28 replies to this topic

[jprocess]

Dear All:
Please find attached a P&ID of a de-pentaniser unit.
First of all as can be seen the basic engineer has been put the tail pipe isolation valves of all psvs before the expander but the inlet line isolation valves of all psvs before reducer.
As I considered some other projects and procedures,the tail pipe isolation valve of psv is located after expander and I think this is a more convenient configuration and follow the concept of easy psv relieving.
A justification for this configuration can be cost because if we locate the isolation valve after the expander the size will be higher.
And my second question is about spare psv of reboiler on steam line.As can be seen the related tail pipe is attached to by-pass line(and not to flare sub-header) which usually is 2".I guess that this configuration be a drawing mistake anyway I can not find any concept or justification behind of it.
Your valuable comments are appreciated.

Attached Files

•  PSV_Configuration.pdf   92.5KB   234 downloads

[pleckner]

A justification for this configuration can be cost because if we locate the isolation valve after the expander the size will be higher.

Yes as well as ease of access to the PSV and block valve. The back pressure contribution of the smaller (rather than the larger) block valve against the PSV is generally too small to worry about in the scheme of things.

The block valve upstream of the reducer in the inlet line is obviously to minimize its effect on the 3% Rule.

And my second question is about spare psv of reboiler on steam line.As can be seen the related tail pipe is attached to by-pass line(and not to flare sub-header) which usually is 2".I guess that this configuration be a drawing mistake anyway I can not find any concept or justification behind of it.

I would have to see the isometric or piping drawing to make a real comment. But saying this, it looks like the piping is the same size and I don't see the difference by the way this is drawn.

By the way, and again I don't have the details but per this drawing, why is there even a PSV on the steam piping, not to mention why is there a spare and why is the relief going to a flare; it's steam?

And why is there a PSV on the reboiler shell, the vent line is locked opened to the column and thus is protected by the column PSV (if this is because your company standard is to include a PSV at the bototm of a column as well as the top, then I actually agree with this; most do not do this and ignore the possibility that the trays may be an obstruction)? And based on the philosophy depicted by this P&ID, why isn't the PSV on the reboiler shell spared, you spared the PSV on the steam?

You are going to put a note that the specticle blinds on the line from the reboiler shell to the column need to be some how locked open also, right?

Does someone own stock in the relief valve and piping manufacturer?

[jprocess]

Dear Phil,
Thanks a lot for your valuable comments.And sorry for my late reply.
Let's review your comments:

I would have to see the isometric or piping drawing to make a real comment. But saying this, it looks like the piping is the same size and I don't see the difference by the way this is drawn.

The by-pass line size is missed.Suppose that the related size be 2".Now we should connect the spare psv tail pipe to subheader and not to by-pass line.Yes?

By the way, and again I don't have the details but per this drawing, why is there even a PSV on the steam piping, not to mention why is there a spare and why is the relief going to a flare; it's steam?

We should protect the tube side against any probable overpressure scenario.So a pasv on steam piping is needed.Do not you agree?
And yes the tube side fluid is LP steam.Do you recommend to route the steam line psv to safe location instead of flare network?

And why is there a PSV on the reboiler shell, the vent line is locked opened to the column and thus is protected by the column PSV (if this is because your company standard is to include a PSV at the bototm of a column as well as the top, then I actually agree with this; most do not do this and ignore the possibility that the trays may be an obstruction)?

Generally I agree with you but for this special case as can be seen tube rupture scenario is probable(Shell side design pressure=4.5 barg and Tube side design pressure=12 barg).So a protective device should be installed on shell.Right?

And based on the philosophy depicted by this P&ID, why isn't the PSV on the reboiler shell spared, you spared the PSV on the steam?)?

There is a common sparig philosophy which is proposed by Technip company that for fire psv we do not consider any spare but for non-fire psv we should consider a spare psv.Maybe this was the reason.
As I told you in my other topics we have not access to basic engineer to ask about these problems.

You are going to put a note that the specticle blinds on the line from the reboiler shell to the column need to be some how locked open also, right?

Could you please explain more about your question?As I checked the mentioned line have a blind and not a spectacle blind.

Does someone own stock in the relief valve and piping manufacturer?

Not yet!

[pleckner]

The by-pass line size is missed.Suppose that the related size be 2".Now we should connect the spare psv tail pipe to subheader and not to by-pass line.Yes?

Sorry, but I don't know if we are talking about the same thing. I still don't see any difference in the way this is piped. Perhaps a quick hand sketch of just this area would be a good idea with the lines in question identified.

We should protect the tube side against any probable overpressure scenario. So a pasv on steam piping is needed.Do not you agree? And yes the tube side fluid is LP steam. Do you recommend to route the steam line psv to safe location instead of flare network?

The reboiler shell appears to be designed for 5 barg and the tubes side is designed for 12 barg. What scenario do you see that will over pressure the steam side? There is no code that requires protection of the tube side of a S&T heat exchanger unless that side is part of a larger system that falls within ASME BPV sections. And of course, the scenario has to be credible. And yes, if you want to aid in the annual income of the PSV vendor, then pipe the relief to a safe location, it does not belong in your flare system.

Generally I agree with you but for this special case as can be seen tube rupture scenario is probable(Shell side design pressure=4.5 barg and Tube side design pressure=12 barg).So a protective device should be installed on shell.Right?

Because of this horrible design a tube rupture scenario is credible but you don't need an individual PSV on the shell if the shell can be protected by the PSV on the column as yours can be. The vapor line is wide opened to the column even though it has a block valve because you show this valve as being locked opened. The only caviet would be if the vapor line from the shell to the column is too small to handle the relief (very doubtful).

Personally, I would like to see that PSV on the shell but I haven't done a project yet that would allow me to it with this type of arrangement. The clients seem to feel the PSV at the top of the column, sized correctlyl, is all that is needed. My argument is that the trays (or packing) can not only act as a blockage but adds a lot of pressure drop. The way around the pressure drop issue it is to lower the set pressure of the PSV at the top of the column, which is what I do.

You mention that Technip is or was the contractor on record for these designs. The last I heard they are still around. I would have your managment talk to them about many of the issues and not accept no for an answer. Besides, if they sized the PSVs then they must have the calculatiions and scenarios available. There has to be a name attached to these documents. There had to be a company project manager associated with this project. There are many ways to get your questions answered. It may not be easy but with enough persistance, you should be able to succeed.

That symbol shown on the line, call it a blind, call it a spectable blind, I don't care. It needs to remain opened at all times without being accidentally put into position to block the line without adminisrative procedures being followed. If not, then throw away everything I've said about the PSV on the shell and make sure it gets installed.

[jprocess]

Dear Phil,
Thanks a lot for your valuable comments.

Sorry, but I don't know if we are talking about the same thing. I still don't see any difference in the way this is piped. Perhaps a quick hand sketch of just this area would be a good idea with the lines in question identified.

Yes we are talking about the same thing.I repeat my question that if we suppose that the by-pass line size be 2" then we should connect the spare psv tail pipe to subheader instead of by-pass line?

The reboiler shell appears to be designed for 5 barg and the tubes side is designed for 12 barg. What scenario do you see that will over pressure the steam side?

Blocked outlet scenario !

There is no code that requires protection of the tube side of a S&T heat exchanger unless that side is part of a larger system that falls within ASME BPV sections.

The tube side is a part of steam distribution network.But I wonder about this sentence of you!
Could you please explain more about it?Generally protection of tube side is not a necessity?

The only caviet would be if the vapor line from the shell to the column is too small to handle the relief (very doubtful).

The line size is 12" and is not small but you believe that should be checked for governing relieving load.Right?

Personally, I would like to see that PSV on the shell but I haven't done a project yet that would allow me to it with this type of arrangement.

Why do not they allow you?

[pleckner]

Yes we are talking about the same thing.I repeat my question that if we suppose that the by-pass line size be 2" then we should connect the spare psv tail pipe to subheader instead of by-pass line?

You have a relief line between the TW and reboiler steam inlet. Two PSVs are installed on this line. On each tail pipe there is a locked-opened block valve and a reducer then they combine into a common line that is indicated to go to the flare header (FLW). I see nothing wrong with this. So, maybe I'm mis-interpreting what your "bypass line" is?

Blocked outlet scenario !

I have to take your word that there is a blocked-outlet scenario on the steam line because this is not the norm. The line must be designed to handle the steam pressure. If it is not, then yes, you could have a credible scenario but then I would fire the engineer who did this!

The information we are missing with this entire reboiler is a detailed description of how this system operates. Without that, I can't see going any further with this discussion because it doesn't make a whole lot of sense to me.

The tube side is a part of steam distribution network.But I wonder about this sentence of you!
Could you please explain more about it?Generally protection of tube side is not a necessity?

I didn't say it wasn't a necessity to protect the tube side, I only said there is nothing in the ASME codes that says you need to protect the tube side. You still must make sure that the piping is protected per ASME B31.3. The tubes are not an ASME coded vessel, only the shell is, so the protection is for the shell. ASME only says you need to protect the heat exchanger against internal failure ,

"Heat exchangers and similar vessels shall be protected with a pressure relief device of sufficient capacity to avoid overpressure in case of an internal failure."

But again, if the tube side is connecting other pressure vessels, then you must consider the entire system because you are no longer just dealing with piping but vessels as well. In your specific case, the tube side is associated with a piping system so it does not fall under Section VIII but B31.3.

The line size is 12" and is not small but you believe that should be checked for governing relieving load.Right?

That's for you to decide. I did say, "...if the vapor line from the shell to the column is too small to handle the relief" and I added, "(very doubtful)". Meaning, it is very doubtful that it is too small to handle the relief.

Why do not they allow you?

They don't believe it is a necessary expense. They feel the system is well protected with the PSV at the column top.

[fallah]

Yes as well as ease of access to the PSV and block valve. The back pressure contribution of the smaller (rather than the larger) block valve against the PSV is generally too small to worry about in the scheme of things.

Dear pleckner
1-When we have outlet block valve in LO position what is the difference between back pressure in the case of small valve(valva installed before expander) and large valve(valve installed after expander)?
2-In the p&id when the tower design pressure is less than 7 bar as per API 521 do we need to BDV for depressurizing (the BDV on the top of the tower)?
Regards

[pleckner]

@fallah:

1-When we have outlet block valve in LO position what is the difference between back pressure in the case of small valve(valva installed before expander) and large valve(valve installed after expander)?

You will have to go through the numbers to determine this yourself. My statement is purely qualitative, not quantative and was made in the context of the back pressure created from the entire downstream system, not just the block valve.

2-In the p&id when the tower design pressure is less than 7 bar as per API 521 do we need to BDV for depressurizing (the BDV on the top of the tower)?

You do what your company standards tell you to do. Also, I am curious as to what significance 7 bar has on performing depressurizing using a BDV per API 521. Can you quote that section as I couldn't find that in doing a quick scan of the document? Thanks.

[fallah]

[quote]You will have to go through the numbers to determine this yourself. My statement is purely qualitative, not quantative and was made in the context of the back pressure created from the entire downstream system, not just the block valve.[quote]
I mean when in any case we have LO valves what is the difference between backpressure values in the cases the valve before and after expander? I accept your statement generally, but i mean the position of LO valve can't affect the backpressure value.


[quote]You do what your company standards tell you to do. Also, I am curious as to what significance 7 bar has on performing depressurizing using a BDV per API 521. Can you quote that section as I couldn't find that in doing a quick scan of the document? Thanks.[quote]
In API RP 521(at this moment i have'nt for addressing) the final pressure of depressurising in fire case is 50 percent of design pressure or 6.9 barg, whichever is lower.My question is: in P&ID that jprocess is attached to his post, the DP of the tower is 4 barg and why do we need to BDV in this case? Do the issuer of that P&ID work under another standard or practise?
Regards

[pleckner]

I mean when in any case we have LO valves what is the difference between backpressure values in the cases the valve before and after expander? I accept your statement generally, but i mean the position of LO valve can't affect the backpressure value.

Assuming the same flowrate, you don't think a pressure drop through a 3" valve is less than it would be through a 2" valve?

In API RP 521(at this moment i have'nt for addressing) the final pressure of depressurising in fire case is 50 percent of design pressure or 6.9 barg, whichever is lower.My question is: in P&ID that jprocess is attached to his post, the DP of the tower is 4 barg and why do we need to BDV in this case? Do the issuer of that P&ID work under another standard or practise?

First off, I cannot find your absolute values (6.9 barg) stated in either the older API RP 521 or the newer API Standard 521. What you will find is the following (and is practically the same in both editions):

5.20 Vapour depressuring
5.20.1 General

"A vapour-depressuring system should have adequate capacity to permit reduction of the vessel stress to a level at which stress rupture is not of immediate concern. For pool-fire exposure and with heat input calculated from Equations (6) or (7), this generally involves reducing the equipment pressure from initial conditions to a level equivalent to 50 % of the vessels design pressure within approximately 15 min. This criterion is based on the vessel-wall temperature versus stress to rupture and applies generally to carbon steel vessels with a wall thickness of approximately 25,4 mm (1 in) or more. Vessels with thinner walls generally require a somewhat faster depressuring rate. The required depressuring rate depends on the metallurgy of the vessel, the thickness and initial temperature of the vessel wall and the rate of heat input."

As far as why there is a blow down in the design, that will have to be answered by @JP.

[jprocess]

First off, I cannot find your absolute values (6.9 barg) stated in either the older API RP 521 or the newer API Standard 521.

Dear Phil,
The 6.9 barg value is a requirement for fire case in 1997 edition but in 2007 edition only is a requirement for leak detection case and not for fire.

As far as why there is a blow down in the design, that will have to be answered by @JP.

It should be noted that based on our depressuring philosophy document the criteria to make a decision that where to install a BDV or not states that for two phase vessels(like concerened De-Pentaniser) the design pressure should be above 7 barg and P*V(gas)>100 bar.m3 so we recently omitted the De-Pentaniser BDV.But regadless of 7 barg value sometimes we need to depressurize the column down to atmospheric pressure.With new configuration this should be performed with the aid of manual by-pass line.And as I said earlier we have not access to basic engineer to ask about why did he consider a BDV for this column?

[jprocess]

You mention that Technip is or was the contractor on record for these designs. The last I heard they are still around. I would have your managment talk to them about many of the issues and not accept no for an answer. Besides, if they sized the PSVs then they must have the calculatiions and scenarios available. There has to be a name attached to these documents. There had to be a company project manager associated with this project. There are many ways to get your questions answered. It may not be easy but with enough persistance, you should be able to succeed.

I did not say that Technip is or was the contractor.I said that this sparing philosphy is recommended by Technip.I think this is necessary to discuss about this subject in a new topic and I will write about it ASAP.

[jprocess]

I have to take your word that there is a blocked-outlet scenario on the steam line because this is not the norm. The line must be designed to handle the steam pressure. If it is not, then yes, you could have a credible scenario but then I would fire the engineer who did this!

The MP steam operating pressure is 10 barg and based on this value the steam network design pressure is 12 barg.So the tube side design pressure is 12 barg.The line can handle the steam pressure but suppose that blocked outlet scenario occure and pressure goes higher than 12 barg.In this case we should not protect the tube side?

I didn't say it wasn't a necessity to protect the tube side, I only said there is nothing in the ASME codes that says you need to protect the tube side. You still must make sure that the piping is protected per ASME B31.3. The tubes are not an ASME coded vessel, only the shell is, so the protection is for the shell. ASME only says you need to protect the heat exchanger against internal failure ,

"Heat exchangers and similar vessels shall be protected with a pressure relief device of sufficient capacity to avoid overpressure in case of an internal failure."

But again, if the tube side is connecting other pressure vessels, then you must consider the entire system because you are no longer just dealing with piping but vessels as well. In your specific case, the tube side is associated with a piping system so it does not fall under Section VIII but B31.3.

Thanks for clarification.Just one question:
Suppose that you have a shell and tube heat exchanger that for which the low pressure side is tube and its design pressure is lower than 2/3 of shell side design pressure.In this case you do not consider any protective device for tube side?

[jprocess]

Personally, I would like to see that PSV on the shell but I haven't done a project yet that would allow me to it with this type of arrangement. The clients seem to feel the PSV at the top of the column, sized correctlyl, is all that is needed. My argument is that the trays (or packing) can not only act as a blockage but adds a lot of pressure drop. The way around the pressure drop issue it is to lower the set pressure of the PSV at the top of the column, which is what I do.

You mean you do not set the psv at the top of column @ column design pressure?!

[pleckner]

1. Thanks for pointing out where this depressurizing value comes from. I couldn't find it in API RP521 4th edition because it is not given in bar but psig or kpag.

2. How is the pressure in the tubes of this exhanger going to see a pressure greater than 12 barg? The way this is approached in the documentation is that you say you have a blocked-in scenario but there will be no relief.

@JP, from what you write, there should be a relief valve on the steam header system so the steam piping and heat exchanger tubes cannot see a pressure greater than 12 barg. What you have here is a credible blocked-in scenario but with no relief. I repeat, the scenario is credible but there will be no relief required on the tube side as the tubes are designed for the same pressure as the steam and the steam system is already protected by a PSV (or should be).

3.

Suppose that you have a shell and tube heat exchanger that for which the low pressure side is tube and its design pressure is lower than 2/3 of shell side design pressure.In this case you do not consider any protective device for tube side?

Again, if the tube side is part of a system connected to ASME Stamped equipment then yes, you would have to take into account a tube rupture scenario in your PSV sizing calculations for the PSV on the protected equipment. But you are not required by Code to protect the tubes of a shell and tube heat exchanger.

4.

You mean you do not set the psv at the top of column @ column design pressure?!

Let me clarify one thing in my response. I mis-stated when I wrote "...which is what I would do." I really meant, "...which is what I would like to do.". This is a subject opened to interpretation and a lot of argument, one I always seem to loose. As I see it, it depends on the system and on how much pressure drop there would be between the reboiler and the top of the column at the time I need to protect the reboiler. My argument is that if I have to pass vapors from the reboiler through column packing or trays then my reboiler may see pressures exceeding design (or MAWP) plus allowable over pressure and this must not be allowed to happen. Where I can show this to be a real problem then a PSV goes on the shell, end of discussion.

In summary, I agree with whoever showed this PSV on your reboiler shell, keep it there.

[JoeWong]

Just chimed in with additional inputs...

The 6.9 barg value is a requirement for fire case in 1997 edition but in 2007 edition only is a requirement for leak detection case and not for fire.

API Std 521, 2007, Section 5.20, Page 56 stated

"Emergency depressuring for the fire scenario should be considered for large equipment operating at a gauge pressure of 1 700 kPa (approx. 250 psi) or higher. The effect of heat input to process vessels is discussed in 5.15.2 and 5.20.2. Depressuring to a gauge pressure of 690 kPa (100 psi) is commonly considered when the depressuring system is designed to reduce the consequences from a vessel leak."

5.15.2 is referred to fire. Please be noted.


Please also read the explanation note by API in http://committees.ap.../docs/521ti.xls

5.20 Vapour depressuring
5.20.1 General

"A vapour-depressuring system should have adequate capacity to permit reduction of the vessel stress to a level at which stress rupture is not of immediate concern. For pool-fire exposure and with heat input calculated from Equations (6) or (7), this generally involves reducing the equipment pressure from initial conditions to a level equivalent to 50 % of the vessels design pressure within approximately 15 min. This criterion is based on the vessel-wall temperature versus stress to rupture and applies generally to carbon steel vessels with a wall thickness of approximately 25,4 mm (1 in) or more. Vessels with thinner walls generally require a somewhat faster depressuring rate. The required depressuring rate depends on the metallurgy of the vessel, the thickness and initial temperature of the vessel wall and the rate of heat input."

Shall keep in mind that the 1" criteria is only applicable to Pool-fire.

[jprocess]

1. Thanks for pointing out where this depressurizing value comes from. I couldn't find it in API RP521 4th edition because it is not given in bar but psig or kpag.

You are welcome

@JP, from what you write, there should be a relief valve on the steam header system so the steam piping and heat exchanger tubes cannot see a pressure greater than 12 barg. What you have here is a credible blocked-in scenario but with no relief. I repeat, the scenario is credible but there will be no relief required on the tube side as the tubes are designed for the same pressure as the steam and the steam system is already protected by a PSV (or should be).

Sorry but I did not understand your mean from a psv with no relief!

Again, if the tube side is part of a system connected to ASME Stamped equipment then yes, you would have to take into account a tube rupture scenario in your PSV sizing calculations for the PSV on the protected equipment. But you are not required by Code to protect the tubes of a shell and tube heat exchanger.

You do not take into account the tubes as a part of a ASME stamed S&T heat exchanger?

This is a subject opened to interpretation and a lot of argument, one I always seem to loose.

I feel the necessity of a separate topic to discuss about this subject

As I see it, it depends on the system and on how much pressure drop there would be between the reboiler and the top of the column at the time I need to protect the reboiler. My argument is that if I have to pass vapors from the reboiler through column packing or trays then my reboiler may see pressures exceeding design (or MAWP) plus allowable over pressure and this must not be allowed to happen. Where I can show this to be a real problem then a PSV goes on the shell, end of discussion.

So how could not you convince your clients with this logical reason to use separate psvs for reboiler and columns?

[jprocess]

Dear JoeWong,
One new correction in 2007 edition is that API limit the depressuring rules to carbon steel vessels and I do not know about the reason and also what about the other vessels?
This criterion is based on the vessel-wall temperature versus stress to rupture and applies generally to carbon steel vessels with a wall thickness of approximately 25,4 mm (1 in) or more. "

shall keep in mind that the 1" criteria is only applicable to Pool-fire.

And even in this new edition they do not speak explicitly about the depressuring time for vessels with thickness less than 25 mm!
"Vessels with thinner walls generally require a somewhat faster depressuring rate. The required depressuring rate depends on the metallurgy of the vessel, the thickness and initial temperature of the vessel wall and the rate of heat input."

Technip company criteria for concerned vessels:
For wall thickness smaller than 25 mm, following rule shall be applied:
- Wall thickness < 25 mm : 15 minutes minus 3 minutes for each 5 mm decrease in thickness

Total company criteria for depressuring time:
"As a general rule, time to achieve the final pressure level after an EDP has been initiated (1)
shall be, by default:
· 15 minutes for piping and vessels containing hydrocarbon, both gas or liquid
· 8 minutes for vessels containing LPG's or light condensate to avoid the risk of BLEVE.
Note 1: These requirements are applicable only to emergency depressurisation and are not
valid for depressurisation imposed by process reasons

If these criteria were to lead to unacceptably large hydrocarbon disposal devices (either flare or
cold vent) then the two following exceptions could be envisaged:
· Depressuring time for capacities with a wall thickness larger than 25 mm could be
enlarged on the basis of 3 more minutes for every 5 mm in excess of 25 mm and with an
absolute maximum of 30 minutes. This approach is allowed only if one vessel is
concerned (or one group of vessels with similar characteristics served by a common BDV)
and if it is demonstrated that nozzles, instrument tappings and other possible spots where
metal thickness is less than 25 mm do not represent a weak point, likely to leak before full
depressurisation is achieved.
· Credit can be taken for passive fire protection when provided. In this case the time to
achieve full depressurisation shall be as per requirements above, lengthened by the time
it takes for the vessel (or piping) wall to reach its critical temperature (generally 400°C)
and considering the characteristics of the fire to which it will be submitted.
Sizing of BDV's to match the above criteria shall be based on the assumption that during a fire,
all streams incoming and outgoing the system are shutdown'd and all internal heat sources
within the process, if any, have ceased."

[JoeWong]

One new correction in 2007 edition is that API limit the depressuring rules to carbon steel vessels and I do not know about the reason and also what about the other vessels?
This criterion is based on the vessel-wall temperature versus stress to rupture and applies generally to carbon steel vessels with a wall thickness of approximately 25,4 mm (1 in) or more. "
And even in this new edition they do not speak explicitly about the depressuring time for vessels with thickness less than 25 mm!
"Vessels with thinner walls generally require a somewhat faster depressuring rate. The required depressuring rate depends on the metallurgy of the vessel, the thickness and initial temperature of the vessel wall and the rate of heat input."

JP,
Earlier revision, API RP 1997 has stated depressuring time of 15 minutes. This is a hard rule to be complied.

Latest API Std 2007 has provided a more reasonable approach. "Vessel-wall temperature versus stress to rupture". Thicker wall and higher material strength will be less likely to rupture and allow longer depressuring time and lower depressuring rate. There is no hard rule i.e. 15 minitues in latest revision.

[pleckner]

Sorry but I did not understand your mean from a psv with no relief!

Actually, I got to stop responding at night while I'm on some new medication because I'm not writing clearly enough!

There are credible scenarios but not every scenario results in a relief. For instance, blocking in the tubes of this heat exchanger is a credible scenario as it can physically be done but what is the relief? The supply pressure can't exceed 12 barg because that is the system design and the tubes are designed for 12 barg so again, where is the relief? What are you going to set this PSV for, 12 barg, yes? Then where is the driving force to produce a relieving flow? I don't see it.

You do not take into account the tubes as a part of a ASME stamed S&T heat exchanger?

I do not. The stamp is for the shell only. Again, look at the wording I gave you way back in this post, "Heat exchangers and similar vessels shall be protected with a pressure relief device of sufficient capacity to avoid overpressure in case of an internal failure."

The internal failure that can cause over pressure in a shell and tube heat exchanger is the tubes. The tubes are considered the cause of over pressure of the vessel, this being the shell.

So how could not you convince your clients with this logical reason to use separate psvs for reboiler and columns?

I guess because there hasn't been any documented cases of catastrophic failure due to this. That's a hard argument to beat.

[jprocess]

There are credible scenarios but not every scenario results in a relief. For instance, blocking in the tubes of this heat exchanger is a credible scenario as it can physically be done but what is the relief? The supply pressure can't exceed 12 barg because that is the system design and the tubes are designed for 12 barg so again, where is the relief? What are you going to set this PSV for, 12 barg, yes? Then where is the driving force to produce a relieving flow? I don't see it.

In these cases(credible scenario with no relief) you consider a psv or not?

I do not. The stamp is for the shell only. Again, look at the wording I gave you way back in this post, "Heat exchangers and similar vessels shall be protected with a pressure relief device of sufficient capacity to avoid overpressure in case of an internal failure."

I checked all of our NGL project P&IDs and found lots of S&T heat exchangers provided with a rupture disc on tube side to protect against tube rupture scenario.I attached a sample for your kind attention.
As can be seen the shell side design pressure is 40 barg and tube side is 12 barg.I am interested of knowing your valuable comments.
Thanks in advance.

Attached Files

•  Tube_Side_Protection.pdf   80.46KB   62 downloads

[pleckner]

In these cases (credible scenario with no relief) you consider a psv or not?

When analyzing an entire system you have to evaluate all credible scenarios, do you not? Not every scenario has to have a relief so document that you looked at the system, came up with this scenario but you just say no relief is possible. If you can't come up with a relief for a credible scenario then we go back to our trusted thermal relief scenario as the sizing basis and normally put in a minimal sized PSV. Of course this is just for those systems where a PSV is required to comply with ASME Section VIII Divisions and again, the system in question is not.

I checked all of our NGL project P&IDs and found lots of S&T heat exchangers provided with a rupture disc on tube side to protect against tube rupture scenario. I attached a sample for your kind attention. As can be seen the shell side design pressure is 40 barg and tube side is 12 barg. I am interested of knowing your valuable comments.

This is a totally different case than what we've been discussing up to now IF the sea water is discharging into a pipe header system and not to any equipment that falls under ASME Section VIII, Div 1.

(1) The tubes are liquid filled and are cold, the shell being hot.
(2) The rupture disk is not for a tube rupture scenario but for liquid thermal expansion.

BUT....You don't need anything here either because your isolation valves are locked-opened, aren't they? This implies that to close them you require administrative protocols and this should also require the operator to drain the tubes if they are to be blocked-in.

I call this design belt and suspenders...better known as overkill.

[jprocess]

Dear Phil,
Each day this discussion become for me more attractive!

This is a totally different case than what we've been discussing up to now IF the sea water is discharging into a pipe header system and not to any equipment that falls under ASME Section VIII, Div 1.

(1) The tubes are liquid filled and are cold, the shell being hot.
(2) The rupture disk is not for a tube rupture scenario but for liquid thermal expansion.

You first state that the function of RD is to relieve the credible overpressure due to thermal expansion...

BUT....You don't need anything here either because your isolation valves are locked-opened, aren't they? This implies that to close them you require administrative protocols and this should also require the operator to drain the tubes if they are to be blocked-in.

But after that you make the thermal expansion scenario, a doubtful one!
As you know one of the requirements of thermal expansion is that the liquid can be trapped but as you stated this situation is not credible for my case.

I call this design belt and suspenders...better known as overkill.

You believe that despite of this fact the basic engineer have followed a conservative approach?

I attached another sample from a De-Propaniser column for your kind attention.For reboiler the tube side design pressure is 12 barg and for shell side the value is 24 barg.So the tube side design pressure is less than 2/3 of shell side design pressure.What is your idea about this case?We do not need to protect the tube side?Suppose that for any reason(vibration,corrosion,shock and etc) a tube rupture occures.The high pressure(here the shell side) fluid will flow to low pressure side and results in overpressure for low pressure side.So the low pressure side will impose to a pressure higher than its design pressure! I can not understand that why ASME did not include these cases for tube rupture scenario?!

Thanks in advance.

Attached Files

•  Sample.pdf   83.29KB   63 downloads

[pleckner]

You first state that the function of RD is to relieve the credible overpressure due to thermal expansion...

But after that you make the thermal expansion scenario, a doubtful one!
As you know one of the requirements of thermal expansion is that the liquid can be trapped but as you stated this situation is not credible for my case.

NO, I didn't say the scenario is credible at all. I mearly state that it is obvious that the designer is usng the rupture disk to protect against over pressure due to thermal expansion. I do not think it is a credible scenario because the exchanger should be drained per administrative protocols. The designer is going beyond conservatism and is using over-kill.

I attached another sample from a De-Propaniser column for your kind attention.For reboiler the tube side design pressure is 12 barg and for shell side the value is 24 barg.

You are not telling me what the MP steam header is designed for. If it is at this lower design pressure too than you need to read what ASME B31.3 says about exceeding pipe design pressure; it is allowed depending on expected frequency and other factors. I don't have time now to pull the info but I'll try to get it later and post it.

I can not understand that why ASME did not include these cases for tube rupture scenario?!

You are mixing ASME Section VIII, Div. 1 requirements with those of ASME B31.1 requirements. Section VIII is only for equipment and not piping and the tubes of a shell and tube exchanger is not considered integral with the shell but with the attached piping.

[pleckner]

Additonal information to my immediate previous post:

Per ASME B31.3, you may exceed design pressure based on frequency of events and duration but must never exceed test pressure. ASME B31.3 does not specifically talk about scenarios so I would fall on both ASME Section VIII, Div 1 and API Std 521. In this system it appears you may be able to exceed the test pressure of the MP steam pipe if there were a tube rupture so in this system I would say you need a PSV on the MP steam line to limit this pressure excursion.