4 replies to this topic

[vinod]

Hi

I have a liquid at 10 barg in the tube side of a chiller..........this is being cooled using R22 (in the flashed form form a control valve........R22 in the upstream of the control valve is at 12 barg and being flashed to 1 barg) in the shell side......the MAWP of the shell side is 12.5 bar..........we want to provide a PSV in the shell side of the exchanger.................i have some doubts.........

1)In the case of a tube failure(due to the vibration along the tube sheet) will the tube side fluid overpressuirise the shell(since the shell side MAWP is 12.5 barg and the tube side supply pressure is just 10 barg)

2)In case the tube side outlet is blocked the pressure will reach the pump shut of head 18.5 barg and at the same time if the tube fails(assuming that the pump discharge block is not sensed by the operator and that there is no cntrol system to know it).....during the tube failure will the flowrate be the flowrate from the pump or we need to calculate based on the pressure differential and the size of the tube orifice formed.

3)Other than total tube failure cant we take that due to corrosion a hole is formed on the tube and the fluid escapes through that?

4)What will be the fluid being relieved in case of the tube failure...i mean will it be the tube side fluid or will it be the R22 present in the shell side(is there any differnce in having a liquid in the tube and gas in the shlele .........and that of a gas in the tube and liqyuid in the shell)


Please help

Thanks

Vinod

[Art Montemayor]

Vinod:

In response to your 4 questions I offer my following comments:

1) No. The driving force is in the opposite direction. R22 refrigerant will flow into the tube side.

2) You have described a double jeopardy situation. It is not likely that this situation will occur.

3) Why do you insist on debating about a “partial” flow through the leaking tubes? Are you not interested in protecting the R22 vaporizer shell? You are describing a maintenance leak that should be repaired – not a relief scenario.

4) I consider this a silly question. Are you designing this vaporizer at this time or is it already existing? Please be specific. The quality of the fluid that is relieved would depend on whether the fluids are miscible, where the PSV is located, is the flow 2-phase, etc., etc.. You determine this with your design or where the existing PSV is located. If you are trying to determine how to design the vaporizer (where to place the refrigerant), then state so. Of course there is a difference where you place the fluids. You are supposed to make that decision based on the application and on the fluids. We can’t determine that. We don’t know enough about your application or your scope of work. We don’t even know if this is an existing R22 vaporizer or if this is just the design stage. All I know is that you have a submerged coil in a R22 liquid at -25 ^oC (which you have not pointed out). I hope that your tube side liquid has a freezing point well below -25 ^oC. If it doesn’t, then you have another problem - that of the leaking, tube-side liquid instantly freezing in the shell.

[vinod]

Sir

For the 4th point...............the fluid in the tube side is not miscible with R-22 and the chiller is existing........and we have been asked to check the performance of the PSV .........

Thanks...........i am clear with the 1st 2nd and 3rd point...................but for the "double jeopardy" answer ........isnt it possible ....

Regards

Vinod

[Art Montemayor]

Vinod:

If your R22 vaporizer is existing, then you have the option to locate the PSV where you believe it will function in the best and safest manner. This option is open to you and I would select a position where I can only expect single-phase, vapor flow relief.

By not responding to our inquiries, you are starving us.

Is the R22 shell side at -25 ^oC or not?

Is your R22 vaporizer a kettle type? Is the tube bundle totally immersed? Does it have a required vapor disengagement space above the liquid level?

What is the fluid in the tube side? Is its freezing point below or above the R22 shell side liquid?

Why are you interested in the effects of a leaking tube? Why don’t you just design for the worse case scenario – as you are supposed to?

Did you purchase a complete, engineered chiller system or did you (or your company) design it yourselves? Do you have the engineering calculations for the system and the relief valves?

Your application is a pretty well, worked out model. Mechanical refrigeration systems like the one you describe have been around for many, many decades and the PSV scenarios are well-worked out. The vaporizer is usually designed to have an MAWP that is equivalent to the H.P. liquid refrigerant reservoir. The chiller system is usually designed to "settle out" in pressure throughout the system in accordance with the refrigerant's vapor pressure at the expected settle out temperature (usually ambient). This usually means that there will never occur any relief of the vaporizer shell side – although it is still protected with a PSV, according to ASME requirements.
Do you agree with these design principles and does your unit follow suit?

By your asking where to place the refrigerant in an existing unit, you have confused the real issue. You obviously are not free to debate that theoretical question. What exists, exists. You would require a lot of engineering calculations and capital money to reverse the fluid locations in the vaporizer. So, why do you ask this non-issue, theoretical question? You obviously have to use what you have in front of you - unless it is dangerous or it simply won't work (which is something you have not stated). Please explain this.

[pleckner]

You only gave us the MAWP of the shell side but not the tube side. If what I suspect is true, you don’t have a credible tube rupture scenario to consider. If the MAWP of the tubes and shell are within 2/3 of each other (assuming a system test pressure of 150% of design) or are within 10/13 of each other (assuming a system test pressure of 130% of design), you meet the exclusion for not having to consider a tube rupture scenario.

For your question #2, this is NOT double jeopardy. Double jeopardy is defined as simultaneous and unrelated events. A pump dead headed increasing system pressure to that of the shut-off head is event one. Tubes in the exchanger weakened over time and break because of this increase in system pressure is event two. This is cause and effect, sequential events that are related. If you do indeed have a credible tube rupture scenario then you must calculate the relief requiremnts for this case.

But Art brings up a very important point for this whole discussion. If this is a unit bought from a vendor, the vendor is required to provide you with the proper protection and it is their responsibility to provide you with the calculations.

By the way, this should have been posted in the Relief Devices Forum, not in the Industrial Professionals Forum.