21 replies to this topic

[Bill B]

I am evaluating the pressure drop for several hundred RVs. My understanding of API-520 is that the dP for vapors is computed based on the maximum relief capacity of the valve, not the required rate. I also understand that the inlet dP spec is <3% and the outlet is the Overpressure (e.g. 10% for a conventional valve used in single-valve service, 21% for a fire case). Please confirm this.

Here is the issue - the range of dP is all over the map. In some cases the valve is so oversized that the bp = 60%, but at the scenario rate it is 1-2% - i.e. a smaller valve would suffice.

My questions are:
1. Would you ever replace a RV with a smaller one to eliminate an inlet or back pressure issue? I hate to spend money on this.
2. For existing cases, is there a good rule of thumb to apply on "livable" inlet (say 5%) or back pressure (say 15%).
3. Can anyone tell me how long the use of the maximum capacity has been in use - it appears that the original calcs on a lot of these valves was based on the scenario rate?

Any guidance would be helpful. Thanks

[Lowflo]

The dP limit on the inlet is 3%. For the tailpipe, the % dP is limited to the % overpressure of the valve. In process upset cases that's 10% and for fire it's 21%. Also, if the PSV has a set pressure lower than the MAWP, then the %tailpipe loss can be higher than 10%, provided that it's no higher than the % overpressure.

These pressure drop limitations are all based on the rated flow for the PSV. That's because these rules are for the purpose of ensuring that the valve remains open and stable while flowing at its flowrate. The valve's flowrate (rated flowrate) is the only relevent value for the stability test, unless it's a modulating valve. Since the valve will flow whatever it can flow, that's the rate that you want to use for the inlet and outlet loss calculations. ASME and other application codes all require that these calculations be performed at that flowrate.

As for your other questions:
1. Would you ever replace a RV with a smaller one to eliminate an inlet or back pressure issue? Sometimes we have to install a smaller valve to solve a dP problem. It seems counter-intuitive, but an oversized PSV isn't safe unless it satisfies the stability tests (inlet/outlet loss limits).

2. For existing cases, is there a good rule of thumb to apply on "livable" inlet (say 5%) or back pressure (say 15%). My observation is that most companies allow up to 5% inlet loss for existing PSVs. To my knowledge, there's no consensus on outlet loss limitations. From a severity perspective, inlet losses are the greatest hazard. High outlet losses can cause the valve to close too soon, but equilibrium will be reestablished at a higher pressure. That means the accumulation limits might be exceeded. That's a code violation but it doesn't present as much of a safety risk as excessively high inlet losses. High inlet losses cause chatter. It's easy to think that chatter will have a relatively small affect on capacity, but that can be fatal mistake. A severely chattering PSV will have very little capacity as compared to the expected (rated) capacity of a fully open PSV.


3. Can anyone tell me how long the use of the maximum capacity has been in use - it appears that the original calcs on a lot of these valves was based on the scenario rate? This has been the rule since the inception of these rules. If a relief designer used a different basis, then that was simply an error. That said, there are cases in which one can arguably use the sizing flowrate (required flowrate for a relief scenario). There's no concensus on this but many will argue that it's OK to use the sizing flowrate whenever the fluid is all-liquid. That makes sense to me because the valve obviously can't remain full open in an incompressable service unless the system demands the full flow.

[fallah]

1. Would you ever replace a RV with a smaller one to eliminate an inlet or back pressure issue? Sometimes we have to install a smaller valve to solve a dP problem. It seems counter-intuitive, but an oversized PSV isn't safe unless it satisfies the stability tests (inlet/outlet loss limits).

As far as i know for solving a dP problem we should normally enlarge the inlet/outlet lines sizes or shortening the length of those lines rather than using smaller size PSV.

[fallah]

Here is the issue - the range of dP is all over the map. In some cases the valve is so oversized that the bp = 60%, but at the scenario rate it is 1-2% - i.e. a smaller valve would suffice.

What do you mean by above statement?

Actually,PSV size would be calculated based on relief load and dP in inlet/outlet lines is a line sizing issue.

[rxnarang]

My questions are:
1. Would you ever replace a RV with a smaller one to eliminate an inlet or back pressure issue? I hate to spend money on this. NEVER - the PSV orifice size is based on relieving load scenarios. DO NOT ATTEMPT TO CHANGE PSV SIZE WITHOUT PROPER CALCULATIONS.
2. For existing cases, is there a good rule of thumb to apply on "livable" inlet (say 5%) or back pressure (say 15%). 3 % of set pressure as only frictional losses is a good rule to use based on blowdown. Try to live with it. If it is not possible then consult the vendor for a higher blowdown, which in some cases is adjustable via the blowdown ring. For back pressures stay within the allowable limits. If that is not possible then the valve will need to be derated, or change the valve to balanced bellows to negate the effect of higher back pressure. Use the PSV rated flow for both the calcs. No exceptions
3. Can anyone tell me how long the use of the maximum capacity has been in use - it appears that the original calcs on a lot of these valves was based on the scenario rate? As long as I remember

Any guidance would be helpful. Thanks
[/quote]

3 % of set pressure as only frictional losses is a good rule to use based on blowdown.

Dear Sir,

As per API 520&521 recommendation the 3% pressure drop belongs to "Non Recoverable" pressure drop which in my view is the summation of pressure drop due to friction and elevation change and not only the frictional loss. In some of the cases that I have calculated the importance of static head pressure drop is more vital because of the layout limitations.

Let me know.

[fallah]

in my view is the summation of pressure drop due to friction and elevation change and not only the frictional loss.

It is including frictional loss and dosen't cover elevation change due to its recoverable nature.

[JoeWong]

3 % of set pressure as only frictional losses is a good rule to use based on blowdown.

Dear Sir,

As per API 520&521 recommendation the 3% pressure drop belongs to "Non Recoverable" pressure drop which in my view is the summation of pressure drop due to friction and elevation change and not only the frictional loss. In some of the cases that I have calculated the importance of static head pressure drop is more vital because of the layout limitations.

Let me know.

3% frictional loss is good rule of thumb recommended in most design. However, adjusting blowdown enable a PSV tolerate higher frictional loss. I have previous experience upto 5.5%. However, this exercise shall always involve PSV manufacturer.

Static head would probably affect the set pressure. However may not affect frictional loss. Therefore 3% shall refer to non-recoverable loss.The difference in static pressure between the PSV and the protected system has been accounted for during selection of PSV set pressure. Not sure if this post "Only Consider Non-Recoverable Losses in PSV Inlet Line Loss Determination" will help you.

3% frictional loss is good rule of thumb recommended in most design. However, adjusting blowdown enable a PSV tolerate higher frictional loss. I have previous experience upto 5.5%. However, this exercise shall always involve PSV manufacturer.

Static head would probably affect the set pressure. However may not affect frictional loss. Therefore 3% shall refer to non-recoverable loss.The difference in static pressure between the PSV and the protected system has been accounted for during selection of PSV set pressure. Not sure if this post "Only Consider Non-Recoverable Losses in PSV Inlet Line Loss Determination" will help you.

Dear Sir,
I reviewed your post at mentioned post. I see this logical but in my view my trend is more logical and conservative because during the design stage the process engineer usually is not aware about the psv inlet line route so he may has no sense from the order of pressure drop. So it is logical to specify the psv set pressure as the equipment design pressure (or MAWP if available)and after that whenever he is fed with line isometrics he can check the 3% rule consideing both frictional and static head losses. Any hint?

It is including frictional loss and dosen't cover elevation change due to its recoverable nature.

Why do you consider static head loss to be recoverable in nature?

[rxnarang]

Joe is right. Static head - ( liquids only) is used to adjust set pressure.
Rajiv

[fallah]

Why do you consider static head loss to be recoverable in nature?

Static head loss,contrary to dynamic head loss (friction),as potential energy could be naturally recovered and also set pressure of relevant PSV could be adjusted based on without any negative effect.

Static head loss,contrary to dynamic head loss (friction),as potential energy could be naturally recovered and also set pressure of relevant PSV could be adjusted based on without any negative effect.

How this energy is recovered at psv inlet line?

Joe is right. Static head - ( liquids only) is used to adjust set pressure.
Rajiv

As I asked before what to do if there is not any available data about the inlet pressure drop which is used to adjust the psv set pressure?

[fallah]

Static head loss,contrary to dynamic head loss (friction),as potential energy could be naturally recovered and also set pressure of relevant PSV could be adjusted based on without any negative effect.

How this energy is recovered at psv inlet line?

In liquid relief inlet lines,it is not recovered,but is considered in PSV set pressure finalization.

In fact "recoverable" in this case means a potential energy as static head permanently exists in both static (no relief) and dynamic (relieving) conditions.

[JoeWong]

It is not matter of conservative to consider the static loss. It is matter of illogical to include the static head in the PSV inlet line loss.

For example.
You have a vessel with design pressure of 10 barg. A PSV is protecting this vessel. This PSV is located at XXm (say 10m) above the vessel. The vessel is normally contain water (density = 1000 kg/m3). PSV set pressure would be approx. 9 barg as you have approx. 1 bar static head loss. So that...whenever the vessel experience 10 barg, the PSV disc will see 9 barg and open...When PSV open, water passing PSV and maintain 10 barg in the vessel. Frictional loss due to water flow at the inlet line (say 3% of 9 barg = 0.27 bar) results PSV disc seeing 8.73 barg. The PSV normally still at open position as normal PSV blowdown is approx. 6-7% (PSV for liquid service sometime even higher). Once vessel internal pressure start to decrease to 9.7 barg and PSV disc seeing 8.4 barg (reach PSV blowdown), the PSV disc is reseat and PSV close.

If we follow the way you analyse the system by inclusion of static loss in the inlet line loss calculation...
inlet line loss will be 1.3 bar. It will be more than 3% of 9 barg. Even though you have a perfect pipe without any friction loss, the static loss is always 1 bar which is approx 10% of 9 barg. No matter what you do, you can not meet 3% rule. UNLESS, you mount the PSV on the vessel...

As I told in my previous post I do agree with your justification but you did not answer to my main question:
What to do if the inlet pressure drop can not be estimated?
Consider the process engineer estimate a value for inlet pressure drop and specify the psv set pressure. Relief system design is affected by psv set pressure. Suppose that inlet line pressure drop change based on the isometric drawings, piping stress calculation and any probable change in plot plan. So the psv set pressure should be changed as well and it may be necessary to repeat the psv sizing, flare network sizing and other calculations again. Any hint?

[fallah]

As I told in my previous post I do agree with your justification but you did not answer to my main question:
What to do if the inlet pressure drop can not be estimated?
Consider the process engineer estimate a value for inlet pressure drop and specify the psv set pressure. Relief system design is affected by psv set pressure. Suppose that inlet line pressure drop change based on the isometric drawings, piping stress calculation and any probable change in plot plan. So the psv set pressure should be changed as well and it may be necessary to repeat the psv sizing, flare network sizing and other calculations again. Any hint?

Set pressure of PSV would be equal to MAWP of relevant vessel or lower than that value and isn't considered to be affected by inlet line pressure drop.Thus,PSV set pressure can be fixed regardless of inlet line pressure drop in FEED stage,but the size of that line may be considered as HOLD till detail engineering stage.

Before issuance of AFC isometrics and as per the relevant route to be proposed by piping department,the pressure drop of inlet line would be checked by process department and proper size ,such that meets 3% rule, to be selected.

[JoeWong]

As I told in my previous post I do agree with your justification but you did not answer to my main question:
What to do if the inlet pressure drop can not be estimated?
Consider the process engineer estimate a value for inlet pressure drop and specify the psv set pressure. Relief system design is affected by psv set pressure. Suppose that inlet line pressure drop change based on the isometric drawings, piping stress calculation and any probable change in plot plan. So the psv set pressure should be changed as well and it may be necessary to repeat the psv sizing, flare network sizing and other calculations again. Any hint?

Pressure drop can be estimated with good assumptions at the time of you make it. The difference is how accurate is your assumptions.

Design subject to changes since inception. All design is improved continuously with up-to-date information. No one calculation is once and all. All subject to update. That's the meaning of REVISION. For example, for critical PSV, you may do once during feasibility stage to estimate the flare load. During Conceptual stage, you make another calculation with better inputs. Similarly for FEED and Detail design stage. Similar PSV calculation is updated. When the plant is in operation, the PSV calculation is update due to properties change (change in feed composition). PSV calculation may change due to operating mode change, plant modification, upgrading... code change...It subject to continuous change.

I am not sure how many years of experience you have in this line. You post implies that you are young. This is OK. You can learn and you have time and opportunity to learn. The only advice right in front of you is "DO NOT STOP YOURSELF TO LEARN".
If you think the responses are NON-SENSE, leave it to others.

[Bill B]

I want to express my appreciation for the responses given to my initial question. Even the long series of posts regarding liquid case dP was of value!

That said, I would like to share a related back pressure issue and would like your comments (I apologize if I should have started a new thread).
On several occasions I have run into the situation where the back pressure exceeds 10%, even with a pipe stub on the outlet of the RV. Further analysis confirms that the flow is choked at the outlet flange. So simply increasing the outlet venting line would solve the problem.

Normally the only resolution is replacement with a bellows valve. Another solution involves installing the same body (with same outlet flange) but with smaller orifice, and use the same piping; this is rare solution since the existing valve must be oversized, plus the manufacturer must make a body that houses a smaller orifice.

I guess my problem is why would a vendor sell a product that is hopelessly designed to fail the back pressure specifications? Anyone else encounter this problem? I have experienced this enough times over past couple of months in compressible gas processes that this is certainly not rare - more like highly possible.

[fallah]

I guess my problem is why would a vendor sell a product that is hopelessly designed to fail the back pressure specifications? Anyone else encounter this problem?

Normally you should specify fixed/variable back pressure in your data sheet and vendor would size/select the PSV as per these information.

Thus,your problem should be more clarified.

[rxnarang]

Bill ,

This is an interesting question, whose answer I have been seeking for some time. The closest I have got is as below

a) CCPS guide on Pressure Relief and Effluent does mention this exact problem, and says

Quote

3. In rather rare cases, the calculated back pressure is excessive even for
zero tailpipe length of valve outlet size (or no length value is computed
at the maximum back pressure; COMFLOW reports instead that
specified flow is too high):
A dilemma is encountered.
The calculation model says that the valve outlet area cannot handle
the nozzle flow at the allowable back pressure. This may well represent
an inadequacy of the model rather than of the valve (see
§2.4.2.2.3 for further discussion for both gas/vapor and flashing twophase
flow). The computational problem can be resolved by either:

Selecting another valve (or set of valves in parallel) with higher
outlet vs. nozzle flow area.
• Selecting a valve with a larger outlet for the same nozzle size
(or with restricted lift if the nozzle is larger).

Unquote


I personally think that this is a limitation of compressible flow equations, rather than valve body.

2) Some company I worked for has guidelines, that if the inhouse hydraulic program gives excessive backpressure, then unstandrd PSV size needs to be selected. e.g instead of a 3K4 standard size, choose a 3K6. In practice this valve is either not availble or is too expensive, so the work around is to igonore the choke at the valve outlet.

I hope some PSV vendor will contribute to this topic.

Regards
Rajiv