Exactly when are PSV's needed? 1

[djack77494]

I need to move into some gray areas. We have a number of cases where we are flowing through multiple (large) heat exchangers and then into a vessel. I would consider the exchangers to be pressure vessels since they don't meet the exclusion criteria. Some of these circuits are quite complex, with exchangers in series and in parallel. They can often be blocked in with isolation valves that are NOT car sealed, and they are located where fire must be considered.

Most of the exchangers do not have PSV's. Instead, we would often have a PSV at/near the vessel. The vessel and exchangers are protected against "Blocked Outlet" conditions with this PSV, which also protects against fire. (We do add the exchanger areas to the vessel area in sizing.)

I would contend that since there are intervening block valves, each item that could be isolated should have its own PSV for (at least) the fire case. Some say that is not needed, and others contend that if we car seal open the intervening valves, then it would not be needed.

Carrying this concept a bit further, for some systems we have a high pressure fluid on the tube side. I need to consider tube rupture with the higher pressure escaping into the shell. Should I consider each exchanger separately?

I know these are big questions. Thanks for any guidance you can provide.
Doug

 

[pleckner]

Doug: If you don't my 2-cents worth...

and since you already determined that the shells are to be considered pressure vessels, I will direct my answer with that in mind.

Code is very clear on the need for protection (unless your particular municipality doesn't follow ASME or equivalent), and this is independent of the scenario; we addressed this in previous posts. You are correct in that if each exchanger can be isolated from any common PSV, then that exchanger MUST be protected by its own PSV. So I don't see how those who say one is not needed can justify violating code at the most, and sleep at night at the least.

Those who say car sealing the valves open is a valid means to not having to install a PSV for each individual shell are indeed correct.

BUT, and this is a big but, one must then be very careful not to exceed the design pressure of the vessel in question (plus accumulation). Assuming shells in series, the relieving vapor from the furthest shell must travel through several shells and piping before reaching the PSV. This requires pressure drop from the furthest shell to the PSV. If this pressure drop is even somewhat significant, the pressure in the furthest shell during relief can exceed MAWP plus accumulatiion; unless you stagger the design pressure of the shells. The furthest shell gets the highest design pressure and the closest shell gets the lowest. The PSV is then set at the lower design pressure. Alternatively, just set the PSV for what would be the equivalent design pressure of the closest shell if you had actually done so. This will allow for the built-up backpressure against the furthest shell not to exceed that which is allowable.

API RP521 address this and unfortunately, most of us have been violating this. AND this is precisely why I don't like using a common PSV for vessels in series.

Now, if you want to arrange the common PSV so that it is in a parallel configuration, fine. BUT then don't forget to maintain reasonable compliance with the 3% rule by making sure the PSV inlet piping is adequately sized.

Addressing your tube rupture question, as I said earlier, the need for a PSV and its placement is independent of the scenairo you use to size it.

 

[djack77494]

Thank you, Phil. I knew I could depend on you for a high quality response.

 

[CJKruger]

Pleckner/Doug,

1) If fire case is the ONLY overpressure case for an exchanger, then it is allowed to have block valves (that are not locked open and used for isolating the exchanger) between it and the PSV. Refer to Appendix M of the ASME code for more details. In some jurisdictions you have to apply to governing body to use Appendix M.

2) Similarly for tube rupture case. Refer to API 521. And again, you may have to apply to use it.

3) There are a great many plants where multiple equipment are protected by a single PSV (almost all hydrotreaters). Because there are many overpressure cases, no block valves are allowed in these circuits. Some companies stagger the design pressure, and some just use the higher value for all.

4) For tube rupture, the preferred solution is to make the low pressure side design pressure 10/13 of the high pressure side design pressure. The difficulty is in deciding how far to extend this higher pressure, which depends on the type and phase of the fluids.

 

[pleckner]

CJKruger:

Responding to your points (1) and (2): As far as Appendix M goes, I want to make it clear that what it says about isolation valves pertains to all vessels, it does not specifically single out heat exchangers.

I know you said to refer to the Appendix for more details but leaving the sentence hanging the way you did is very dangerous. Those who don't have access to ASME will take what you said and run with it because they won't know that the detail you are leaving out is extremely important and in my opinion, makes the exclusion practically meaningless (I say "practically", not totally).

What Appendix M (section M-5.8 Stop Valves Provided in the Relief Path of Equipment Where Fire Is the Only Potential Source of Overpressure) says is the following:
"Full area stop valve(s) located in the relief path of equipment where fire is the only potential source of overpressure do not require mechanical locking elements, valve operation controls, or valve failure controls provided the user has documented operating procedures requiring that equipment isolated from its pressure relief path is depressured and free of all liquids."

The important part is "... requiring that equipment isolated from its pressure relief path is depressured and free of all liquids." and that there are documented operating procedures to carry this out. This is the same thing as saying I am going to fully drain my equipment. And yes, this would even apply to heat exchangers.

Note that this does not preclude the use of "administrative controls". This is critically necessary especially in this case to prevent the isolation valve from being closed during normal operation.

Your point (3): I just want to clarify that I didn't say you can't have a single valve protecting multiple equipment, I only said I don't like to do it...my preference. But go back and check to see if you stay within reasonable compliance of the 3% rule. Most designers forget about this small detail.

Your point (4): I think we all agree.

General, Appendix M is non-mandatory and the designer should check with the local governing body to see if they do indeed allow its application.

 

[CJKruger]

Pleckner,

1+2) Thanks for quoting M-5.8. When I wrote my reply above I did not have access to it.

I do not think it is "practically meaningless" though. It is an important exclusion, that is frequently used and enable us to treat exchangers similar to some other equipment (such as pumps and unit piping that can be blocked it).

3) As far as I know the 3% rule only apply between the last vessel and the PSV (on say hydrotreaters). Is your interpretation that you have to look at all the equipment protected by that PSV ?

 

[pleckner]

Interesting, I never saw a relief valve on a pump, except for PD Pumps that is. PSVs on pipes, quite frequently.

Unless you are going to perform maintenance, I wouldn't think the rule is to completely drain each piece of equipment within the system each time something goes down, especially for only a brief time period...or perhaps for your facility it is (I'm not very experienced with hydrotreaters)? So that is why I think the exclusion is, practically speaking, practically useless. You still need to protect the equipment from accidental closure of the isolation valve during normal operation. How best to do that than putting in a locked opened isolation valve?

The 3% Rule is there to protect against the pressure at the inlet of the PSV from dropping below the blowdown pressure, causing the valve to close. Let's say you have an exchanger, tank, exchanger and relief valve. If the first exchanger goes into relief then flow starts. It must pass through the tank and the next exchanger to get to the PSV. What is the pressure drop? Can you prevent the pressure at the PSV from dropping to blowdown? If you design the system correctly, staggering the design pressures of the equipment and setting the PSV correctly, then you probably won't have the problem. But as I've said above, I've seen this done wrong in the past. I just bring it up as a precautionary note. Perhaps it is my conservatism but better to check and make sure; it doesn't take much to do so.

An no, from my explanation above of what the 3% rule is for, it has nothing to do with only the last piece of equipment and the PSV. It has to do with making sure the inlet pressure at the PSV doesn't fall below blowdown until the over pressure cause is eliminated.

 

[CJKruger]

pleckner,

Thank you for the comprehensive reply.

Typically PSVs are not used to protect piping against a fire case in refinery process units (their area is just added to the equipment areas). PSVs are however used on piping in offsites or tank farms.

Which is why I object to installing a PSV on the hot side of an exchanger for which fire case is the only case. There are thousands of feet of piping that can be blocked in and no-one cares about fire case, but those same people want to put a PSV on a channel head that only has a few square feet of area.

 

[pleckner]

PSVs on pipes can be for any number of reasons. I'm not a piping expert, I'll leave that up to BigInch if he chimes in. But you cannot exceed the desgin pressure of a pipe so if it takes a relief valve to accomplish this, so be it. And, BigInch and I have had some interesting discussions on the criteria for pressure relief in pipelines. There appears to be some disconnect between ASME B31.x and the BPV Codes.

I understand your point but if the heat exchanger is an ASME Coded vessel, then it must have a PSV associated with it; note I said associated, not necessarily right on the shell as we've already discussed. There is no argument, it must be done if you are to follow Code.

 

[CJKruger]

pleckner,

Yes, maybe my reply wasn't clear. The exchanger must always be protected by a PSV. But if fire case is the only case, then I prefer to have the PSV not directly on the exchanger, but on a vessel in the same circuit. Even if the exchanger has isolation valves that are not locked open.

 

[pleckner]

CJKruger: For the sake of your operators and engineers that may be in the vicinity, I just hope you follow through with ASME requirments about having the procedures in place to drain and depressurize the equipment if not in operation and making sure the isolation valves can't be closed by accident during operation.

It may sound like a highly unlikely thing to happen (close those valves during normal operation) but the people in BP Texas, City didn't think they would or could ever overfill that column that eventually caused the loss of 15 lives and injured over 150 more.

If something happens, I won't have to tell the loved ones why daddy isn't coming home just because I blindly followed a recommendation rather than installing a locked opened valve.

 

[CJKruger]

pleckner,

Yes, we all agree the procedures for draining should be in place.

My only comment is that you seem to imply what I propose for exchangers is unsafe. This is not the case. It is done frequently and is safe.

You also do it (no PSV between block valves) for pipes and centrifugal pumps the whole time, but because ASME doesn't care about them, you ignore it.

 

[pleckner]

For pipes, I'll leave that up to BigInch.

You are 100% correct, I am implying that it is unsafe practice not to lock open the isolation valves. Just because it is always done that way in your neck of the woods, doesn't make it safe; and neither does it make it safe just because it shows up in a section of ASME...that is non-mandatory no less. I wish I was on the governing board of any municipality that allows this practice.

I guess we'll just have to agree to disagree.

 

[CJKruger]

pleckner,

Jip, I think we have to agree to disagree.

And next time you have a centrifugal pump that can be isolated (and almost all can), I expect you to install a PSV to protect the pump and connecting piping against a fire case, else your design is unsafe by your definition. Of course I think it is perfectly safe, but that seems only to be the case in my "neck of the woods".

 

[pleckner]

You're going off topic. We're not discussing blocking in equipment that falls outside ASME scope and really doesn't call for any protection anyways, nor are we discussing piping. We're talking about isolating a pressure relief valve on a piece of equipment that does fall within ASME Section VIII, Div. 1 scope.