supersonic flux across a DN 32 valve 1

[engy74]

I have a natural gas pipe line, operating pressure 25 bar-a, operating temperature 60°C, fluid natural gas with density 0.8 kg/Nm3. On this line I have a vent line to the atmosphere with a ball valve DN 32.
Let's suppose this DN 32 valve remains open; in this situation the main pipe line is connected to the atmosphere.
The pressure in the main pipe line is constant due to a gas compressor before the vent.
How can I calculate the gas flow across the DN 32 vent line?
obviously I have a sonic bloc inside the DN 32, but I can not undertand how can I calculate the size of the flow cross section...
Thanks for Your help...
Engy 74

 

[hacksaw]

with a ball valve you do not necessarily have sonic flow in the valve. that can only occur where you have a change in diameter. you may be experiencing sonic conditions at the vent discharge.

 

[Latexman]

First, you probably do not have supersonic flow with standard valve, pipe, and fittings as your title indicates. You need a converging diverging cross section that approaches isentropic conditions for that.

Is the ball valve reduced port or full port? Get the model and serial number off the valve and consult the valve manufacturer.

If it's full port, you probably have sonic flow at just the vent pipe exhaust to atmosphere.

If it's reduced port, you may have sonic flow at the valve and at the vent pipe exhaust to atmosphere.

How can you calculate the gas flow across the DN 32 vent line? There are entire texts dedicated to this problem. I recommend you fine one and start from there. The simplest (assumes a lot of knowledge on the readers part) is Crane Technical Paper 410. You can Google that and get to the website where it's sold. I think ABZ sells it. The holy bible of compressible flow is written by Shapiro. It's out of print, but can be bought on the used boob market. There are many texts in between these two. You'll just have to search for them.





Good luck,
Latexman

 

[quark]

I agree with the above two posts. It may not be the valve which creates the required pressure drop for a sonic flow.

The link below would give you some understanding about choked flow.

http://www.air-dispersion.com/usource.html

Regards,

 

[engy74]

uhm...
dear sirs, I thank YOu for the explanations suggested, but I disagree with YOu.
First: the DN 32 valve permit to vent a line with an operating pressure of 24/25 bar-g: so the ratio between internal and outside pressure is more than the critical ratio = sonic bloc inside the pipe!!!!
The problems becomes then this
1) to define the speed of the gas inside the valve: this is easy, probably, as the speed corresponds to the Mach number for the gas considered
2) to define a cross area: and this may be very difficul!!
Let's suppose in fact to have the valve completetly open, for some reason... which is the cross area? is the area of the full bore (i.e. of the open orifice of the valve)? or, considering that in sonic bloc we have a situation of "vena contracta" (as from Perry) tmy doubt is this : the real cross section is less (=vena contracta due to sonic bloc) the the available cross section inside the valve.
These are, then, my hipoteses... please, if somebody else has another idea, we can discuss together...
Regards,
Engy 74

 

[engy74]

PS:
I have to thank quark: the we paged linked is very usefull, and mentions all the information necessary for my calculation.
The only doubts is about the discharge coefficient, which is arbitrarly placed equal to 0.72
What is it???

REgards,
Engy 74

 

[quark]

Technically speaking, we can say flow is choked across a device if the device creates enough pressure drop to cause a choked flow. i.e the pressure downstream the valve should be atleast 17 bar. But the chances for flow to be choked at the exit are high, for you are discharging to a pressure of 1 bar.

You have to consider the smallest area.

Finally the opening from where the gas exits is similar to an orifice when compared to bigger pipe size and that is why the discharge coefficient of 0.72.

Regards,

 

[Latexman]

2) That's why the equation has the two terms C x A. A is the actual flow area thru the valve. C is the flow coefficient. The product of C x A is the flow area of the vena contracta. C is not always 0.72, it depends on beta and Re, so beware. Of course, this depends on the accuracy needed or the cost of conservatism whether you want to refine the flow estimate or not.

Also, if you'll be very specific on the explanation you disagree with (copy and paste it in), we can explain further or see the error made. It could be as simple as a communication problem.

Good luck,
Latexman

 

[engy74]

thanks, latexman.
As I have to verify the max gas flow in case of accident (release valve DN that suddenly opens with the gas compressor before the valve at full load) I need to understand wich is the worst case: so probably C= 1 (???) may be the most conservative value for calculating the flow across the valve; my further question is this: is this value too much conservative??
reguarding the point on which I disagree
"with a ball valve you do not necessarily have sonic flow in the valve": false: the pressure ratio I have between pressure pipe AND the atmosphere is more than the critical pressure ratio= sonic bloc inside the pipe and the valve on the vent line (hacksaw)
With reference to Your note, latexman, You are right, I don't have Supersonic flow; this is correct.
thanks again and have a nice day