33 replies to this topic

[wojtar]

Dear CHeresources Users

 

Could anyone explain to me the following problem:

 

Basing on Vendor's offer for safety valve I'm trying to verify the calculations for discharge piping.

This is due to Customer suggestion that no balanced bellows is required (costs).

 

My calculations show that the backpressure exceeds 10% of set pressure -> balanced bellows required.

 

To simplify the problem I made a simulation of real safety valve (Vendor's offer) together with pressure drop calculation in discharge piping. This is done for 1m long discharge pipe which diameter is equal to safety valve outlet -> just to check what happens if I have only short pipe without any additional pressure drop sources like elbows, expanders, etc.

 

The simulation shows that:

-  flow in such pipe is critical.

-  backpressure less than 10% of set pressure (I adjusted pipe lenght)

- the pressure at outlet is higher than atmospheric even I specify atmospheric at the outlet.

 

My simulation software calculates flow through the system knowing he following:

 

- valve data (coefficient of discharge, flowing area);

- pipe geometry (friction losses);

- fixed pressure at inlet (relieving pressure);

- fixed pressure at discharge pipe outlet (atmospheric).

 

Are my results reasonable or maybe there is something wrong with them?

 

Previously I was using real piping geometry (designed by piping department) as short as possible but quite longer, with two elbows and pipe enlargement. The flow was subcritical (through pipe) but backpressure was higher.

 

Now, the Customer tries to convice me that my calculation is  wrong because it requires balanced bellows which cost too much (there are several identical valves).

 

My explanation that safety goes first seems to mean nothing to them. Have you ever faced such situation??

 

At the end once again my question: is my flowsheet  for with reduced piping (attached) reasonable or not.

I'm trying to verify that the simulation software does its work right or maybe there is some mistake in it.

 

Thanks in advance for replies.

 

 

 

 

 

Attached Files

•  PSV Flowsheet.PDF   20.78KB   204 downloads

Edited by wojtar, 03 December 2013 - 09:13 AM.

[Bobby Strain]

So, what is the basis for the relief rate? Fire? Other?

 

Bobby

[PaoloPemi]

wojtar,

critical flow condition for a gas will depend from density in your case at relieving conditions,

make sure to simulate correctly the PSV as temperature has influence on density (ok. dp is limited),

you may assume a backpressure and calculate t (outlet), since you know mass flux it

is easy to verify critical flow (you need only a EOS to calculate the speed of sound),

at that point you may decide to allow the back pressure to increase until velocity reduces below

critical flow or perhaps you may redefine piping.

[wojtar]

Fire case is the base for required flow.

I guess you have in mind that 21% overpressure is allowable for such case.

I used 10% instead.

 

Anyway the problem is in calculation of the backpressure.

 

The Vendor of specific valve submitted his own calculation which show much lower discharge  pressure drop.

That's because Customer wanted to check me -> they don't want balanced bellows. The reason is price which is high but only 10% of the total safety valves cost.

 

I have a feeling that Vendor's calcs are somehow wrong. They use simple equations for incompressible flow pressure drop.

 

For example:

 

pressure drop for pipe enlargement:

Dp = x  u²/2 x ro

 

where: x - pressure drop coefficient (constant taken from some literature)

           u - velocity

           ro - density (inlet or average i don't know

 

I'm using instead:

 

Dp = (m²/ro₁ s) (ro₁/(ro₂ s) -1)

 where: m - mass flux [kg/h m²]

             ro1 - density of gas at inlet to enlargement

             ro2 - density of gas at outlet from enlargement

             s - area ratio outlet/inlet

 

The results done by hand are very close to those from simulator.

 

Finally, is it possible that the pressure at straight pipe outlet is higher than atmospheric (critical flow in pipe).

Pressure at pipe outlet minus pressure drop in pipe gives us higher back pressure than expected by valve Vendor.

 

The same for case elaborated by me for real piping (enlargement, elbows) where flow is subcritical.

Discharge pipe lenght is not so big -> just 1m of straight vertical pipe and 1,3 m of staright horizontal pipe (plus mentioned 2 elbows and enlargement) -> we had to direct flow from valve to so called "safe location".

 

For that case (not attached here) the total pressure drop and resulting build-up backpressure is higher than calculated by valve Vendor.

I think that is because they don't take into acount the flow is compressible.

[PaoloPemi]

however, considering your pdf, I would say that 1-3-Butadiene at 72.41 C  9.64 bar.a should be very close to liquid state,

in these conditions for pout = 0.8 Bar.g the predicted temperature at PSV outlet is about 45 C,

density about 2.08 Kg/m3 and speed of sound about 230 m/s

from these you should be able to verify the design (see my previous post).

Edited by PaoloPemi, 03 December 2013 - 11:43 AM.

[wojtar]

wojtar,

critical flow condition for a gas will depend from density in your case at relieving conditions,

make sure to simulate correctly the PSV as temperature has influence on density (ok. dp is limited),

you may assume a backpressure and calculate t (outlet), since you know mass flux it

is easy to verify critical flow (you need only a EOS to calculate the speed of sound),

at that point you may decide to allow the back pressure to increase until velocity reduces below

critical flow or perhaps you may redefine piping.

Ok.

From the other hand -> is that backpressure still applicable for critical flow in discharge piping (not in valve) or maybe I should just forget about it . What would happen with my set pressure. I have a feeling that valve would open at lower pressure than expected.

 

I designed discharge piping to avoid sonic velocity (anlargement and higher diameter of discharge pipe). Everything's ok if we use balanced bellows. Back pressure is about 1,2 bar G (set pressure 8 bar g) But Customer/EndClient says that I'm wrong - "everybody knows that there is no pressure in such short pipe directed to atmosphere. Why should we use balanced bellows and spend a lot of money".

 

I could say: "You will spend much more money and we all end in jail if your tank fails". But I'm trying to convince them.

 

So I tried to use discharge pipe with the same diameter as PSV outlet -> flow is critical in some segments of piping and backpressure even higher (if calculated correctly).

 

Then I tried to use bigger pipe than initially assumed -> flow is critical in enlargement, backpressure still high.

 

Have you ever faced a situation when build-up back pressure for pipe of about 2,5 m is as high as 1,2 bar g?

Additionally -> I don't want my pipe to be as large as 12" or 14". Discharge from PSV is 8".

[wojtar]

however, considering your pdf, I would say that 1-3-Butadiene at 72 C  9.64 bar.a should be in liquid state, am I missing something ?

It's close to equilibrium. I'm using SRK EOS and specify in CHEMCAD inlet conditions: (fluid state: gas, and relieving pressure -> temperature is a result).

[PaoloPemi]

correct, I solved at 72 C and my software (Prode Properties) did return liquid,

however at 72.41 C is vapor, see my previous post,

for pure fluids phase equilibria I think both Prode and Chemcad do use vapor pressure correlation, not SRK,

my software predicts 45 C as outlet temperature, this means higher density,

Edited by PaoloPemi, 03 December 2013 - 11:50 AM.

[wojtar]

OK. I'm attaching now my designed system flowsheet (hope not boring all of you).

 

Discharge piping is enlarged here from DN200 (8') to DN250 (10").

Flow is subcritical and higher than required.

Back pressure is calculated (not acceptable to EndClient because of balanced bellows need).

 

The whole piping lenght is not more than 2,5 m.

 

The main pressure drop is on that enlargement (but still subcritical flow > that software generates warnings if critical flow occurs).

 

Is it reasonable or maybe I'm doing some big mistake here?

 

Valve Vendor calculates much lower pressure drops! They are assuming in calculations some constant pressure drop coefficients for fittings.

Attached Files

•  flowsheet2.PDF   16.5KB   82 downloads

Edited by wojtar, 03 December 2013 - 11:49 AM.

[PaoloPemi]

if pipe is relatively short and you have a feeling that dp is correct  I would verify only critical flow condition, that is easy as you need only density and speed of sound.

[latexman]

Please describe PSV - inlet nominal diameter, flow nozzle diameter, outlet nominal diameter.

[fallah]

wojtar,

 

In a choked flow through a PSV the critical flow pressure will exist at the exit plane of the valve flow nozzle while the back pressure is the pressure at the valve outlet flange. You fail to specify the size of the PSV inlet/outlet flanges. If the PSV outlet flange size will be low enough, most critical flow pressure can be transferred to the valve outlet flange as the point of back pressure indication. Then even with a short discharge line the back pressure might be much higher than 10% of the PSV set pressure and to overcome this excessive back pressure you may have to select a larger valve outlet size or balanced bellows valve type...

 

On the other hand, choked point might occur at PSV downstream piping. Then the pressure upstream of the choked point and downstream of the PSV to be increased to provide the conditions for passing required mass flow through the choked pipe. This situation might lead to PSV chattering due to repeated valve closing and opening...Then it might you have to enlarge the PSV discharge line size to prevent excessive back pressure than allowable value at PSV downstream... 

[latexman]

That's why I asked for info. on the PSV.

 

“On certain occasions, the designer can encounter a dilemma when sizing a tail pipe to meet the built-up back pressure limitation for conventional pressure relief valves. Calculations for a tail pipe of the same size as the outlet flange can say that the built-up back pressure exceeds the allowable limit even if the length is reduced to zero."

 

Reference: Guidelines For Pressure Relief and Effluent Handling Systems, Center For Chemical Process Safety of the American Institute Of Chemical Engineers, 1998, pages 38-39.

 

Please describe PSV - inlet nominal diameter, flow nozzle diameter, outlet nominal diameter.

 

[Bobby Strain]

So, you should use 21% as the allowable backpressure. If your calculation for backpressure is less than this, don't worry about the details; move on to the next task. But remember that you, not the vendor, are responsible for calculating the backpressure. If I get some time, I'll check it with my trusty, self developed software used to design hundreds of relief systems. I checked your calc OK. Note I used 100 ft 0f 10" pipe with 2-90deg ells.

 

Bobby

Attached Files

•  BD_PSV_Compressible Flow NEW-IMPROVED1.pdf   4.03KB   115 downloads

Edited by Bobby Strain, 03 December 2013 - 05:26 PM.

[ChemEng01]

Why not calculate your relief rate at 21% overpressure instead of 10%. If the standard your using says use 21% overpressure for the vessel in question why use 10%?

 

From this you will get a lower relief rate. You can size your tailpipe using the smaller required relief rate which hopefully will give you a backpressure less than 10% of set pressure.

 

Also, Your tank must have a huge surface area exposed to fire to give you a relief load that high!!

[wojtar]

Why not calculate your relief rate at 21% overpressure instead of 10%. If the standard your using says use 21% overpressure for the vessel in question why use 10%?

 

From this you will get a lower relief rate. You can size your tailpipe using the smaller required relief rate which hopefully will give you a backpressure less than 10% of set pressure.

 

Also, Your tank must have a huge surface area exposed to fire to give you a relief load that high!!

Well, probably I should have do this (21% overpressure). Since the tank design pressure is 8 bar g. and test pressure is 1.5 x 8 we are still on the safe side. Anyway I was suggested by existing tanks safety valves settings (10% overpressure).

My tank is a sphere 1000 m3 -> required flow was calculated according to API std fire case.

Additional info:

 

Selected valve:

inlet nozzle - DN150

outlet nozzle - DN200

flow area - 8741.6 mm2

coefficient of discharge - 0.801

 

Thanks for all answers.

[wojtar]

wojtar,

 

In a choked flow through a PSV the critical flow pressure will exist at the exit plane of the valve flow nozzle while the back pressure is the pressure at the valve outlet flange. You fail to specify the size of the PSV inlet/outlet flanges. If the PSV outlet flange size will be low enough, most critical flow pressure can be transferred to the valve outlet flange as the point of back pressure indication. Then even with a short discharge line the back pressure might be much higher than 10% of the PSV set pressure and to overcome this excessive back pressure you may have to select a larger valve outlet size or balanced bellows valve type...

 

So, may I assume that such big backpressure is nothing uncommon even for short discharge pipes?

I designed my discharge piping so that no critical flow occurs. Backpressure is high but the use of balanced bellows protects me against that. Am I right?

 

Anyway I'm curious about the results which I presented on the first flowsheet (1 m pipe with diameter the same as PSV outlet flange).

The simulator indicates backpressure at PSV outlet (inlet to discharge pipe) and the pressure at discharge pipe outlet - higher than atmospheric.

Is it also possible? The flow in pipe is indicated as critical.

 

On the other hand, choked point might occur at PSV downstream piping. Then the pressure upstream of the choked point and downstream of the PSV to be increased to provide the conditions for passing required mass flow through the choked pipe. This situation might lead to PSV chattering due to repeated valve closing and opening...Then it might you have to enlarge the PSV discharge line size to prevent excessive back pressure than allowable value at PSV downstream... 

 

Yes. I'm also checking that. Inlet pressure drop is less than 3% of set pressure and no critical flow is indicated.

[PaoloPemi]

to put it shortly it means that you should verify mach number on PSV outlet flange, too,

some manufacturers do verify that mach number < 0.7 at 10% but this doesn't seem the case,

perhaps a DN250 or higher would have been more suitable.

[fallah]

 Selected valve:

inlet nozzle - DN150

outlet nozzle - DN200

flow area - 8741.6 mm2

coefficient of discharge - 0.801

 

Thanks for all answers.

 

 

wojtar,

 

Considering the orifice area, appears the selected valve isn't an API designation one...

[ChemEng01]

Yeah check what code your vessel is to. Usually ASME vessels allow 21% overpressure for fire, however BS code vessels allow 10% overpressure for fire.

[fallah]

 wojtar,

 

My comments as red color...

 

 

 

wojtar,

 

In a choked flow through a PSV the critical flow pressure will exist at the exit plane of the valve flow nozzle while the back pressure is the pressure at the valve outlet flange. You fail to specify the size of the PSV inlet/outlet flanges. If the PSV outlet flange size will be low enough, most critical flow pressure can be transferred to the valve outlet flange as the point of back pressure indication. Then even with a short discharge line the back pressure might be much higher than 10% of the PSV set pressure and to overcome this excessive back pressure you may have to select a larger valve outlet size or balanced bellows valve type...

 

So, may I assume that such big backpressure is nothing uncommon even for short discharge pipes?

It depends on the proper valve selection and discharge line sizing...

I designed my discharge piping so that no critical flow occurs. Backpressure is high but the use of balanced bellows protects me against that. Am I right?

 Using Balanced bellows is a better option...

Anyway I'm curious about the results which I presented on the first flowsheet (1 m pipe with diameter the same as PSV outlet flange).

The simulator indicates backpressure at PSV outlet (inlet to discharge pipe) and the pressure at discharge pipe outlet - higher than atmospheric.

It might multiple choked points to be created along the discharge line...

Is it also possible? The flow in pipe is indicated as critical.

 Yes, of course...

On the other hand, choked point might occur at PSV downstream piping. Then the pressure upstream of the choked point and downstream of the PSV to be increased to provide the conditions for passing required mass flow through the choked pipe. This situation might lead to PSV chattering due to repeated valve closing and opening...Then it might you have to enlarge the PSV discharge line size to prevent excessive back pressure than allowable value at PSV downstream... 

 

Yes. I'm also checking that. Inlet pressure drop is less than 3% of set pressure and no critical flow is indicated.

I meant PSV downstream piping not upsteam one...

 

Edited by fallah, 04 December 2013 - 04:24 AM.

[ChemEng01]

Fallah

 

The area will be the actual orifice area.

 

For a Q orifice valve the API effective discharge area is 7129mm2. The ASME actual discharge area is 8742mm2

[wojtar]

The tank is not according to ASME. Anyway PSVs are of "full nozzle" type.

[fallah]

 

The area will be the actual orifice area.

 

For a Q orifice valve the API effective discharge area is 7129mm2. The ASME actual discharge area is 8742mm2

 

JRudd,

 

If so, as per API 526 there would be no room to increase the size of PSV outlet flange. Then, it might increasing the discharge line size or using the balanced bellows valve be the applicable options...

[PaoloPemi]

to explain my previous post #18 (at last I got some time...)

assuming that DN200 -> 200 mm internal diameter, area 0.0314 m2,

with a mass flux of 21.05 kg/s density 3.85 kg/m3 (T discharge= 45 C, P =0.8 bar.g) ,

volumetric flow 5.46 m3/s and speed of sound 225.5 m/s

mach number should be about 0.8 which is a bit high in my opinion,

(you may repeat the calc's at exact operating point to make sure about that)

Edited by PaoloPemi, 04 December 2013 - 05:52 AM.