Method for rating relief requirements for a CO2 vaporizer?

[hollerg]

I would like advice for the blocked in case for the CO2 side of a vaporizer. The control logic will allow the steam side to reach the header steam pressure. As I understand it the blocked-in venting starts as a liquid expansion case, but as soon as the valve opens vaporization begins and heat transfer increases. Since the heat transfer coefficient on the steam side is the same order of magnitude as on the CO2 side, I'm guessing the overall U value from the exchanger rating form is not necessarily at it’s peak. It underestimates if it includes fouling factors and it may incorporate the subcooling zone and a superheat zone.

I'm thinking the rigorous approach is to do a sensitivity study with heat exchanger rating software, omitting fouling factors, to find the peak local vaporization heat transfer coefficient from the incremental analysis, with the outlet at 110 % of set point and the steam side at full steam pressure. The peak local value from both sides of the tube would be used to compute the maximum overall heat transfer coefficient.

Is there a simplification I can make that I have overlooked?

 

[georgeverghese]

Sounds good. Also take into account the max steam flow possible through the supply control valve.

Tube rupture case with steam bursting into liquid CO2 not valid, and why ?

 

[don1980]

Yes, there's a simplification that yields perfectly acceptable results. Use the design Q but multiply it by a correction factor (dT_relief/dT_design) to account for the reduction in delta-T at the relieving pressure. At relief pressure, the dT is smaller. I would not spend time attempting to refining the overall U value - it's not going to vary enough to justify that effort.

 

[hollerg]

The tube rupture case is valid. In this case the steam side has the lower MAWP so the broken tube scenario sets the steam side relief. Is there some factor that influences the relief for the CO2 side of the vaporizer?

 

[hollerg]

don1980
Is there any literature you could point me to about using the design U as an acceptable short cut? Maybe I overlooked it, but I didn't find API 521 discussing that approach. They explicitly do remark on ensuring the fouling factors are removed from the calculation.

 

[don1980]

API 521 is not that prescriptive. It provides high-level guidance and leaves the detailed decisions to the user.

 

[EmmanuelTop]

I find this to be a 'chicken and egg' kind of a problem; careful analysis is required for each system. As the outlet of CO2 vapor is blocked, the inflow of liquid CO2 is reduced and it becomes dependent on upstream pressure and the existing CO2 pressure/flow controls or whatever is available in the DCS.

Steam flow is driven by the amount of energy that can be consumed on the process side (similarly to distillation column reboilers). Once the CO2 flow reduces, the flow of steam will reduce as well. Assuming equal Q (heat exchange) in the blocked-in case would probably result in a greatly oversized PSV.

Dejan IVANOVIC
Process Engineer, MSChE

 

[don1980]

Dejan - One of the high-level guidelines in API 521, applicable to any type of system, is that one shouldn't take credit for a favorable response from the control system. If the control system respond unfavorably, then you need to design the relief for that, but you shouldn't reduce the relief design for favorable action, even if it's likely to occur.

I agree that there are some cases that justify a more rigorous analysis. Specifically, those are cases in which a rigorous analysis could be justified based on cost (e.g. much smaller relief device) and/or risk reduction (e.g. less hazardous material released to atm). However, in most applications I find that rigorous analyses are not justified. In every case the user is responsible for assessing whether that applies to the specific case at hand.

 

[georgeverghese]

An intractable problem with pure CO2 relief is when solid CO2 forms on the LP side of an RV - this happens when the LP side of the RV is less than approx 5barg. Have we checked for this ?

 

[EmmanuelTop]

See attached paper

http://smithburgess.com/wp-content/uploads/2013/07/Cristea-Making-Relief-Loads-Match-Reality.pdf

- it does not explain the exact same scenario but there is similarity of your case to the first example in the paper. And perhaps an idea how to employ simulation in relief load calculations.

don1980, I fully agree with you - he can't go wrong with the PSV designed for the full/design heat transfer rate (reduced for DeltaT ratio at Relief/Design conditions) - if he wants to make sure the equipment is protected against overpressure. I just wanted to make a point that the PSV will likely end up greatly oversized, because heat transfer normally diminishes substantially if the flow of cold fluid is reduced or stopped completely.

We all know this from operating distillation columns, when e.g. reboiler outlet gets flooded and the circulation stops. The amount of heat pumped into the system at these conditions is well below the design rates. Now, this is a different systems and capable of generating very high pressures upon absorption of heat, but some general heat transfer concepts still apply, in my opinion.

Dejan IVANOVIC
Process Engineer, MSChE