to take corrective action. Now I will consider it. Note that API made an addendum to its standard 521 as a result of the accident and the following investigations. See API STD 521 addendum 2008. Please check out this addendum.

My approach here would be: do a HAZOP (HAZard and OPerability) study in combination with a Layer Of Protection Analysis (LOPA), to identify scenarios and to see whether the risk (e.g. of overpressuring the tower, but also of sending liquid hydrocarbons to the flare) has been reduced to an acceptable level.

Just saying "it is not credible if you have an independent high level trip" does not make sense to me. How much risk reduction does this high level trip provide?

Also the "double jeopardy" approach does not make sense to me here. In LOPA you consider failure of several independent protection layers. If the outcome of a scenario could be severe (e.g. multiple deaths), you will need several independent protection layers to bring the mitigated risk to an acceptable level.
Note that API 521 section 4.2.1 as quoted by fallah deals about unrelated causes of overpressure. The high level trip is not a cause of overpressure but a protection layer. If you would follow fallah's approach, having one protection layer would always be sufficient...

I would need much more information (e.g. P&ID, plotplan, instrument and relief valve datasheet etc.) before I can even start to give you a concrete recommendation.

Edited by gvdlans, 27 September 2009 - 12:54 AM.

[aju_1807]

Before Texas City Refinery accident, see Wikipedia page, I would not have considered an overfilling scenario if the operators would have > 20 minutes to take corrective action. Now I will consider it. Note that API made an addendum to its standard 521 as a result of the accident and the following investigations. See API STD 521 addendum 2008. Please check out this addendum.

My approach here would be: do a HAZOP (HAZard and OPerability) study in combination with a Layer Of Protection Analysis (LOPA), to identify scenarios and to see whether the risk (e.g. of overpressuring the tower, but also of sending liquid hydrocarbons to the flare) has been reduced to an acceptable level.

Just saying "it is not credible if you have an independent high level trip" does not make sense to me. How much risk reduction does this high level trip provide?

Also the "double jeopardy" approach does not make sense to me here. In LOPA you consider failure of several independent protection layers. If the outcome of a scenario could be severe (e.g. multiple deaths), you will need several independent protection layers to bring the mitigated risk to an acceptable level.
Note that API 521 section 4.2.1 as quoted by fallah deals about unrelated causes of overpressure. The high level trip is not a cause of overpressure but a protection layer. If you would follow fallah's approach, having one protection layer would always be sufficient...

I would need much more information (e.g. P&ID, plotplan, instrument and relief valve datasheet etc.) before I can even start to give you a concrete recommendation.

Could you please provide me your email id so that i can forward you relevant details.

My email id is ajay.sharma@petrofac.com

Thanks

Aju