37 replies to this topic

[rohollah]

Dear members

As you can see in the attached picture, there is a by-pass line for PSV.

Can anybody tell me the reason of that?

because I have not seen anything like that in API 520 II.


Regards

Rohollah

Attached Files

•  Photo-03491.jpg   293.06KB   348 downloads

[fallah]

Rohollah,

Bypass line provision isn't a main matter of PSV installation that would need to be included in API 520 Part II. It is usually considered for maintenance purposes and should be specified with a single isolation valve only where there is a requirement to vent the gas in the system to flare while there is no other connection to flare nearby.

Fallah

[rohollah]

Dear fallah

Thank you for your perfect replies.
In this case (single isolation valve) if we remove the PSV, gas goes out from the inlet flange.
How can it be justified.

Regards
Rohollah

[S.AHMAD]

1. Obviously, the bypass line in the picture is not for PSV onstream maintenance. The PSV requires isolation valves for the purpose.
2. Most probably the bypass is for purging the system during start-up and shutdown. Otherwise the bypass line redundant.

Edited by S.AHMAD, 24 January 2012 - 10:02 PM.

[fallah]

Dear Rohollah,

The bypass line around a single PSV should normally be taken from upstream of the PSV inlet isolation valve and terminated to downstream of PSV outlet isolation valve.

Hence, if the bypass line in the photo is considered for maintenance purpose seems it had to be taken from upstream of PSV inlet isolation valve and present configuration may be wrong from this standpoint.

Fallah

[ZBR]

when you take the PSV out next time try to install a valve on psv inlet side too this is what happens when you dont check your P&IDs

[paulhorth]

Rohollah,
It is usual to install a small line for purging while the vessel is off-line, before and after manned entry. This line is usually taken from the PSV inlet and is routed downstream. It is nothing to do with maintenance of the PSV, because maintenance of the single PSV requires the vessel to be out of service (in most oil company standards).
Your photo has a number of other features to comment on:

• The bypass line seems to have been retrofitted, judging from the blackened paint at the welds, so its use was not required or identified to start with and may be for some specific function
• The PSV inlet line is the same size as the outlet line. This is contrary to anything I have seen. Look at that huge inlet isolation valve. It should have been on the small bore section. But why is it there at all?
• I can see no lock on either the outlet or inlet isolation valve.This is UNSAFE.
• The bypass should have been routed to fall into the header, as it is there is a dead leg which will fill with liquid, with a risk of corrosion or freezing

Don't do what ZBR says. Isolating the inlet to a PSV is dangerous, against good practice, and you already have one valve there.

Paul

Edited by paulhorth, 29 January 2012 - 05:00 PM.

[S.AHMAD]

1. Agree with Paul thar carrying out onstream maintenane on single PSV is not a good practice.
2. However, I have seen such maintenance was carried out succesfully and sefely
3. Extra precautions are required during such activities such as ensuring pressure alarms are installed and functioning as expected.
4. Other means of controlling pressure is available and functioning. Such a s pressure control valve. The purpose of the bypass is for depressuring the vessel manually if needed to do so.
5. Operator must be extra alert and must be prepared to take emergency shutdown if so required.
6. The job was carried out after conducting risk assessment. In most cases the risk of doing onstream maintenance of PSV is remote since the probability for vessel overpressure is remote eventhough the consequence is high. Past history of overpressure should be analyzed to establish probability of occurence. Some PSV never pop in its entire life. In fact PSV is actually the 'last line of defence'. Other means of avoiding overpressure are normally available. For those who understand LOPA will understand what I mean.
7. Please take note that in this world, there is no such thing as "risk free'. It all depends on the management whether there are risk takers or risk averts.
8. Some of you may be suprise that the risk of travelling in a bus is higher than the risk of working in a chemical plant or travelling in a plane.

Edited by S.AHMAD, 29 January 2012 - 07:44 PM.

[Dacs]

Just to be off-topic for a bit:

The PSV inlet line is the same size as the outlet line. This is contrary to anything I have seen. Look at that huge inlet isolation valve. It should have been on the small bore section. But why is it there at all?

Off the top of my head, I'm thinking of the (oversized) inlet line size to meet the 3% pressure drop losses.

Maybe the vessel is located near grade (sump?) and the discharge header (assuming this is a closed system) is located on top of piperack.

Edited by Dacs, 30 January 2012 - 03:40 AM.

[Zubair Exclaim]

Paul:

isolating a PSV inlet is abolutely safe and as per design standards, the right way is to have locked open valves on psv inlet and outlet and locked close valve on the bypass line. This configuration is usually used when you dont have Spare PSV of N+1 capacity installed. this practice is common in refineries and gas plants. This PSV bypass can be used in maintainence operation as well as for purging and provides operator intervention option in dire need

[rohollah]

Paulhorth:

Regarding to your point of view PSV inlet and outlet lines can not be the same size.
is there any reference for that?

Rohollah

[rohollah]

S.AHMAD:

Thanks a lot for your reply.

What do you mean by "LOPA"?

Rohollah

[paulhorth]

Rohollah

Regarding to your point of view PSV inlet and outlet lines can not be the same size.
is there any reference for that?

The reason is obvious - gas expands as it goes from high pressure upstream of the PSV to low pressure downstream. So for a given mass flow the velocity will be much higher in a given pipe size downstream than upstream. This higher velocity causes a higher DP in the pipe, so to avoid excessive back pressure, as well as excessive velocity ( maximum 0.7 Mach), the outlet line size is almost always larger than the inlet line.
Pressure drop is proportional to (density) x (velocity) squared. This is proportional to (mass flow)squared /(pressure).

So, if your PSV is set at say 20 barg, the pressure drop downstream will be about 20 x upstream in the same pipe size. The velocity downstream will also be about 20 x upstream.
This is not the case for liquid relief, because liquid does not expand, but the relief valve in your photo does not look to be in liquid service.

On the question of isolation for a single PSV, I quote the Shell DEP 80.45.10.10 in the attachment. This prohibits isolation of a single PSV unless the vessel is out of service. I have similar quotes from Total, and BG, and I am confident that this is also the case for BP.

Paul

Attached Files

•  From Shell DEP 80.45.10.10 PSVs.doc   22KB   326 downloads

[S.AHMAD]

1. LOPA - Layer of Protection Analysis. This a semi-quantitative risk assessment where the probability of occurance is measured quantitatively while the consequence is measured qualitatively.
2. You can search internet for more info.
3. Attached some picture showing layers of protection

Attached Files

•  SIS&LOPA.pdf   387.68KB   297 downloads

[S.AHMAD]

1. The guideline for the discharge line size is the requirenent for the line pressure drop for not to be more than 10% of set pressure.
2. If the line is short, for example, and the pressure drop is much lower than 10%, there is no reason for the line cannot be of the same size as the inlet.

[fallah]

1. The guideline for the discharge line size is the requirenent for the line pressure drop for not to be more than 10% of set pressure.

AHMAD,

Above mentioned limitation is applicable just for conventional PRV at 10% allowable overpressure.

Fallah

[Zubair Exclaim]

Paul:

I didn't find the text in section 3.4 you quoted kindly see attachment .

Also, Appendix A is attached for your info which is a typical arrangement of a PSV. Which edition of Shell DEP are you using because i cant find the clause?

Attached Files

•  section3.4.doc   43.5KB   149 downloads
•  appendix-1.JPG   96.24KB   140 downloads

[paulhorth]

Exclamation,

I apologise, I had the section number wrong, it's in 2.8.2.
The diagram in the appendix is a schematic for the usual situation of two or more PSVs. The para that I quoted prohibits isolation of a single PSV unless the upstream system is out of service and depressured.

Paul

[Zubair Exclaim]

There is no section 2.8.2 either . However i did found attached section which doesnot prhibit use of inlet isolation valve ....

Attached Files

•  section2.2.doc   31KB   142 downloads

[Dacs]

Just to chime in on the discussion if it's acceptable to isolate a PSV with no spare online, with my previous experience with Middle East refineries (notably Saudi Aramco), they will never allow that philosophy.

Any provision for bypass line in this case will be used for other purposes, such as startup.

They will only do PSV maintenance during turnaround (for a single PSV with no spare) and they will likely request to have a spare PSV if there's any reason to do PSV maintenance without taking the associated equipment offline.

[fallah]

Dears,

IMO, for PSV isolation valve and bypass line, each company has its own interpretation on relevant codes and standards may result in practices included a little bit different points in details. Even when we study such practices our understanding about a point may be different. For example, when Paul says: "...The para that I quoted prohibits isolation of a single PSV unless the upstream system is out of service and depressured..." he may means isolation of a single PSV without a bypass line is prohibited, that at a glance seems to be logical.

Anyway, regarding the matter perhaps general guidelines of some codes and standards such as ASME Sec VIII (is attached) have been considered by engineering companies as a base and they did extend those general guidelines in details then to be formed as their own practices being used in their EPC projects.

I intended to conclude that some differences in various practices to be considered as a natural matter and the responsibility of the executive practice of each company might be left for that company against relevant project client.

Fallah

Attached Files

•  ASME SECTION VIII D1 PSV Isolation Valve.pdf   44.18KB   340 downloads

[Zubair Exclaim]

Thanks DAC for your input.

Ok lets move a little further... lets forget codes and standards

Suppose i am an operator working on a reactor vessel, etc. and have no spare PSV and bypass. This PSV is some how choked and reaction is going wild, presure is on the rise. Considering that there is a power/instrument air failure etc, and no instrumentation for normal control is functioning.

So, how should i intervene - in lets say, 10 min?

[paulhorth]

Exclamation,

Suppose i am an operator working on a reactor vessel etc and have no spare psv and bypass.This PSV is some how choked and reaction is going wild, presure is on the rise. Considering that there is a power/instrument air failure etc and no instrumentation for normal control is functioning.
so how should i intervene in lets say 10 min.

In this hypothetical case, opening a manual 2 inch PSV bypass is unlikely to save the situation. The case as you describe it supposes several failures in the design process.If a reaction can run away, there should always be failsafe means provided to control the situation. These can include depressuring, cooling, gas blanketing, external deluge, etc, which have to be available in an emergency. The case should be identified during design, in particular in a HAZOP, and suitable safeguards put in place. if you think the PSV could stick, then put in a spare - or a bursting disc. Don't rely on the operator in an emergency. This basic principle has been understood since the first steam boilers were built over 200 years ago.

In any case most PSVs on oil and gas plants are not for reactors but for simple pressure vessel systems.

Referring to your quote from the DEP. you have a more recent version, which seems to have been renumbered. No problem. In the very next sentence from the one you highlighted, we read:

"A single relief valve (without a spare) connected to a relief header which cannot be shut down will have only an outlet isolation valve."

that is, no inlet isolation valve. That's what I was saying. You need inlet isolation valves only when you have more than one PSV.

Paul

[S.AHMAD]

Dear all
1. The scenario that was posted by Exclamation what we called "double jeopardy" which is not being considered during HAZOPS and risk assessment.
2. As I have mentioned earlier, we have several layers of protection:
Layer 1 - Process Design
Layer 2 - Basic Process Control
Layer 3 - Critical Alarms and Human Intervention
Layer 4 - Safety Instrumented Function
Layer 5 - Relief Device
In the scenario, Layer 2 and Layer 5 failed. We still have Layer 1, 3 and 4 as a mean to prevent catastrophic failure due to overpressure.
In the case of utilities failure e.g instrument air, critical control valves such as flow of reactants to the reactor will be design to be "FC". So all reactants flow to reactor will stop and no additional reactions taking place. Instrument air failure will also trigger the emergency shutdown valve to close. That is Layer 1 protection will prevent catastrophic failure due to overpressure.
3. Under the situation where the relief device (Layer 5) is not available, we have 4 layers of protection. Are we still worry?
4. From this discussion, I am no longer wonder why top management always in the opinion that "engineers" cannot be good "managers".
5. Have anyone heard of RISK BASED DECISION MAKING (RBDM)? It is a good tool for managers (why not engineers use it too?). Most major companies has RISK MATRIX in their possession.
6. Fallah has mentioned ASME Code. API RP 521 Part II also recognizes the need of isolation valves for onstream maintenance.
7. Hey guys, don't you want to be managers? If you want then think like one.

Edited by S.AHMAD, 01 February 2012 - 08:03 PM.

[S.AHMAD]

1. The principle stated by Paul "Don't rely on the operator in an emergency" is debatable.
2. A better statement should be" Don't rely SOLELY on operator during an emergency". This means that, in addition to operator intervention,we shall install other devices for plant protection.
3. In LOPA, operator intervention that is coupled with critical alarms is the Protection Layer No. 3.
4. Engineer should not install a "Gold Plated Plant" just because of corrosion potential. SS316 or CS with corrosion nihibitor could be a better choice.

Edited by S.AHMAD, 01 February 2012 - 10:28 PM.