3 replies to this topic

[PhilippM]

Hello everyone,

we have the following situation as shown in the attached sketch:

The distillation column C-1 can be overfilled with liquid when valve V-3 is closed. The liquid (density = 600 kg/m³) then flows over into B-1, and at some point the entire system is liquid-full. Feed pump P-1 has a shut-off pressure of around 18 barg, so the pressure rises in the system. Relief valve RV-2 opens, but is too small to handle this relief load, so the pressure continues to rise until RV-1 opens (the allowable accumulation of B-1 is exceeded at this point due to the additional hydrostatic  head of 1,2 bar). RV-1 is much bigger than RV-2 and can handle the relief load.

 

In order to prevent the pressure in B-1 from exceeding the allowable accumulation, it was proposed to change the set pressure of RV-1 to 13,8 barg (15 barg minus 20m hydrostatic head), so RV-1 and RV-2 would open at more or less the same time.

If however a vapor relief case occurs (e.g. excessive heat input by reboiler E-1), RV-1 will always open first (hydrostatic head of gas is negligible), which somewhat defeats the purpose of having multiple relief valves, as usually the smallest relief valve should be set to the lowest pressure (risk of chattering?)

 

I have the following questions:

Is this a widely used relief valves design, or has anyone designed/operated something like that?

Can the set pressure of RV-1 be changed to 13,8 barg as mentioned above, or should there be a larger difference between set pressures ? (we can go as low as 13 barg without causing problems during normal operation)

Are there any standards/guidelines that describe such a system and how the set pressures should be staggered?

I only found section 5.4.2.2 (Multiple Device Installation) of API 520 Part 1, but this seems to be applicable to multiple devices that are installed on the same vessel, not different ones. Also, the applicable code here in Germany makes no distinction between single or multiple device installation and only allows 10% accumulation, not 16%.

 

Any help would be appreciated

Attached Files

•  Sketch.jpg   47.21KB   8 downloads

Edited by PhilippM, 24 January 2020 - 07:42 PM.

[Sharma Varun]

Normally in cases of blocked outlet at column bottoms there is sufficient hold up volume available in the column bottom sump providing operator with sufficient time to act before the relief occurs. Generally liquid filling scenario can at the max dislodge the column internals and entire column being filled is very rare, however same can’t be ignored. Please see at what rate the column is being fed and calculate the time required to fill the column completely, if the time is sufficient you can take credit of level alarms on the column to manually trip the pumps. Column pressure and temperature profile will also indicate the same. If found credible, instead of sizing the PSV and thus the entire flare header for this liquid relief case, HIPPS mechanism with 2oo3 level trip of charge pump can be considered.

 

Having said that, if you decide not to put the HIPPS or consider the operator intervention, Column overfilling scenario load should relief from column overhead PSV only and not the reflux drum PSV. Your configuration tells me the condenser is total condenser so please check the relief rate w.r.t. condenser outlet. Please note that these multiple valves are provided for dedicated system, they may operate together but there purpose is different.

 

Coming back to your question overhead drum is generally designed for column top design pressure + static head. Thus safety valve on overhead drum is set at higher set pressure than the column overhead PSV. In short 13.8 barg & 15 barg seems logical to me. 

[breizh]

Hi ,

The document attached may help you for this project . It's about pressure control on distillation columns. It could help you to optimize the number or location of PSVs .

Good luck

Breizh

Edited by breizh, 29 January 2020 - 10:04 PM.

[PhilippM]

Thank you Sharma Varun and breizh, it was decided to lower the set pressure to 13,8 barg and simply check if the inlet losses and backpressure are okay for that pressure. For vapour relief cases it was accepted that the larger relief valve opens first.