API 3% Inlet Loss for PSVs Protecting Piping

[kyu]

API RP 520 Part II para. 2.2.2 recommends that the total non-recoverable pressure loss between the protected equipment and the pressure relief valve should not exceed 3 percent.

I have a situation where I have a pressure regulator and a high-to-low pipe code break at the regulator. A PSV is located downstream of the regulator to protect the (low-class) piping in the event that the regulator fails in the wide-open position.

Clearly, the non-recoverable pressure loss in the inlet piping to the PSV needs to be taken into account. Does the process line--in other words, that segment of piping that starts immediately downstream of the regulator and ends at the PSV inlet piping--also need to be taken into account?

Forum member EGT01 helpfully pointed out to me in another thread that it is important to look at the (static) pressure drop in the process line in order to ensure that the pressure immediately downstream of the regulator does not exceed MAWP+10% when the PSV is relieving. However, static pressure is not the same as dynamic pressure, and it is possible that this criterion can be (largely) satisfied and yet the 3% rule is still violated.

If I can show that the static pressure in the process line is kept at an acceptable level even when the PSV is relieving, can I subsequently remove the process line from the 3% calculation? (In other words, is it okay to perform the 3% calculation for the inlet piping only?)

Thanks to everyone for their time and attention.

 

[EGT01]

kyu,

To help get you started, here's my 3/21/03 post from the other thread thread798-40798 . There is one other post from pleckner dated 3/20/03 relating to the thread you have started here.

EGT01
I totally agree with pleckner's response about starting a new thread (Re: 3% inlet loss when protecting a pipeline) but let's consider the situation when protecting piping such as might occur when a piping spec break occurs at a pressure reducing valve (PRV), say 600# class upstream and 150# class downstream as an example.

In this situation you should also consider the effects of pressure loss from the starting point of the lower class piping to the point of intersection with the relief valve inlet line. In that upstream segment of lower class piping, I think you could argue that the flow would only be the max flow you get through the pressure reducing valve. Of course you need to check the relief valve inlet segment at the rated capacity of the relief valve.

If the relief valve is set at the MAWP of the lower class piping, then the total pressure loss (upstream segment + relief valve inlet) should figure into the 3% allowable loss. If you wanted to exclude the upstream segment from the 3%, I suppose you could have the relief valve set below MAWP so that at relief conditions (Set+10%) the additional pressure loss in the upstream segment doesn't exceed MAWP+10% for lower class piping at its starting point.

Piping codes may actually allow set pressure or overpressure other than I've indicated above but I tried to keep it simple to make a point.

 

[pleckner]

I don't think we have enough information at this point to make a sound judgement. Nothing has been said of the downstream part of this system. Is this line blocked-in at the time the PCV fails? If yes, then the process line might be looked at as a contained "vessel". In this case, my judgement and experience says to start the 3% rule where the flow starts and this is at the inlet line to the PSV (just as you would start at the vessel nozzle).

If there is a vessel downstream which is being protected by this relief valve then that is where the flow calculation MUST start from.

You can't carry this too far or you my not meet a 6% pressure drop let alone a 3% pressure drop.

One last point. There very well may be MANY opinions on this one and because of this I suggest the following to avoid making this thread as long as "War and Peace". There are only two opinions that really matter. One is the plant safety people and Kyu should be asking them, not the INTERNET. The second is the authorities (and I include the insurance people as well).

 

[hacksaw]

This case is for protection of the pipe. Any vessels or equipment connected to the pipe must satisfy their requirements separately.

When a regulator fails, the flow is limited by the wide open valve (relief case). The PSV at its location, must limit the pipe pressure (the so called static pressure) to a 10% accumulation (gas case).

Additionally, in limiting pressure rise to 10%, the pressure drop from the failed regulator to the relief valve must be less than 3% of the static pressure to insure stable operation of the valve.

 

[pleckner]

Well, I'm confused. Is this question being asked by "Kyu" or "hacksaw"? or are you one in the same?

Nevertheless, my opinion still stands. The pressure drop calculation should start at the pipe inlet to the relief valve and does not have to start at the PCV. Also, and again, since there are going to be many opinions on this (and most different), plant safety needs to be the refree and not us folks on the INTERNET.

 

[MarkkraM]

When a PSV begins discharging, the pressure acting on the valve disc is reduced due to the frictional losses in the inlet piping. If this loss is too great the valve inlet pressure may fall below reseating pressure. This will then cause a chattering effect.

I can't think of a situation off the top of my head where you would not design for the boxed in case of the PCV. Here the static pressure is the same in all the piping. To me the pressure sink is at the interface of the PCV and thus pressure loss should be considered from that point.


Pleckner,
Are you claiming we should not be discussing anything safety related on this forum? I would agree that no safety related thread or response should be acted on without proper cross referencing to the appropriate codes or conformation with a plant safety expert. With that said I do believe there is a lot of understanding to gain through discussions in this forum. Perhaps it can trigger people to question/rethink their interpretations of various guidelines (that are possibly incorrect).

 

[kyu]

pleckner was certainly correct in predicting a myriad of opinions...

By the way pleckner, kyu and hacksaw are two different people.

hacksaw was right in stating that the PSV is protecting the piping only, not a vessel. As I stated in my original post, the PSV is there to protect the piping in the event that the PCV fails wide-open. The downstream piping is not blocked-in, nor is there anything downstream of any consequence except for the piping itself.

What makes it very difficult to satisfy the 3% rule if we include all the piping back to the PCV is that the PCV body size is smaller than the pipe. Of course, what that means is that there is an expander in the line between the PCV and the PSV inlet piping. Much of the dynamic pressure drop is associated with the expander.

I would like to assure everyone that I take the safety aspect of my job very seriously indeed. I have discussed this issue at length with several of my co-workers, but we are finding it difficult to come to any kind of consensus. I wanted to know how others might tackle the problem, and so I posted the question in this forum.

 

[pleckner]

To MarkkraM,

I think talking about Process Safety is one of the most important things we can do on these message boards and one in which I love discussing. I was only pointing out that this particular question has the potential of generating a multitude of varying opinions and that the only refree can be the plant safety people. However, in looking back on one of my responses I can see where you are coming from.

To Kyu,

You are indeed doing the correct thing by trying to solicit opinions from others but again, you must satisfy your plant safety people. Saying you discussed this with co-workers is not the same as discussing it with plant safety (unless you are including them). In any event, they have the last say.

Now, as far as the 3% rule goes:

1. It can be "violated" if you can show the pressure drop you calculate will not adversely affect the valve's operation. For an existing installation, I wouldn't hesitate to go up to 3.5%.

2. The calculation should contain only NON-RECOVERABLE losses. For gases, you DO NOT include acceleration losses as these are recoverable losses. Frictional losses in piping (including entrance and exit losses), fittings and valves, equipment and rupture disks if there was any are the only losses you need to consider. This means you should only be using the incompressible flow equation (i.e. Darcy) in calculating the piping losses.

3. Is this PCV a true control valve system, i.e. not a self contained regulator? If so, the affect of reducers are already included in the overall pressure drop and sizing of the control valve. This means you DO NOT take the downstream reducer into account as it is already included in the Cv of the PCV. If this is a self-contained regulator, then it may be a different story.

4. However, my original opinion still stands, the pipe is pressurized as a contained vessel and I would start my calculation at the PSV inlet pipe tie-in with the process line. I guess I'll just have to agree to disagree with some of the other respondents.

 

[EGT01]

Let's consider a simple graphic representation of 2 cases, one with a PSV on a vessel and the other with a PSV on a pipeline.
[tt]
PSV
|
|
-----
| |
------PCV-----| |
| |
-----

PSV
|
|
------PCV-----------------------------
[/tt]
Let's also say the piping segment between PCV and nozzle on the vessel in the first case is the same length as the piping segment between PCV and intersection of the PSV inlet line in the second case.

For the first case, in defense of pleckner's position, I would only check the pressure loss in the piping between the vessel and the PSV to ensure the 3% rule is met and would not consider the segment between PCV and vessel nozzle.

However, for the second case, I've stated that the same segment that I ignored in the first case needs to be considered. Granted this seems somewhat contradictory but my feeling is that in the second case, the pressure at the start of the PSV inlet line is more directly affected by the pressure loss in the pipe segment downstream of the PCV and would be more likely to cause relief valve instability.

If this is true, then I may need to look more closely at the volume of the vessels when considering cases like the first one.

Pleckner does raise an interesting point about not counting acceleration losses which I have heard on a number of occassions but, as embarrassing as it may be, not sure I fully understand.

Kyu,

If this is an existing installation that you are dealing with, then there's probably some interest to use any trick (within reason) to minimize having to modify the existing installation. Have you accounted for the actual angle of diversion in the expansion or are you treating it as a sudden expansion? What's your relief valve set pressure? What's the break down in pressure loss?

If you could live with it, I still think a reduction in relief valve set pressure should allow you to comfortably ignore the pressure drop in the segment downstream of the PCV if you are not willing to accept the position that the 3% rule starts at the intersection to the PSV inlet.

If it is a new installation, I'd say try to cover all bases without going overboard.

 

[MortenA]

One more thing with regards to the 3% rule - if your PSV have adjustable blowdown you might be able to go higher. Your valve vendor will have to confirm this.

Best Regards

Morten

 

[hacksaw]

Greetings Mr. Kyu,

I don't know how the confusion arose, my appologies (and sympathy).

Pleckner is right on the mark about considering only the non-recoverable pressure loss. I was attemptng to emphasize the same point by focusing on the "static pressure".

No doubt that you will have some pressure recovery after the valve and the expander (vena contracta recovery for non-sonic flows). The static pressure at some point down stream of the valve (this is a specific definition of how the pressure measurement is obtained) is important in your relief valve calculation.

I understand the concerns about safety issues(raised by pleckner), but as in any dialog or exchange of views regarding engineering subjects, the final responsibility falls to the design engineer who signs off on the work.

 

[kyu]

Thanks to everyone for your varied and insightful responses.

Quote from pleckner:
"For an existing installation, I wouldn't hesitate to go up to 3.5%"

It is an existing installation, and I too wouldn't hesitate to go up to 3.5%, but it's quite a bit large than that, unfortunately.

Quote from pleckner:
"Is this PCV a true control valve system, i.e. not a self contained regulator?"

As I stated in my original post, it is a regulator, and so I do believe that I should be taking the expander into account separately.

Quote from EGT01:
"Pleckner does raise an interesting point about not counting acceleration losses which I have heard on a number of occassions but, as embarrassing as it may be, not sure I fully understand."

I will try to explain, and others can correct me if I'm totally out to lunch.

Static pressure is probably what we usually think of as pressure: it's the pressure we read off gauges, or the pressure that causes a relief valve to open. However, once a relief valve is open and relieving, there is another form of pressure that serves to keep it open; this other pressure is the velocity head (think of someone aiming a firehose in your direction and how much "pressure" you would experience). The sum of the static pressure and the velocity head is known as the dynamic pressure. In a perfect world, the dynamic pressure will not change because the static pressure and the velocity head are reversible. So if you lose velocity head going through an expander, you haven't actually lost that pressure because it's all been transferred to static pressure; therefore you don't take these losses into account -- this is what pleckner was talking about, and what I alluded to in my original post. Of course, the world is not perfect, and we always lose pressure to friction, and it is these frictional, non-recoverable losses that need to be taken into account. In addition, the effects of elevation are also non-recoverable (as far as the system is concerned), but these effects are usually dominated by the frictional losses and can often be neglected.

Hopefully, I didn't just make things more confusing for you...

Quote from EGT01:
"Have you accounted for the actual angle of diversion in the expansion or are you treating it as a sudden expansion? ... What's the break down in pressure loss?"

Yes, I have accounted for the angle of expansion. Roughly 2/3 of the non-recoverable losses are associated with the expander.

Quote from EGT01:
"If you could live with it, I still think a reduction in relief valve set pressure should allow you to comfortably ignore the pressure drop in the segment downstream of the PCV if you are not willing to accept the position that the 3% rule starts at the intersection to the PSV inlet."

I'm probably missing something obvious (my apologies) but I don't see how this helps me. By reducing the RV set pressure, the flow through the failed regulator will either increase (if the flow is subsonic across the regulator) or the flow will stay the same (if the flow is sonic across the regulator). In either case, the required relieving flow is at least as big as it is currently, and my frictional losses will not be reduced.

Actually, what I'm thinking of doing is raising the RV set pressure. This is an option, because we think the piping can handle more pressure than what the RV is giving it credit for. I've done some preliminary calculations and it seems that this will reduce the flow through the regulator by an amount sufficient to allow us to use a smaller orifice in the RV. Of course, this reduces its rated capacity and the frictional losses are also correspondingly reduced to a level that (barely) satisfies the 3% rule. I can't bring myself to confine the 3% calculation to the inlet piping only -- obviously, otherwise I would never have spent so much time trying to rationalize myself out of my original position -- and so this might be how I deal with it.

Thanks to all of you.

 

[EGT01]

Sounds like you've found a workable solution but wanted to try to explain my logic for reducing the set pressure

Let's say your system MAWP=150 psig
then allowable inlet loss is 0.03*150=4.5 psi

Now let's say your total pressure loss between PCV and PSV is dPt=10 psi
your pressure loss between PCV and intersection to PSV inlet is 2/3 of total then dP1=10*2/3 = 6.7 psi
this leaves pressure loss in the PSV inlet line as dP2=10-6.7=3.3 psi

Now let's adjust Pset=150-6.7=143.3 psig
Allowable inlet loss is now 0.03*143.3=4.3 psi
Since you have accounted for dP1 by the reduction in set pressure, the only other pressure loss that would affect relief valve operation is the inlet line loss and that would be the only segment that needs to be included in the 3% evaluation.

Recognizing the pressure loss in the PSV inlet line will increase somewhat for a decrease in set pressure, dP2 will still most likely be less than allowed.

The drawback is the potential to relieve a little earlier. Don't think this would work when pressure loss in the PSV inlet alone is 100% of the allowable and may not work at lower set pressures or higher pressure losses in the segment downstream of the PCV.

Of course this is only a small part of the total evaluation that would be required and you would need to check all aspects to make sure it would work. The low pressure system maximum relieving pressure at the PCV outlet doesn't change, still MAWP+allowable accumulation so I don't think required flow changes; a lower set pressure would mean reduced valve capacity for a given %overpressure; etc. The key would be to make sure the pressure at the PCV outlet does not exceed MAWP+allowable accumulation.

 

[SooCS]

Hi engineers,

When I first saw this threat, I thought this a simple solution but surprisingly it turn out to be such a lengthy discussion. But looking into the threat, I may be wrong but it seem you guys are debating like scientists on the accuracy of 3 or 3½ %.

May I suggest you guys look at it in another aspect.

API RP 520 is a recommended practice and instead of debating whether you meet the 3% or not, why not look at what is intended of the 3 % rule.

May I suggest using HIPPS approach.

First, I will categorized the failure modes into 4 groups (5 if you consider chattering is a safety problem; to me, a small PSV can be replaced easily and PSV relieving is a rare occurrence i.e. I assume you have an SDV to protect your system).

Cat 1: The resulting stress will be equal or lower than the specific allowable yield stress. Very unlikely to cause any problem.
Cat 2: The resulting stress will be higher than the allowable yield stress but below the max yield stress of the equipment. May cause only small leak but unlikely.
Cat 3: The resulting stress is such that the equipment may sustain plastic deformation. Low probability of bursting but likely to have small leakage.
Cat 4: The resulting stress is such that the busting of the equipment is likely.

Do a hydraulic calculation of the worse case scenario of your system and see which group does your system fall into. If it falls under Cat 1, is there a real concern?

Waiting you guys to fire me.

 

[pleckner]

I've come up against this situation during detailed design for a number of chemical plants (I'm a process design engineer in an E&C environment) and in all cases, plant safety has approved what I've been saying....no need to include pipe segment from PCV. So, I will just have to agree to disagree.

Oh, my apologies to Mr. Kyu on the regulator. I lost sight of the original post.

 

[MarkkraM]

According to my plant guidelines if the set point is below 340kPag and a vapour service a 5% inlet pressure drop may be used if necessary. I do not know the rationale behind this. Perhaps it has to do with having less energy to cause damage from chattering.

On the argument of where to start applying the rule.
What we need to consider is the change in dynamic pressure at the inlet to the PSV before and after it lifts.


PSV
|
|
-----
| |
(1)------PCV-----| |
| |
-----

Case 1
Before PSV lifts:
The vessel and inlet line to PSV is static while the line from the PCV to the vessel is flowing due to the PCV opening fully.
After PSV lifts:
Inlet line to PSV begins flowing and associated frictional losses reduce the dynamic pressure at the PSV inlet. No change occurs to flow rate in the line from the PCV to the vessel therefore the 3% rule applies only to the inlet piping to the PSV.


PSV
|
|
(2) ------PCV1-----------------------PCV2----


Case 2:
I am not sure if my reasoning is correct but here goes.
We can not claim to know the requirements of PCV2 as it is a control valve and may be fully shut as a response to high pressure in the downstream system therefore in design assume PCV2 is fully shut. With that said, when PCV1 will open, the system of piping between PCV1, PSV and PCV2 will pressurise to the point of the PSV lifting. Once lifted a flow will establish from PCV1 to PSV and thus the pressure loss should be considered from the PCV.