20 replies to this topic

[Kryz]

Dear Forum,

I found your opinion very useful. Very good experience.

We have problems with our two lateral lines from PSV. I run Model in Hysys initially and found all lateral line and header size is OK, except two PSV discharge lines.

Then we use our consultant who has license to AFT Arrow, who done more detailed model for those two, as HYSYS can not clearly shows chock points.

Because design of some part plant was done and manufactured by contractor, we get vendor package (skid). Their battery limits include their equipment and also PSV plus short discharge pipe and was done by them. The rest parts of lateral are ours and also flare header, stack, etc.

They undersized discharge line, made it the same as PSV outlet (2 inch). We already fabricated flare header. It was assumed whole line will be the same 2 inch, so it was made tie-in point between lateral and header the same size. It is weldolet, weld to main header 8 inch, then other end of weldolet is connecting directly to 2 inch flange (also weld). We cannot change this part as already fabricated, so we always assuming in our hydraulics 2 inch flange/weldolet going to 8 inch header. If we increasing size to larger we attached reducer 3 inch x 2 inch and include in Arrow model. After that 2 inch we have sudden expansion to 8 inch header.

When I start work for this organisation the task was to perform sizing of all lateral lines and main header. When already some part was fabricated.

2 inch lateral is too small; backpressure is greater than 10% of Set pressure. We are using conventional spring loaded PSV. We investigate to change line size to 3 inch. Initially we change our part to 3 inch, it does not improve, after we change vendor part pipe as well. We still have chocking on weldolet part. Do you think we can accept sonic velocity (mach=1) for just only this fitting. There are two similar PSV with the same problem. One is even more problematic because mass flow is greater by 25%.

We try also when we redesign pipe work, and make interconnection between two laterals. Because both PSV discharge line are routed to tie-in points very close to each other like 2 meters away, we decide to connect them by bypass pipe. We assume always only one PSV discharge to flare. No coincident scenario of two or more PSV open occurs. We want to fix and split the flow, so one PSV discharge to two places in header. Firstly my question is if such solution is acceptable by API standard and good engineering practice. Anyone have seen this in refineries or gas plant. We are deailing with LNG plant.

Second problems is however, we test various configuration, like increasing pipe size from 2 to 3 inch, plus making bypass also 3 inch instead of 2, changing our vendor (in skid) part of PSV discharge to 3 inch as well. Always we get similar answer - Chocking in weldolet 2 inch part. We know that it is bottleneck of the flow.

I wonder then if Forum can accept such design. Please be also be advised that I always look at backpressure. And after increasing pipe size we managed to drop below 10% of Set P. Bypass options decrease even more backpressure, somethiong less than 5%.

So, again question can we accept such design if we are moreover safe in regards of backpressure (no any risk to PSV capacity problem or disk lifting). I read a lot Forum discussion regarding lateral by other authors. Mach = 0.7 is maximum aaceptable, some of you accept 0.9. Our hydraulics is based on exact isometrics. Length and fittings accuracy is say 1 millimetre.

Physically high velocity means vibration problems from mechanical point of view. But from the capacity and flow point of view (hydraulics) it is no problem. You can pass the same amount of flow; the matter is that velocity is very high. Are there any risk of damage pipe by vibration, and this is really not pipe but weldolet - small segment/fitting only. We call this also our bottleneck.

Thanks in advance for any comments.

Kryz.

[JoeWong]

Kryz,
Not sure but unlikely API 521 allow such configuration but it is definitely not a good engineering practice. I have tried to do this once but failed...

I hope flowrate used to estimate flowing velocity was based on selected PSV rated flow...

If both PSVs are within single fire zone, both PSVs may expose to fire and relieve together. Please review this point.

Assuming there is no event that both PSVs relieve together, you may needs to check if flow distribution between 2 outlets will not result more flow one laterial and sonic flow at 2" connection.

If you convinced yourself to procede with your proposal (1 x PSV with 2x 2" outlets), you may needs to reinforce the 2"-to-8" connection parts...Not sure if you can do this as construction are completed.

Previous experience showed that flare piping failed under choking condition generally occur at lateral tap-in point (welded part)...and your weak point is at tap-in point...please reconsider twice if you really wanna to do so...

Failed to find the flare piping failure report, will upload it once i found it.

regards,

JoeWong

[Kryz]

Hi Joe,

Thanks for your notes.

Just advising you that both PSV are not fire case, but full flow. They have both tap in point at flare header very close to each other, but location of PSV itself are far away in plant layout anyway. We definitively excluded scenario if two PSV relief the same time by analysing process basis.

I am sizing this lines based on rated PSV capacity.

At this moment we found solution like disconnect both PSV from flare and discharge them to atmosphere via 3 m pipe with elbow to safe location. However this is interim solution, there is environmental issue obviously. We have done dispersion modelling and it is acceptable from safe point of view.

My idea is probably we should go to bellows option, they can tolerate higher backpressure. However it will not solve the problem with chocking in 2 inch weldolet anyway. And also problem is long delivery time, extra cost plus we have tight schedule (can not wait).

So at this moment I can not find any better solution, our consultant with AFT software either. They offer us bypass option, however that one stil din not solve chocking problem. So we need to think if bellows is better or bypass. Bellows stil require change pipe size from 2' to 3", as bellows PSV has no 2" outlet available for this backpressure.

Kind Regards,

Kryz.

[JoeWong]

Hi Kryz,

Just advising you that both PSV are not fire case, but full flow. They have both tap in point at flare header very close to each other, but location of PSV itself are far away in plant layout anyway. We definitively excluded scenario if two PSV relief the same time by analysing process basis.

Just wanna to remind you that if both located at same fire zone, you may experience both PSVs open simulteanously. The worst is both PSV at it rated flow.

At this moment we found solution like disconnect both PSV from flare and discharge them to atmosphere via 3 m pipe with elbow to safe location. However this is interim solution, there is environmental issue obviously. We have done dispersion modelling and it is acceptable from safe point of view.

You may consider to run a dedicated line to some flare tie-in (if available) near by. Of course this will need you to provide large and longer outlet flare pipe or you may compensate with bellow PSV.

I hope your dispersion has considered zero to maximum wind.

Is the gas MW lower than air MW ? Is there potential liquid formation ? Have you considered radiation in case flame formed due to lightning ? I hope these has been addressed by converting it to normal vent.


No doubt it is a temporary solution. I hope it doesn't turn-out a permanent solution.

As promised, i found the document related to sonic flow failure...Please don't under-estimate the POWER of SONIC FLOW...[attachment=414:InSaleh_...incident.pdf]

JoeWong

[Kryz]

Hi Joe,

Thanks for your educative story about accident happen in Krechba Algeria. That is really what demonstrates the POWER OF SONIC FLOW. Also I remember your suggestion from previous email to make reinforce connection. This is really week point as your report shows.

I spoke with our mechanical engineer. His comments were: "Weld-o-let is considered as reinforced fitting and hence do not require additional reinforcement at branch connection. The problem with not filling the entire weld bevel on a weld-o-let fitting is the creation of a "notch" effect or stress riser. This notch will result in the intensification of thermal/mechanical stresses leading to possible crack initiation in service. But this is just highlighted as care to be taken during welding."

So hi postulated it may give us some safety factor/protection - this is stronger connection than use in Krechba. Additionally we have chocking only in weldolet, when proposed pipe size was increased to 3 inch. The case from Krechba was sonic chocking in undersized pipe segments.

Anyway this is wake up call for us, I knew from my >10 years experience engineering practice that sonic velocity must be avoided. I read it through API 521 more than hundred times. Plus from other sources. But this one is real life example. So we carefully look at 2 inch connection problem again. Main issue that is already fabricated, 2 inch weldolet with 2 inch flange.

In regards of you question of about fire zone. I explained before the contingency for both PSV is not fire. They are size to relief full of the plant. And process design analysis exclude that both discharge the same time. When one is discharging second one will not lift. But please be advised that our bypass option was rejected due to lack of space. It is busy pipe work.

No we are remaining with atmosphere discharge option. And answering to your question regarding dispersion model. Molecular weight is 19.15 and 19.77, therefore lighter than air. No risk of liquid formation. It just natural gas. The modelling conclusion was: "The consequence modelling concluded that the proposed release height of 3 m for the GDS00PSV0106 and GDS00PSV1402 PSVs is appropriate. The releases did not slump below the release height to expose personnel to a flammable cloud at their working height, in the event of an atmospheric PSV release from these two PSVs."

So we see how we go.

Regards,

Kryz.

[JoeWong]

Kryz,

Lookslike you have came to some conclusion...Atm venting...

Just to highlight as you did not response in my earlier post

i) i hope dispersion has considered zero to maximum wind.
ii) Radiation level below acceptable level in case of flame formed due to lightning.

Just recall, you may need to access the noise level as closed system noise level is much lower than open discharge...


If you choose to tie-in 3" discharge piping to 8" header with 2 " neck, i hope the stress due to reaction force does not exceed the piping stress limit (2" conenction). Secondly the there is no issue with the Acoustic Induced Vibration (AIV) (relate to dynamic stress of piping) which normally weak at 8" header.

What's 8" header wall thickness ?


Good luck.

JoeWong

[Kryz]

Hi Joe,

Thanks for comments and questions.

I passed this question to dispersion modelling person, when receive let you know.
I am not expert in this sort of problems.

As as far as wall thickness for flare header. We use 10S schedule. For 200 mm (8 inch) pipe strainless steel the wall thikcness is 3.76 mm

Have you got any opinion about this. Any stress in pipe?

Regards,

Kryz.

[JoeWong]

Hi
If you relieve to ATM, then no issue with the dynamic stress.

In case you tie-in to 8" header...
Maximum allowable SPL for 8" Sch 10S with wt=3.76mm is about 167 dB. Judge from 3" discharge piping (normally small PSV with low rated flow), i believe it shouldn't be any problem.

I can help to make a quick check for you if you can advise me the following info for PSV,

Inlet
W Mass Flowrate, kg/hr
P1 Inlet Pressure (Abs), kPaa
T1 Inlet Temperature, K
MW Molecular Weight

Outlet
P2 Outlet Pressure (Abs), kPaa
T2 Outlet Temperature, K
z2 Outlet Compressibility Factor
k2 Outlet Specific Heat Ratio

Length between PSV and tie-in point.



regards,

JoeWong

[Kryz]

Hi Joe,

Please find enclosed following information:
PSV-0106
Rated capacity 16458 kg/hr
Inlet Pressure (or rather Set pressure) 9308 kPag, or 9409 kPaa
Inlet temperature 27 degree C, or in Kelvin degree = 300.15 K
MW = 19.76

Outlet Pressure at PSV (or backpressure for 3 inch pipe) 1034.99 kPag
(this is assuming we have 3 inch pipe)
Outlet Temperature -25 deg C, or 248.15 K (this after JT effect)
Z factor = 0.9524
K (Cp/Cv) = 1.266
Length 7.4 meters (include fittings like 4 elbows 90 and isolation ball valve)
That PSV more problematic.

I can give you data for second PSV.

PSV-1402
Rated capacity 12996 kg/hr
Inlet Pressure (or rather Set pressure) 8177 kPag, or 8278 kPaa
Inlet temperature 20 degree C, or in Kelvin degree = 293.15 K
MW = 19.15

Outlet Pressure at PSV (or backpressure for 3 inch pipe) 496.68 KPag
(this is assuming we have 3 inch pipe)
Outlet Temperature -30 deg C, or 243.15 K (this after JT effect)
Z factor = 0.9759
K (Cp/Cv) = 1.284
Length 8.7 meters (include fittings like 3 elbows 90 degree and isolation ball valve)

I still did not received yet reply from dispersion modelling person. I chase them up again.

Let me know what you think from data I passed to you.

If you have more question or something is not clear please do not hesitate to ask.

Thanks for you help.

Kryz.

[JoeWong]

Kryz,

Both PSVs will generate about 150-152 dB, as expected lower than 167 dB. No issue from dynamic stress...







regards,

JoeWong

[Kryz]

Joe,

Thanks for your numbers. It looks that we came finally to the right solution.

We can accept therefore 3 inch pipe connected to PSV, and replace existing 2 inch. But keeping existing 2 inch tie-in connection to header as it is, and only add reducer 2 inch x 3 inch.

Eventually if back pressure for PSV-0106 will be higher than 10% then we order bellows and swap.
And PSV-1402 is good as conventional (backpressure 6% now). We can forget now about bypass option for ever.

Finally stress engineer should confirm that design is OK. I already spoke with our mechanical guy, also with external pipe stress engineer, plus your valuable information make me convinced that we should not expect any bad news from mechanical design point of view.

Regards,

Kryz.

[JoeWong]

Kryz,

I thought you are proceeding with the atm venting option ?

Your decision on 2" neck option with expected sonic flow and rely on pipe stress analysis to prove that it is acceptable is really first time (great challenge) for me. You really have to put additional attention to your pipe stress engineer to ensure the pipe stress analysis has no issue with sufficient conservatism...

Good Luck.

JoeWong

[Kryz]

Hi Joe,

Atmospheric option is for now. But it is temporary to avoid any problems.

If we find out from stress engineer, and mechanical guys will not see any issue with 2 inch neck the we reconnect to flare header. It will be our final solution. It will be then more investigation from mechanical point of view. My job is finished, basacilly I need to ensure less than 10% backpressure. We really want avoid bellows (if possible)

So, at this stage we discharge to atmosphere.

I greatly appreciate your help.

I stil did not receive reply from dispersion modelling consultant, when having this anwer I let you know.

Regards,

Kryz.

[JoeWong]

Kryz,
Can't view your response due to recent "IPS" problem at host...


Huh...Awaiting your response on the dispersion...

Please don't limit yourself by considering tie it some point downstream...

Ahh...just recall. PSV discharge temperature is subzero...If you needs to select balance bellow PSV, you may needs to consider balance piston to avoid moisture freezing and PSV stuck...any PSV vendor can advise you...

JoeWong

[Luigi]

Joe, could you post that Algeria accident report again? Seems interesting.

thanks

[JoeWong]

Joe, could you post that Algeria accident report again? Seems interesting.

thanks

Lurgi,
Lookslike the attachment has been removed...not sure the real reason ...If i attach the file again, it may be removed soon...please let me know your email address. I will sent to you.

Those who are interested, you may do the same ...

JoeWong

[bhagat]

Hi
Joe

I am looking forward to have a Algeria accident report.

Email:bhagat.arvind@yahoo.co.in

Thanks and regards.

Bhagat.

[JoeWong]

Hi
Joe

I am looking forward to have a Algeria accident report.

Email:bhagat.arvind@yahoo.co.in

Thanks and regards.

Bhagat.

Hi Bhagat,
I have send a copy to above email address. Enjoy...

JoeWong

[JoeWong]

Dear all,
If you are interested, you may download InSaleh Blowdown Incident...[till May 31, 2007]

JoeWong

[Kryz]

JoeWong I have my reply back form dispersion consultant.

WE to HIM - "A question has been raised regarding the need to assess the thermal
radiation due to a lightening strike during an atmospheric release from a
PSV relieving hydrocarbons? My thoughts are that this should not be
required since a PSV release is a very infrequent safety related event and
the risk of a PSV release coinciding with a lightening strike would be
diminishingly low??"

HE back to US - "Issue of risk to personnel off-site is a non-event as the separation
distances and vertical orientation of the releases means the thermal
radiation will be below harmful levels at boundary (ie less than 4.7
kw/m2).
Risk to personnel on-site, is negligible given as you say due to the two
independent issues having to occur, but also due to the fact that the
personnel on site can evacuate the area of the fire (hence limited
exposure typically set escape failure parameters at 12.5 kw/m2 for 20
seconds), have appropriate clothing on, etc.
Risk to other equipment is the only place this becomes a question and
provided the PSVs discharges are not orientated towards other equipment
will not be a issue."

[JoeWong]

JoeWong I have my reply back form dispersion consultant.

WE to HIM - "A question has been raised regarding the need to assess the thermal
radiation due to a lightening strike during an atmospheric release from a
PSV relieving hydrocarbons? My thoughts are that this should not be
required since a PSV release is a very infrequent safety related event and
the risk of a PSV release coinciding with a lightening strike would be
diminishingly low??"

HE back to US - "Issue of risk to personnel off-site is a non-event as the separation
distances and vertical orientation of the releases means the thermal
radiation will be below harmful levels at boundary (ie less than 4.7
kw/m2).
Risk to personnel on-site, is negligible given as you say due to the two
independent issues having to occur, but also due to the fact that the
personnel on site can evacuate the area of the fire (hence limited
exposure typically set escape failure parameters at 12.5 kw/m2 for 20
seconds), have appropriate clothing on, etc.
Risk to other equipment is the only place this becomes a question and
provided the PSVs discharges are not orientated towards other equipment
will not be a issue."

Kryz,
Radiation due to lightening strike during an atmospheric release from a PSV relieving hydrocarbons to atmosphere is common analysis to ensure the radiation level will not reach harmful level to personel. Thsi level is very much depends on wind speed, wind direction, discharge velocity, etc.

The statement is rather "common" to me. If the consultant is confidence and your budget is allowed, make a quick run to show it definitely no issue.

I hope you have no issue with dispersion...



Good Luck.

JoeWong