What is the effect of choked flow in flare header system?

[Iradah]

Hello,

I need your feedback about choked flow in flare headers.
I am reading a study on our flare header system at the plant made in 2009. There is a PSV that has a 12” line going into a 16” pipe then into 24” header. The study says that there will be choked flow at the end of the 16” header going into 24” header and recommends to replace that portion with 24”. The recommendation was not implemented because the plant flow went down, however, it is going to increase now.
I do not have flarenet to repeat the study and the report does not show the numbers.
What is the effect of a chocked flow at some part of the header downstream? Will this cause the vessel to be at pressures higher than the PSV setting if the back pressure is high? Can I repeat the calcs without flarenet?

 

[BigInch]

Chances are that the flow will be restricted in one way or another. You really do not want that on a vent or flare line.

I hate Windowz 8!!!!

 

[BigInch]

There are good Chances that the vessel could exceed relief pressure.

I hate Windowz 8!!!!

 

[DSB123]

As Biginch says you cannot physically get any more "fluid" down the pipe so that the pressure in the vessel must increase if the mass flow is greater than the choked mass flow. Also generally under choked conditions the velocity is Mach 1 so you could expect high noise levels which could start up vibrations.

 

[davefitz]

From the description it sounds as if you will either need to re-analyze the system or find the original calculations.

Just because there is expected to be choked flow does not automatically mean that there will be problems at the safety valve. Nearly all safety relief vave installations that use a loose connection between the PSV outlet elbow and the exhaust pan have choked flow at 3 locations: At the relief valve orifice ( which defines the flowrate), at the elbow outlet, and at the outlet of the exhaust stack to atmosphere. As long as the pressure at the inlet to the PSV outlet elbow is less than 45% of the lifting pressure of the PSV ( or as defined by the PSV valve mfr) at max overpressure flowing conditions there will be no reduction in rated relief flow. Refer to asme B31.1 app II (aka Liao -Bechtel method)

"Whom the gods would destroy, they first make mad "

 

[LittleInch]

In your post you say that "The recommendation was not implemented because the plant flow went down". I'm far from clear what that has to do with a flare header design as this is normally independent of flow rate.

Design of flare headers and systems is not a simple operation nor one which should be undertaken lightly as it is ia critical safety item.

I would recommend therefore that you obtain a copy of Flarenet or advise whoever asked you to do the work that you cannot answer any questions as you do not have the data and cannot validate the report, especially if any flows or pressures have changed.

Choked flow in flare system is bad news and should not have been allowed to be present unless somehow the plant flow was linked directly to the flare volume. You and only you know what this PSV is connected, what the rest of the flare header is connected to and whether the flow rate at the end of the 16" header is now higher, lower or the same as the original design.

Whether or not your vessel pressure increases above the PSV set point is dependant on many things, but as said, you should avoid choked flow in such systems.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Iradah]

Thanks all for your valuable feedback.

I am afraid I have to re-calculate which is not a quick task as expected by my managers.

My manager said that I can pick one of the cases mentioned in the report and do calculations in excel sheet. Just pressure drop calcs for several lines connected to each other. He wants to validate the accuracy of the recommendations in the report.

There was a recommendation to replace the tail pipe of one of the PSVs into larger size as there will be choked flow in the piping and I have to check its validity.
I think that I need to recalculate the pressure drop from the flare stack back to the PSV outlet. This requires full information of the header size, routing, elbows and tees. Also, as the choked flow is more likely to happen at elbows and tees where a smaller pipe joins a larger pipe, how can I simulate this in excel sheet or even normal simulation software like Hysis without flarenet.

I was thinking of simulating the flow in the pipe tale only and see what is the minimum outlet pressure from the PSV (which is the inlet to the tail pipe) that is required to flow the required relief load. This is because a choked flow in the tail pipe will require higher pressure from the PSV outlet to get the required relief load flow. Do I need to consider the downstream system for this specific check on the tail pipe choked flow?

 

[Iradah]

To add to LittleInch question about the effect of lower plant capacity on the sizing of the flare header: This is because some relief scenarios are blockage scenarios where the plant inlet will be flared because of the blockage downstream and this is why the relief load is less and hence there is no choked flow now compared to early flow rates.

 

[davefitz]

My first guess is that you must assume all releif valves are open and relieving at their nominal max rate * 1.10 ( due to the 90% flow reduction factor the code applies) , then work backwards from the final exhaust toward each pipe intersection back to the relief valve elbow inlet. If you find that the velocity exceeds mach 1 at any location then you must use the Fanno relationships to determine the inlet pressure to that pipe.

"Whom the gods would destroy, they first make mad "

 

[Iradah]

After three months, I am still not sure of the results I got.
I worked the flare header pressure drop calculation from the stack backward to the PSV outlet. I divided the system into sections and used Weymouth equation for gas pressure drop calculations. However, I could find good way to calculate the pressure drop at tees (e.g.; when 6" enter 12"). Shall I use the equation KV^2/2? K is the fitting head loss, but this correlation is derived from Bernoulli which is for incompressible fluid.

Does anyone know which correlation is good for gas pressure drop at fittings? the chocked flow is likely to happen at these locations.
also, I think as was mentioned by litleinch that the PSV will flow when it opens is independent on the plant current throughput, so we still have the choked issue although the plant is operating at lower capacity.

I cannot tell the management only that choked flow is not acceptable. The manager told me that this is not enough and I am supposed to tell them what risks we are exposed to and what are the recommendations. I have less than 1.5 years experience in process engineering and no senior to guide me, for that I appreciate if you can advise.