9 replies to this topic

[benabed]

Dear all,

The attached photos show the supports and anchors of a 42’’ flare header after a power failure in a gas processing plant. This flare header is in operation for almost 4 years. This damage happens just ahead of the flare stack where there is two 90° change in direction of the piping. It is clear that this was due to the shock forces resulted from the sudden release of a huge amount of gas to flare, but I am wondering how such scenario was not taken into consideration in the design stage of the project and what about relevant codes such as ASME B31.3 in this matter. I am looking forward for the experienced community members to shed some light on these photos.

 

Best regards

[benabed]

Sorry here are the attached photos

Attached Files

•  IMAG0122.jpg   84.18KB   64 downloads
•  IMAG0120.jpg   75.93KB   58 downloads
•  IMAG0115.jpg   96.99KB   49 downloads

[paulhorth]

Benabed,

Why is there no flare knockout drum?

If liquid got into the header and smashed into the piping bends that would cause some damage.

 

Paul

[Art Montemayor]

benabed:

 

My comments on your experience and the photos attached are:

• the photos depict two different flare piping setups.  One is a sole pipe and one is for two, parallel pipes.
• I don't have B31.3 handy or with me, but I don't think B31.3 was intended to safeguard against an improper or inadequate piping support application.
• The piping supports - specifically the one supporting the 90o turn on the sole pipe - look to be laterally damaged and out of the perpendicular position.  Therefore, they should be repaired immediately because their support of the pipe has been compromised.
• The concrete support foundation for the piping suports seems to be very cheap - it doesn't seem to be reinforced and of strong concrete mix.  Lateral forces and subsequent motion has shattered the concrete supports and, as you indicate, are probably the result of a suddent  shock wave caused by a sudden and rapid release of flare gases.
• The detailed piping supports' calculations for this part of your unit should be accessed and reviewed.  I am sure the basic scope of work for the calculations did not take into consideration any lateral forces resulting from a sudden, rapid flare release.  And if that is not the case, then the contractor who installed and poured the piping suports did a cheap and out-of-specification job and the results were not inspected carefully by the client/owner upon accepting the installation.

Before doing anything else, I highly recommend the piping installation be inspected and repaired using the above comments to correctly design the supports - including the concrete foundations.  The corner turn (90o elbow) needs to be structurally analyzed to make sure it is supported properly and reinforced with sufficient strength against a water hammer blow as was obviously experienced (although not originating from water, but from a violent gas stream).

[benabed]

Paul,

 

The flare K.O.D is located approximately 100m up stream of the flare.

 

benabed

[benabed]

Good day,

 

After searching the data of the flare header I found the following simulation results done by the contractor:

  Gas MW=30.72

 ρ= 3.78 kg/m3

Pressure downstream flare K.O.D=1.65barg  

Velocity=180m/sec

Flowrate=1083386KG/hr

Rho x V2=122472 KG /m.sec2

From the above it appears that the velocity criteria fixed by the most conservative standards is exceeded whereas the momentum criteria Rho x V2 is les than 200000. Would you please comment on these values.

 

Regards

[paulhorth]

Benabed,

A velocity criterion typically used for flare headers is:  maximum velocity no more than 0.5 Mach. In your case, you do not tell us the temperature of the flowing gas, but at 20 deg C the sonic velocity would be about 320 m/s.So the velocity of 180 m/s corresponds to about 0.56 Mach which is a little on the high side but not excessive.

I think the real problem is the lack of support against horizontal forces. The force on a 90 deg elbow is given by (rho) x A x v2, which comes to about 11 tonnes in the direction of flow. If this force was applied suddenly it would probably knock the piping off its supports and push them over in the direction of flow, which appears to have happened, judging from your photos.  The designer of the pipe supports seems to have thought its only function is to hold the piping off the ground, with no  support against horizontal forces visible in your photos..

Of course if there was any liquid in the flowing fluid, the force on an elbow could be much higher, due to the higher density.

 

Paul

[benabed]

Paulhorth,

 The temperature of the flowing gas is 44°C so what will be the sonic velocity in this case.

 

[paulhorth]

Benabed,

The sonic velocity is proportional to the square root of the absolute temperature - so at 44 C it is higher than at 20 C, but not much higher.

Try it yourself:

C = sqrt ( k.z R.T/M)

where

C = sonic velocity m/s

k = isentropic exponent

z = compressibility factor

R = gas constant 8314 joule/kg,C

T = absolute temperature K

M = mol wt

 

Paul

Edited by paulhorth, 23 April 2013 - 05:09 PM.

[chemsac2]

Benabed,

 

Mach number of around 0.5-0.6 should not be a problem for a well designed and supported flare header. Considering flowrate, molecular weight and velocity, I feel size of 42" is adequate.

 

You need to post these photos in a piping stress analysis forum for better understanding of which mechanism is playing or rather has played the role in support damage. Once a certain mechanism has been identified by stress engineers, we process engineers can help explore root cause further.

 

Few mechanisms that can affect flare header integrity are:

1) Choked flow somewhere in flare header due to flow more than design flow i.e. flow more than design flow.

2) Temperature of relief stream more than considered for stress analysis

3) Slug flow due to liquid relief or liquid condensed due to contact of hot relief with cold piping

4) Acoustic induced vibration

5) Reaction forces

 

In my opinion reaction forces are responsible for support damage in your case for reasons given by Paul above. For a PSV relief, this reaction force would be transient i.e. till the time flow establishes. For systems with depressurization facility and that too with a sequential depressurization, force would continually change with time due to change in flowrate.

 

Hope it helps.

 

Regards,

 

Sachin