PSV Inlet Piping dP Question 4

[RVAmeche]

I've been discussing a relief valve installation with a colleague and questions about inlet piping drop came up. In the existing installation, due to the large flow capacity the flow through a reducing tee sees a high pressure drop. However my colleague suggests the dP being reviewed is only between the outlet of the tee (branch to PSV inlet) and PSV inlet.

Please see the attached sketch for reference. Thanks!

 

[The Obturator]

Where does your colleague get this notion from?

The Pressure-relief Valve is there to protect against overpressure occurring in the vessel.

The PRV should not see an inlet pressure drop of more than 3%.

To me that means a pressure drop between vessel and PRV regardless.


*** Per ISO-4126, the generic term
'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[RVAmeche]

In this example it's protecting the header/system, but I agree. Unless you have a pilot PSV with remote sensing, it seems to me that pressure drop through the branch of the tee should be included since the pressure at the PSV inlet will be effected by it.

 

[georgeverghese]

Where is the source of this overpressure? Not shown on sketch. The entire segment from this source should be also checked for permissible built up backpressure due to the high flow, for the case when the PSV has reached full lift at 110% of set pressure. Where necessary, the set pressure should be reduced to account for this built up backpressure.

 

[mk3223]

In this example it's protecting the header/system.....

Should be as ... the header/system... of the pressure source to the PSV inlet.

 

[shvet]

In the existing installation, due to the large flow capacity the flow through a reducing tee sees a high pressure drop. However my colleague suggests the dP being reviewed is only between the outlet of the tee (branch to PSV inlet) and PSV inlet.

Answer depends on (a) valve design and (b) what valve protects - piping on picture or a vessel not shown. Note that ASME B31.3 does not contain 3% dP requirements as 3% relates to equipment as equipment (not piping) is overpressure source. So it is debatable to apply 3% to piping only.

Pressure source is critical as 3% relates to pressure source, not an element protected. This is the important point! as many engineers confused the pressure source and the equipment protected.

API 520-2-2015
7.3.4 PRV Inlet Pressure Loss Criteria
The total nonrecoverable pressure loss between the protected equipment and the pressure-relief valve should not exceed 3 % of the PRV set pressure
...
7.3.5 Background on PRV Inlet Pressure Loss Criteria
... many user companies have accepted PRV inlet losses up to 5 % when determining whether modifications to existing installations were warranted.

Note that capacity (rated/actual/certified etc.) is also an issue related. This factor should be reviewed also.

 

[shvet]

questions about inlet piping drop came up

There is not a simple answer as chattering is not well researched. Be aware - 3% was not designed&tested&accepted for piping. Sound engineering recommends to avoid doubtful/unreliable/unverified/unaccepted design/practice/installation/method, so your situation can not be called inherently safer design.

For info

Hydrocarbon Processing "Relief device inlet piping: Beyond the 3 percent rule"

https://www.eng-tips.com/viewthread.cfm?qid=428995

7.3.5 Background on PRV Inlet Pressure Loss Criteria
Limiting the inlet pressure drop to a specific value may not be sufficient to guarantee PRV stability. Recent research and experience indicate that PRV instability is complex and cannot be attributed to just pressure loss in the PRV inlet piping. Limited testing has shown that in many cases PRVs did not chatter when inlet losses exceeded 3 % of set pressure while in some tests PRVs chattered when inlet pressure losses were less than 3 %. Industry experience has shown PRV failures due to chatter are rare. Many existing PRVs in vapor service with inlet losses greater than 3 % of set pressure have not resulted in loss of containment while performing their function. Inlet pressure loss criterion alone is not sufficient to predict PRV stability. There are additional factors that also need to be considered ...

2.4.2.2.1 Inlet Piping Loss
Excessive friction losses in the piping from the vessel to the typical spring-loaded valve (inlet loss) are a possible cause of chatter. There are no losses when there is no flow (when valve is closed); thus, the full pressure of the vessel is exerted on the disk. When the valve opens, flow is established and friction losses occur. The force of the flowing fluid on the disk is reduced by the amount of these losses. The disk lift may decrease, which results in less flow with reduced losses. The force of the flowing fluid on the disk will thus increase with a resulting increase in disk lift, so the cycle starts again. The resulting valve "flutter" causes some reduction in flow capacity. If the losses are sufficient to decrease the forces to the reclosing level, the disk contacts the seat and this dynamic instability becomes chatter. Possible consequences include reduced flow capacity, increased pressure accumulation in the vessel, damage to the valve seating surfaces, and mechanical failure of valve and piping with consequent loss of containment of process fluids.
The "3% rule" (ASME BPVC, Appendix M) is currently accepted as the criterion for the upper limit on inlet losses to safety relief valves. This rule requires that the nonrecoverable (friction) losses be less than 3% of the set pressure when the valve is operating at nameplate capacity, corrected for the properties of the given fluid (see §2.5.2 for the wording of the ASME Code; also see §2.2.2 of API 520-11). This flow capacity is the "relieving capacity" of the valve at 10% overpressure. Some designers and valve manufacturers follow the more conservative practice of using the best estimate flow rate at 10% overpressure for the loss calculation. This flow is about 10% higher than the relieving capacity (see §2.6). Note that standards for countries other than the USA may specify this higher flow basis (British Standard 6759, for example). When a conventional valve is adjusted to correct for service conditions of constant superimposed back pressure, the nameplate show the actual test pressure (cold differential test pressure) and the set pressure in service. Conventional practice in this case is to compute the inlet loss using the calculated relieving capacity for the stamped psi set pressure rather than relieving capacity under test conditions.
Note that the nonrecoverable pressure loss from the vessel to the valve is less than the pressure drop, since the drop includes the change in velocity head from vessel to valve. This velocity head is recoverable (part of the lifting force on the disk), and thus is not included in the determination of inlet loss. Hydrostatic head of two-phase fluids in the inlet piping is typically not included in the loss for either conventional or balanced valves. This head may remain quite constant during normal valve cycling and thus not affect stability, provided that there is enough time for the fluid to drain back between cycles. Otherwise, the pressure drop required to establish the head might conceivably have the same effect on valve stability as nonrecoverable losses. However, test data are not in evidence to show a deficiency in the conventional practice of neglecting the hydrostatic head. See §3.6.2 for methods of application of this 3% rule. Typically a safety valve comes from the manufacturer with its blowdown set at 7% or more. After allowing for the additional losses in the valve nozzle itself (typically about 3%), the 3% limit on inlet piping loss contains a margin of safety. Somewhat higher values of blowdown may be observed for conventional valves in service conditions of constant superimposed back pressure.
A study of the dynamic response to inlet pressure loss has been performed (Kastor 1986, 1986a, 1990, 1994). The proposed computational model is in general agreement with test results for gas flow. The study concludes that the 3% rule is an oversimplified view of the complex dynamic behavior of a valve. Chatter is not observed at higher loss in certain piping configurations, while chatter can be observed at lower loss levels in other configurations. Guidelines for piping layout and sizing based on this work are yet to be developed and accepted by rule-making bodies. Thus, the 3% rule remains as the accepted good practice.

For new installations, to prevent relief valve chatter and damage, total nonrecoverable pressure loss between equipment or pipeline protected (including pipe entrance loss) and a conventional pressure relief valve inlet shall not exceed 3% of set pressure of the valve for flow corresponding to installed valve area (i.e., rated capacity of the valve). For existing installations of conventional pressure relief valves, the nonrecoverable inlet line pressure loss based on rated capacity of the valve should not exceed blowdown or 5% of set pressure, whichever is lower unless justified by a further engineering analysis.

The maximum allowable frictional pressure drop between the protected equipment and the pressure relief valve inlet flange is 3% of set pressure for set pressures equal to or greater than 50 psig(345 kPa gauge) and 5%of set pressure for set pressures less than 50 psig (345 kPa gauge) for all design contingencies including fire. For remote contingencies, the maximum allowable frictional pressure drop between the protected equipment and the pressure relief valve inlet flange is 4% of set pressure for set pressures 50 psig (345 kPag) or higher and 7% for set pressures lower than 50 psig (345 kPag). The pressure drop limitations include only frictional losses and do not include static head or the effects of fluid acceleration. ...
To avoid this mechanism as the cause of chattering, inlet piping to PR valves should be designed for the lowest practical frictional pressure drop (including entrance loss and piping and isolation valve pressure drop). Experience as well as manufacturers' recommendations dictate an inlet pressure drop of no more than 3% of set pressure at the PR valve rated capacity. This 3% limit needs to be met for PR valves in both vapor and liquid service. In the case of PR valves in low-pressure vapor service where the set pressure is below 50 psig (340 kPa gage), 5% inlet pressure drop may be used if necessary. For remote contingencies, an inlet pressure drop of up to 10% of set pressure is acceptable.

3.4.3.3 Inlet Loss
... To avoid this mechanism as the cause of chattering, inlet piping to pressure relief valves should be designed for the lowest practical inlet pressure drop (including exit loss and piping and isolation valve pressure drop). API RP 520 guidelines as well as manufacturers’ recommendations dictate an inlet pressure drop of no more than 3% of set pressure at maximum valve relieving capacity. This 3% limit is particularly important for pressure relief valves in liquid service. In the case of pressure relief valves in low pressure vapor services where the set pressure is below 50 psig (3.4 barg, 3.5 kg/cm2 G), or for certain valve sizes (particularly larger valves), this may be difficult to achieve and the manufacturer should be contacted for guidance. Loss up to 5% may be acceptable for certain valves.

6.2.1 Inlet piping
... The total pressure drop between the vessel and the relief valve, including the pressure drop due to entrance, contraction, fittings, etc. should be kept below 3% of the PRV set pressure (in gauge units). The pressure loss should be calculated using the rated capacity of the PRV. This pressure drop limitation is related to the PRV blowdown characteristic (6-7%), and is imposed to prevent valve chattering. This may be a critical factor for the set pressure below 50 psig . Some client may require that it be calculated on basis of the actual capacity of the PRV: rated capacity is generally 10% less than actual. The 10% margin is required by ASME VIII.

2.1.2 Inlet piping pressure losses
Hydraulic calculations shall be performed for all non-recoverable inlet line losses at each relevant allowable accumulation. All relief devices, except directly mounted tank pressure vacuum vents and TERVs (where heat gain is via the environment and not via the process), require inlet loss calculations. Based on the rated relief valve capacity at the allowable overpressure, the pressure drop in the inlet piping and fittings for new facilities shall not exceed 3 % of the valve set pressure. Exceptions to this requirement are allowed in the case of a pilot-operated relief valve with a suitably arranged remote pilot sensing connection close to the source of overpressure.

Pressure Drop at Valve Inlet Line
Pressure drop between the protected equipment and the pressure relief valve should not exceed 3% of the set pressure.
...
The application of pilot-operated valves should be considered in the following cases:
- Where the pressure drop in the inlet line of the valve exceeds 3% of the set pressure, since the pilot sensor tapping could be located close to the protected equipment.
...

 

[pierreick]

Hi,
Consider this link:

https://www.linkedin.com/pulse/pressure-relief-why-3-inlet-piping-drop-brett-mahar

Pierre

 

[Latexman]

due to the large flow capacity

Care to quantify this by providing the worst case sizing flow rate and the capacity of the installed PSV?

Good Luck,
Latexman

 

[RVAmeche]

In the vein of chattering, do any of you know if that 3% rule is changing anytime soon?

A few years ago (I think 2017-2018) I had a call with DIERS and they mentioned that they were updating their literature to recommend something more restrictive than 3% to prevent chattering, but as far as I can tell no updates have occurred.

 

[Latexman]

The wheels of progress turn slowly, so the wiggling can of worms that shvet descibed so well is what we have, until someone with authority makes a move. I have not heard of any upcoming changes. ioMosaic has some good white papers on it.

If the PSV lift could be mechanically restricted to a lower capacity, would that help you? IIRC, some vendors can do that on some models, but not all. The PSVs with restricted lift option is definately in the minority.

Good Luck,
Latexman

 

[RVAmeche]

For this example no, I've convinced my colleague that the tee needs to included in the calculations. It looks like it may go to standard tee followed by a reducer instead of a reducing tee.

I'll check out those white papers. Thanks!

 

[Latexman]

Oh, I missed that it was a reducing tee.

Good Luck,
Latexman

 

[shvet]

The wheels of progress turn slowly, so the wiggling can of worms that shvet descibed so well is what we have, until someone with authority makes a move. I have not heard of any upcoming changes.

Practice accepting requires wide industry applied researches, expensive ones. Some of those was performed during 70-80s (see references in sources I have posted above). I have not heard such was performed since 2000s. Time has gone, industry has changed, people gone, focus has moved. It doesn't look like today anyone is ready to spend that much one's time and portion of profits on issues this level. These days focus of money and attention moved to IT and similar and there are no signs situation is going to change.
It is times to interpret known facts, not discover new ones.

 

[RVAmeche]

ioMosaic that Latex mentions has a whitepaper they released in 2021. From a quick skim, it sounds like the answer isn't as easy as "it's updated to the 2% rule" and everything would be fine. So obviously it takes longer to implement a more difficult change and do all the testing.

Hell, it took ASME the better part of 30+ years to update B31.3's SIFs after it was known they had issues. And those were all based on one set of tests done in the 40s.

 

[Latexman]

Yes, some of the more recent PSV inlet dP work is not that simple. Some get into acoustics and flow dynamics. I just hope "at the end of the day" we can have a fairly simple rule that works > 95% of the time.

Good Luck,
Latexman

 

[Chachacha1000]

Following API, you use a rated flow to work out inlet losses to PSV; ie rated flow = required flow times de-rating factor (which is = actual orifice A / required A).

Following ISO 4126, you have to use instead “actual flow”/0.9. Actual flow is defined as capacity calculated with certified discharge coefficient (ie needs vendor data).

Which approach is more conservative to work out inlet losses & why?

Thanks!