Over sized relief valve with reduced discharge piping

[arbor]

I am new to the pressure relief valve world and having an issue on a relief valve that has already been installed. The valve is a conventional relief valve for pipe thermal relief of vapor and the outlet is piped to a vent. The capacity of installed valve is about 4 times of the flow rate needed. The valve inlet/outlet sizes are 1/2" and 3/4". However, the discharge line is reduced to 3/8" due to limited space. I am concerned about if the valve will still function properly. I understand the valve capacity will be reduced due to the smaller discharge line and the backpressure probably way beyond 10% of the set pressure. Since the valve is oversized, is it possible to use the backpressure correction factor to calculate the reduction on the capacity? Are there other concerns, chattering?

Any education and advice are greatly appreciated.

Arbor

 

[pennpiper]

What is the commodity in the line?
Does the valve vent (discharge) to atmosphere?

Sometimes its possible to do all the right things and still get bad results

 

[arbor]

Yes, the outlet is piped to a vent stack vented to atmosphere.

 

[Latexman]

It has been my experience that the backpressure correction factor, K[sub]b[/sub], is NOT applied to conventional PSVs. The designer varies other parameters to meet the 10% pressure drop rule. If that is not possible, the designer goes to a balanced bellows or pilot operated PSV. K[sub]b[/sub] can be applied to balanced bellows PSVs up to P[sub]B[/sub]/P[sub]S[/sub] = 50%. Pilot operated PSV are designed with K[sub]b[/sub] = 1.

Have you found the outlet pressure drop on the PSV record, or calculated it yourself? Is it a problem or not?

Chattering and fluttering are the concerns. Unstable operation which can lead to failure. Depending on size, pressures, and the fluid handled, sometimes the failure can be catastrophic!

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[arbor]

Latexman - thank you for your reply. I calculated the pressure drop through the discharge line use simplified equation for gas flow,

P1^2-P2^2=25.2*(S*Q^2*Z*T*f*L)/D^5.
p1: upstream pressure psia
P2: downstream pressure psia
S: Specific gravity
Q: gas flow MMSCFD
Z: compressibility factor
T: gas flow pressure
f: friction
L: pipe equivalent length
D: pipe ID

The pressure drop is quite low, below the 10%. However, I am not really confident if my approach is correct,i.e., are there any other items contributing to the backpressure? Is the capacity really OK even the line size is half of the outlet size? Am I missing something?

Thanks

 

[Latexman]

I don't recognize that equation at first glance. Maybe others here do.

For further help we would need a LOT more detail. In fact, we'd need all the detail you needed to do the calc. Why don't you attach that?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[georgeverghese]

There are formulas in Perry Chem Engg Handbook (in the chapter on Transport of Fluids) for isothermal compressible flow, if you want a comparison to the result you get from the formula you posted. From the info you have provided so far, agree that there is nothing more to do than the check that built up backpressure <10 % of SP for a conventional RV - obviously you'll need to run at a low Mach no for these small lines. Also account for all fitting losses , including the exit loss at the elevated vent tip.

If the fluid disposed poses some HSE issues, then all related design mitigations should be in place. The fluid released should also not condense in the exit line at the lowest ambient temp ( and create a liquid static head backpressure) , else you'll need some kind of low point drip leg in the exit line.

 

[Latexman]

If "The pressure drop is quite low, below the 10%", run it through the incompressible flow pressure drop equation (Darcy–Weisbach equation) and see how it compares.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[arbor]

Thank you all for the advices. The equation I used is the simplified form for isothermal compressible gas flow. Both Crane book and Perry book has similar equations. I used 200 psi as the upstream pressure and I only got a pressure drop of 0.7 psi. However, the Darcy equation gives about 150 psi. The gas I have is at cryogenics temperature, 65R. So it is quite dense with a density of about 1 lb/ft^3. Compression factor is about 0.57. Flow rate is about 10 SCFM. The pipe ID is 0.305 in, and the length I tried was 20 ft. Since it is short so I am assuming not enough heat transfer to change the gas temperature.

Thank you,

 

[Latexman]

Instead of assuming an upstream pressure, start at the exhaust (atmospheric pressure, right?) and work backwards to the discharge connection of the PSV. Your first step is to see if you have sonic flow (Mach 1) at the exhaust.

The 0.7 psi vs. 150 psi dP indicates problems in the methodology to me.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[EmmanuelTop]

This is system of two equations (orifice flow, downstream piping flow) and two unknowns (flow, pressure downstream of PSV).

As Latexman says, the receiver (landing) pressure is known. Assume flow, calculate pressure drop (= get pressure downstream of the PSV), then calculate flow for given pressure differential upstream/downstream PSV. The calculated flow must be equal to the assumed flow.

In this way you can confirm what the valve (and the piping) will actually pass through, and then it will be easier to see if other phenomena will occur as well.

Dejan IVANOVIC
Process Engineer, MSChE

 

[Latexman]

Dejan, not really in this case, right? Or, the answer could be maybe, it depends. The orifice flow (I prefer isentropic nozzle flow because that is how it is modeled) is known, as long as all the PSV rules-of-thumb are met. The isentropic nozzle flow is the capacity of the selected PSV at 10% overpressure.

Take that flow at atmospheric pressure and at whatever temperature assumption you are going to make (adiabatic to isothermal), and see if the exit plane velocity calculates to Mach << 1, Mach < 1, Mach 1 , or Mach > 1. That screening will give you a hint how the problem should be handled.

If Mach << 1, probably incompressible. This method may have to be adjusted later depending on the results.

If Mach < 1, Mach 1 , or Mach > 1; definitely compressible flow.

In all cases though, it's easier IMO, to work backwards, from the atmospheric exhaust to the PSV exit nozzle connection.

If it's not possibly to accommodate a conventional PSV and you have to use a bellows PSV, then flow is an unknown since K[sub]b[/sub] can change.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[EmmanuelTop]

Latex - agree. As you say the downstream pressure is known for all cases and it should be the starting point in calculations. I think the piping downstream will see choked flow in any case (the valve is oversized) but this should be verified of course.

Dejan IVANOVIC
Process Engineer, MSChE

 

[arbor]

Thanks so much for the suggestions. I used the downstream pressure and the pressure drop is 12 psi now. Can this 12 psi be considered as the backpressure to the valve? What would be my upstream pressure, 235 or 26 psi? 12 psi sounds much better than 0.7 psi. But does it make sense for 350 ft/s flow velocity through a 20 ft long 3/8" tube? If the backpressure is higher than 10% in reality, can the valve still be used since it is oversized by more than 3 times?

Thank you again,

 

[Latexman]

arbor,

You ask a lot of good questions, some of which are confusing to me, like "What would be my upstream pressure, 235 or 26 psi?". This makes me think I do not fully understand your installation. And, after about 14 posts we still do not know all that is needed to answer on this specific installation. What is the PSV set pressure? I don't think you ever said. What is the normal operating pressure? What is the gas/vapor, or at least the MW? How else will we estimate and double check the density? A dimensioned sketch would be nice. What about that vent mentioned in post # 1? Diameter, length, etc? For us to be the most help to you, we really do need about the same data as you needed to run the calcs. If not, I'd have to be the Amazing Kreskin to get it right; and I'm not. So, if you want some answers to your questions, provide the information, up front.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[georgeverghese]

The relief flow calc should be done at relieving conditions, which you say is fire case thermal ? ( or is this solar radiation thermal?). So can the upstream temp be at 65degR at this condition? The temp downstream of the RV should be estimated using an isenthalpic flash with P1/T1 ( relieving P and T upstream of RV) and estimated P2. Since this is a thermal relief case, I wouldnt be too fussed about an oversized RV.

The isothermal compressible flow calc should be used with known vent tip exit press to derive the pressure at PSV exit. Max line velocity should typically be less than 0.5Mach. Pressure upstream of the RV would be set press plus accumulation permitted by the Code applicable at your location.

 

[Latexman]

arbor,

Just for reference, from my Crane TP410, for 10 cfm of "free air" (60[sup]o[/sup] F and 14.7 psia, which is close to a lot of "standard conditions") the pressure drop for 100 feet of Schedule 40 3/8" pipe is 1.26 psi. 20 feet would be 0.25 psi.

When I don't know much about the fluid being handled, I look at what seems reasonable with air, water, or steam to try to "rough in" the answer. In this case, air seemed most reasonable to me.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[chicopee]

As a rule, PSV outlets should not be restricted by smaller discharge pipe as they should be full sized openings. In the USA, insurance companies and local jurisdiction regulators would ask to replace the discharge pipes to 3/4".

 

[dungprocess]

I guess from vendor data you can easily find the rated flow thru this PSV, did you check if this rated flow causes excessive back pressure on your psv?