Vapor breakthrough relief 1

[prq123]

Refer to attached schematic.

The scenario: We have an acid gas removal unit which uses a liquid solvent to treat natural gas. There are three columns (absorber COL-1, flash column COL-2, regenerator COL-3). Absorber COL-1 operates at around 70 barg. For the case where the solvent flow to the asborber COL-1 stops (due to FV-1 fails closed), there is potential for vapor breakthrough to the downstream lower pressure columns due to loss of liquid inventory in the columns. We assume that both LV-1 and LV-2 (bottoms level control valves) stay in their original position when FV-1 fails closed (as per API 521, no credit is taken for favorable instrumentation response). The liquid hold up time in COL-1 and COL-2 are about 3 minutes each. First the relief valves on COL-2 would likely lift. In about 3 minutes (when all liquid is gone from COL-2) vapor breakthrough will occur from C0L-2 to COL-3.

The question: Is this a credible relief scenario to be considered for sizing of PRV-2 on COL-3?

 

[prq123]

I would like to add that the current vapor breakthrough case for PRV-2 on COL-3 only considers vapor breakthrough due to fail open of LV-2 on COL-2. The scenario I described above results in a significantly higher relief rate.

 

[chemicalengineerthai]

1) Yes, the gas from CO-1 will go to liquid line to CO-2 and similarity from CO-2 to CO-3, if no liquid seal at bottom column.

2) But the gas rate will limit by valve size & system delta P between valve, You should open the Cv-(Liquid valve coffeficient)valve originally size for liquid , you just convert to Cg-valve ( gas valve cofficient) & use driving force of delta P, for example CO-1-->CO-2 is 7-10=60 barg)

3) If delta P is too high compare to critical valvedelta P, but in the critical senarios, the problem can be happen,especiall if gas outlet at CO-1 may block outlet case.

4) So, Gas will breakthorugh liquid valve but you just change from Cg-->Cg

 

[Ninja1980]

I would consider LV1 failing open 100% and LV2 staying in normal position.

If the backpressure of the flare line was less than the operating pressure of COL3 plus the losses in the inlet line and contol valve wouldn't all the gas flow though the PSV on COL2 anyway?

 

[CMA010]

Why is the case you described much bigger than failing open of LV-2?

 

[JoeWong88]

I am a bit lost on your scenario...anyway...

Some observations :

In case of FV-1 fails closed, no solvent feeding into COL-1. Controller for LV-1 and LOW-LOW level trip for COL-1 still healthy to ensure liquid sealing the COL-1. So...why you have gas blowby ?

LV-1 failure would lead to liquid level in COL-1 lose to COL-2 and subsequently lead to gas blowby from COL-1 to COL-2.
PRV-1 shall be able to take the load for this scenario. Regardless of position of FV-1, this scenario still exist. Don't see any relation of gas blowby with FV-1.

Similarly PRV-2 has to handle relief load due to failed open of LV-2. Again it has not relation with FV-1.




JoeWong
Chemical & Process Technology

 

[prq123]

Joe,
When FV-1 fails closed, we assume that all the other control valves stay in their original position. We do not take credit for favorable instrumenation action in the LV's. This is clearly stated in API 521. Please have a read of API 521 and confirm. With no liquid feeding COL-1 and the LV's still in their original position, this will lead to gas blowby through both LV's.

CMA,
The normal pressure in COL-2 is about 6 barg with very small amount of vapor generated from the flash into COL-2. Also the volume in COL-2 is very small compared to the volume of COL-3. It is doubtful that the amount of vapor in COL-2 would be sufficient even to overpressure COL-3 starting from its normal operating pressure of 1 barg.

 

[JoeWong88]

prq123,
What is the cause of FV-1 failure ? Is this cause can result failure of FV-1 and LV-1 failed simultaneously or sequentially ? If not, can this be considered double jeopardy ?

You consider :
- FV-1 failed to close
- LV-1 stay at position (mean the controller failed to control)
- Low-low level trip in COL-1 failed to take action
- Low-Low pressure trip in COL-1 failed to take action
- LV-2 stay at position (mean the controller failed to control)
- Low-low level trip in COL-2 failed to take action
- Low-Low pressure trip in COL-2 failed to take action

all above cause liquid in COL-1 lost to COL-2 and COL-2 lost to COL-3, gas blowby from COL-1 to COL-2 and to COL-3...and relief via PRV in COL-3.
Please correct me if i mistakenly interpret your scenario.

There are so many failures but not from same cause. Is this credible ?

Just make an analog scenario...
If one of the control valve at slugcatcher failed, do you consider all the control valves downstream of slugcatcher i.e amine, dehydration, dewpointing, NGL extraction, condensate stabilization, etc "stay" in position and inventory slowly lose and finally gas blowdown from slugcatcher all the way to condensate tank ?

I have never consider this scenario is credible.


JoeWong
Chemical & Process Technology

 

[prq123]

Joe,
We are trying to design a plant that complies with API 521. I do not see anywhere in your respone a reference to API 521. Can you provide any reference to API 521 that support your statements?

I can tell you that API 521 explicitly states that credit should not be taken for favorable instrumenation control response when sizing process equipment pressure relief:

API 521 (2007) section 5.10.2 with regards to relief capacity credit states "In other words, no credit
should be taken for any favourable instrument response."

API 521 section 4.2.2 states "For example, latent failures can exist in instrumentation that prevents it from functioning favourably during an overpressure condition. It is not double jeopardy to assume the absence of beneficial instrumentation response in combination with an unrelated overpressure cause. "

API 521 section 5.10.3 with regards to inlet control valves states "If the system has multiple inlets, the position of any control device in those remaining lines shall be assumed to remain in its normal operating position."

My experience is that where we have a high pressure system liquid being letdown to a lower pressure system, the PRV on the low pressure system is sized for the vapor breakthrough with the full open Cv of the control valve and normal operating pressure upstream (in the high pressure system), irrespective of the volumes in the high pressure and low pressure system. This is a conservative assumption, but it avoids discussion on the scenario I am describing. Unfortunately, for PRV's on COL-3 this has not been done. PRV's on COL-3 has been sized for the normal flash vapor in COL-2 which is about 2 t/h. This compares to the vapor breakthrough relief based on the LV-2 CV of about 350 t/h.


To answer your specific points:
1) FV-1 can fail closed for any reason (loss of IA, faulty signal, etc).
2) As explained above, the API 521 requirement is to assume the other control vavles in the system stay in their original position. The LV-1 and LV-2 do not necessarily need to "fail". They could be in "manual" mode by the operator. This is why we do not take credit for the control response.
3)Your understanding of the scenario is correct. We have low level trips on COL-1 and COL-2. However, these trips are SIL 1 and SIL 0, respecivetly. The SIL rating is based on the the downstream PRV being adequately sized for the vapor breakthrough case (which is not for COL-3 based on the scenario I described). None of the safety shutdown systems on the COL-2 and COL-3 are designed to eliminate the vapor breakthrough case for PRV sizing. We do not have low pressure trips.

4) Regarding your analog scenario, API 521, section 5.4 does allow you to take credit for operator response if the operator has enough time. API 521 recommends 10-30 minutes, depending on the complexity of the plant. For the impact on the downstream units you mentioned I would expect the operator would have enough time to respond if all the controllers were in manual mode. As I mentioned in my earlier email, the liquid hold up in COl-1 and COL-2 is about 3 minutes each. So, in about 6 minutes we would lose the level in both columns. This is not enough time to allow for operator response.

 

[JoeWong88]

prq123,
I am aware of all the API statements as mentioned by you.

My experience is that where we have a high pressure system liquid being letdown to a lower pressure system, the PRV on the low pressure system is sized for the vapor breakthrough with the full open Cv of the control valve and normal operating pressure upstream (in the high pressure system), irrespective of the volumes in the high pressure and low pressure system. This is a conservative assumption, but it avoids discussion on the scenario I am describing. Unfortunately, for PRV's on COL-3 this has not been done. PRV's on COL-3 has been sized for the normal flash vapor in COL-2 which is about 2 t/h. This compares to the vapor breakthrough relief based on the LV-2 CV of about 350 t/h.

What you have said is correct. This is gas blowby or gas breakthrough case. The only point i would like to add to your statement is the upstream pressure may be upto maximum pressure before it trip i.e. PAHH as control system and/or operator failed to take any action when PAH is triggered. So ay consider PAHH instead of normal operating pressure.


API 521 section 4.2.4 states "Fail-safe devices, automatic start-up equipment and other conventional instrumentation should not be a substitute for properly sized pressure-relieving devices as protection against single-jeopardy overpressure scenarios. There can be circumstances, however, where the use of pressure-relief devices is impractical and reliance on instrumented safeguards is needed. Where this is the case, if permitted by local regulations, a pressure-relieving device might not be required."

Also refer to Annex E.

API allows Safety Instrumented System in safeguarding a process system.


Let comes back to your system :
I presumed you have SDV1 on liquid line from COL-1 to COL-2. LALL1 in COL-1 will trip this SDV1.
I presumed you have SDV2 on liquid line from COL-2 to COL-3. LALL2 in COL-2 will trip this SDV2.

As you have PSV in COL-2 and COL-3, i presumed you have High-High Pressure trip in COL-2 (PAHH1) and COL-3 (PAHH2) These are required per API14C. PAHH1 will shut SDV1 and PAHH2 will shut SDV2.

Gas breakthrough from COL-1 to COL-2. It only involve 1-level protection (LALL1+SDV1), with SIL-1 level protection system in COL-1, it is insufficient. Thus gas breakthrough from COL-1 to COL-2 is credible.

As there is PAHH1 in COL-2 will trigger in case of gas breakthrough. However, it will only shut SDV1. As it is same SDV used for LALL1 and PAHH1, thus does not increase the SIL level. Thus gas breakthrough from COL-1 to COL-2 is still credible.

Similar analysis for gas breakthrough from COL-2 to COL-3.

BTW, why SIL-1 for COL-1 and SIL-0 for COL-2 ?



Now, let goes a step more to your scenario.
To have gas breakthrough from COL-1 to COL-2 follow by from COL-2 to COl-3.

It involve failure on demand of LAHH1+PAHH1+SDV1+LAHH2+PAHH2+SDV2. The Probability of Failure on Demand (PFD) may drop below 0.001 (please check) which is sufficient for a PSV in COL-3. Breakthrough of LV-1 & LV-2 may not credible.

I may be wrong. Please check and come back for knowledge sharing.

Infact in my earlier responses, there is no or nearly no relation between failure of FV-1 and gas breakthrough via LV-1 & LV-2. Regardless FV-1 fail or not, gas breakthrough via LV-1 and/or LV-2 scenario still has to be analysed.

BTW, i am glad that you have a good understanding in this area.

JoeWong
Chemical & Process Technology

 

[prq123]

Joe,
I am aware that API 521 allows the use of instrumented HIPS instead of a mechanical pressure relief device (PRD) to address a potential overpressure case. However, the use of HIPS is subject to the owner's risk tolerance criteria, as stated in API 521: "The design shall comply with the local regulations and the owner’s risk tolerance criteria, whichever is more restrictive". I can tell you that as owner, our company policies would not allow the use of HIPS for this case as the potential overpressure of the C0L-3 is too high (would be over the hydrotest pressure of COL-3). I should also add that there are 5 pieces of equipment (reboiler, condensers, KO drum) being protected by the PRV's on C0L-3 which have already been built and installed at site. Of course, none of these equipment have complied with ASME Code Case 2211 requirements for use of HIPS.

Even if we could take credit for a HIPS, the point I am trying to make is that based on API 521 guidelines vapor breakthrough from LV-1 cascading to LV-2 is a valid potential overpressure scenario for COL-3 that needs to be addressed (with either a HIPS or PRDs). Do you agree with this point?

 

[JoeWong88]

prq123,

Every project or company has their own believe, risk or tolerance, etc. However, it shall be written down as standard, basis and policies to be complied to.

If these documents do not believe in Safety Instrumented System (SIS) can safeguard the system (where API allow now), and do not accept the fact that probability of failure in demand (where API allow now) lower than certain targeted figure, the scenario becomes not credible, you may go in the way as you doing. Just take the consequence that you will have a very expensive design.

Just add some interesting point...
API 521 stated that check valves (2 or more in series) latest failure, it allow back flow to be calculated based on 10% of cehck valve flow area. But some company do not accept this. The argument is the internal could be missing without knowing. So full opening has to be considered. The consequence is large PSV for back-flow and large tail pipe, flare, etc.

Your consideration of credit for operator intervention for more 10-30 minutes as stated in API, i would put less degree on this. Reason being operator can behave differently anytime, any place, anyway, etc. If your basis document stated ultimate safeguarding can rely on operator intervention, i have nothing more to input but take it.

I have experienced working with many established oil & gas companies. They do accept SIS and i am having same opinion. This give us an opurtunity to have a reasonable and practical design.

Last but not least, your statement "Even if we could take credit for a HIPS, the point I am trying to make is that based on API 521 guidelines vapor breakthrough from LV-1 cascading to LV-2 is a valid potential overpressure scenario for COL-3 that needs to be addressed (with either a HIPS or PRDs). Do you agree with this point?"

Yes. It needs to be addressed. Anyway it will be discussed in HAZOP & IPF review (if your company believe these tools).

JoeWong
Chemical & Process Technology