Silencer sizing and PSV outlet piping

[processeng01]

Dear All,

I am in the process of developing a spreadsheet and would appreciate help on the following topic:

In case of a high pressure safety valve, say boiler superheater PSV, the flow is choked at vent exit as well as at PSV nozzle opening just like in the example given in Nonmandatory Appendix II of ASME B31.1.

If a silencer has to be added to the mentioned configuration, it starts to get puzzling to me.

For argument sake, the vent exit static pressure (without a silencer) is calculated to be 50 psia.

Now, if I add the silencer to existing vent piping, what do I do? Should I just tell the vendor to design the silencer for 35.3 psid(50-14.7)for the calculated flow, temperature, pressure and be done with it? Or am I totally wrong? Sorry for poor English. I would appreciate any insight.

Thanks

 

[insult2injury]

If this is a boiler safety relief valve, I would assume that it would be an open discharge system with a slip joint, in which case the 50 psia outlet pressure you calculate is most likely incorrect. In this type of arrangement, prevention of blowback at the slip joint is important and the additional flow resistance of the silencer must be added to the vent stack to determine if blowback will occur with the initially assumed vent stack geometry.

I2I

 

[processeng01]

Dear I2I,

First, boiler PSV exhaust piping does not have to be open discharge system. There are a lot of projects where outlet piping is routed to roof without open joint-umbrella drip pan design. (though flexibility should be checked carefully)

Secondly if you check the example mentioned in my previous post, code people calculates the exit of vent stack as 51.4 psia.

Thanks

 

[insult2injury]

For open vent stacks, which are used on the vast majority of boiler SRVs, 50 psia would not be the exit pressure.

I2I

 

[processeng01]

Dear All,

The more I think about my original question, the more I come to believe that silencer can be treated as flow resistance in a subsonic flow, since exit velocity at silencer should be low.

Here is my methodology for a spreadsheet solution. (I would like to see what Experts like Latexman and Pleckner have to say about it)

* Assume a initial Mach no (less than 1) at the exit of silencer. (treat silencer as extension of your vent stack with same diameter and certain resistance)
* Knowing stagnation conditions, calculate T* (for M=1) and T for assumed Mach at exit
* Calculate speed of sound and multiply it by assumed exit M to get velocity at the exit
* Knowing the amount of gas discharged, calculate a density from continuity of mass equation
* By doing try-error find correct Mach number that matches density and flow
* With correct Mach number, find P*
* Knowing allowable back pressure from PSV vendor calculate the ratio of P/P* where P=Patm + Pbackpressure, then using fanno line find what mach no this ratio corresponds to.
* Using fanno line calculate fL/D as a total for silencer and piping between PSV exit and silencer exit.
* Find what is left for silencer after deducting fL/D for piping.
* If pressure drop left for silencer is not enough for insertion loss then choose a bigger diameter, shorter route etc.

 

[Latexman]

I try to avoid silencers on PSV tailpipes because in a good overall system design PSVs are rarely supposed to go off.

However, if a silencer has to be added, I hope you are using the absorptive type due to the high freqency of compressible flow noise. The reflective types are for low frequency noise like a compressor.

I also try to avoid sonic flow everywhere in a relief system, except at the PSV nozzle, but sometimes that gets costly.

You will have to characterize the resistance to flow of the silencer and include it in a new calculation to check if the PSV you use can handle the increased built-up backpressure or still has the desired capacity. I recommend you work closely with a reputable silencer vendor on this. If the silencer in compressible flow service (and it is) has a reduced flow area, the flow can choke (at less than you require) and cause the effective N value (frictional loss parameter) to be much greater than K (incompressible velocity head parameter) of the silencer. You may have to ask for technical support from their engineer to get it right. IMO, this is not an opportunity to save cost and make a "home made" silencer. You have a fairly complex compressible flow problem on your hands that needs expert attention.

Good luck,
Latexman

 

[processeng01]

Thanks Latexman,

Some Silencer vendors have diffuser in their design which is a pipe with several holes as first stage of noise attenuation. I suspect the flow can be choked right there and PSV capacity may be impacted, if not designed carefully.

By reading your post, I had trouble understanding your below statement. Looking at Crane paper, aren't K and N same thing? (f*L/D)

If the silencer in compressible flow service (and it is) has a reduced flow area, the flow can choke (at less than you require) and cause the effective N value (frictional loss parameter) to be much greater than K (incompressible velocity head parameter) of the silencer

 

[Latexman]

N = K if there is no reduction in flow area that results in a vena contracta, which results in a further reduction in effective flow area that is usually corrected by a flow coefficient (C), and no expansion (Y). I suspect you'll have one or the other, and near Mach 1 it's effects are significant.

A reference I use is Noise Control in Industry - A Practical Guide by Cheremisinoff, N.P. © 1996 William Andrew Publishing/Noyes.

Good luck,
Latexman

 

[Latexman]

insult2injury,

"The flow is choked at vent exit". The tailpipe will exhaust into atmospheric pressure (the atmosphere). However just inside the tailpipe opening it is not unreasonable to have 50 psia. At the end of the tailpipe there will be a shock wave where the pressure will drop almost instantaneously to atmospheric pressure and the entropy will increase likewise to form an expanding jet.

Good luck,
Latexman

 

[insult2injury]

Latexman,
The method outlined in B31.1 does not look at the pressure just upstream of the shock so I assume he is saying that the exit is 50 psia.

I2I

 

[Latexman]

I2I,

I don't have B31.1, so all I have to go on on this example is what the OP says. I keyed off of "boiler superheater PSV", "the flow is choked at vent exit", and "the vent exit static pressure (without a silencer) is calculated to be 50 psia". He did say "First, boiler PSV exhaust piping does not have to be open discharge system. There are a lot of projects where outlet piping is routed to roof without open joint-umbrella drip pan design", but I assumed that once the roof mounted tailpipe exhausted, the steam discharged into the atmosphere. I think only the OP can clear this up, so . . . .

processeng01 does the tailpipe exhaust to atmosphere or is it contained somehow and the backpressure on the exit is 50 psia?


Good luck,
Latexman

 

[processeng01]

Dear Latexman

Just to summarize the data in example given in B31.1

set pressure of PSV =910 psig
Calculated pressure at discharge elbow exit =118 psia
Calculated vent stack exit pressure =51.4 psia

In this example PSV discharge pipe is free to elongate inside a larger vent stack. (Drip pan umbrella type) The vent stack is open to atmosphere but it is also choked right at the outlet, therefore Pexit is greater than Pambient.

I2I stated that it should always be equal to atmospheric pressure.

 

[Latexman]

Thanks processeng01. My previous posts pertain to the vent stack exit, where a shock wave will exist.

Good luck,
Latexman