Relief Load Distribution 2

[hesanchez]

Hi,
I am designing a relief system for a gas dehydration station, the equipment handling high pressure gas are the well separator receiving the well stream, the inlet dehydration scrubber, the glycol contactor and the outlet dehydration scrubber. All of these equipment will have installed relief valves with the same set point.
In the case of a blocked outlet downstream the glycol contactor, the pressure will rise in all the system, but the first relief valve opened will be the one of the well separator, after, the one of the inlet scrubber and so on.
Taking this into account, it is not probable that the relief load of the outlet scrubber or the glycol contactor will be the total flow rate passing through the vessel, it will be lower because it is relieving in the first relief valves.
In the case of blocked outlet, API 521 recommends to size the PSV to be able to relief all the flow rate normally passing through the vessel.
In this case, API 521 recommendation is not applicable (considering the facts), how can I determine the maximum relief load from each source, taking into account the pressure profile along the station.
I appreciate the help you can give me.

 

[TD2K]

If you block in the station you need to be able to relieving the incoming stream. Does that flow rate fit through the installed relief valve on the first vessel? I'm assuming it would be since there are likely block valves between it and the outlet of the facility.

If it does, the required relief rate through the downstream vessels is zero since they can't be overpressured. The first relief valve is effectively (since all the relief valves have the same set pressure) protecting all the vessels IF it has sufficient capacity.

If it doesn't, then you have to shift the remaining balance over the capacity of the first relief valve onto the next relief valve and so on AND make sure any interconnecting block valves are car sealed open or locked open to ensure the downstream relief capacity is available.

 

[Yuze]

If you can give a drawing to show your system configuration, we may help you to define your problem. The relief capacity is a code requirement. We don't want to guess or mislead you

 

[hesanchez]

Ok, I send you the drawing.
If I understand, according to TD2K, the relief valves downstream the first relief valve must be sized for zero flow in case of blocked outlet, so the sizing of the rest of the valves must be performed only for fire case?
Thank you in advance

 

[hesanchez]

Another question,
I have been sizing the relief header for the same system in the drawing above, the criteria used in the standards is back pressure, however, when i calculate the stream velocity it is above the critical velocity, while back pressure is below the range recommended. Someone told me that these high velocities are acceptable taking into account that the relief event is occasional and not expected in normal operation, but I don't consider it is a safe condition. I hope yuo can help me again. Thank you in advance.

 

[Yuze]

hesanchez,
1, first depending on the pressure rating of the vessels. Let's assume that the vessels have the same pressure rating.
2, depending on the locations of the block valve(s). If there is a block valve located downstream of the well separator, you have to size the PSVs at the piping and at the vessel with full capacity of the flowrate. In such case, the rest of the PSVs will be the supplemental PSVs for fire case (you can size and set the pressure based on fire case requirement).
3, If yor only have block valve downstream of the glycol contactor, you can simplify the system by eliminating some PSVs.

Answer to your another question:
For the gas relief case, the most PSVs are sized in critical flow region. The advantage for the flow in critial region is that the increase in back pressure will have little effect on the relief capacity. It will be safe if you use the formula for critical flow from API520.

 

[TD2K]

Looking at your sketch, I would size the first PSV on in the stream labelled "process stream, flow constant" for the full incoming flow rate. Since that PSV has the same set-pressure as the remaining downstream PSVs (and I'm assuming those set pressures are consistent with the design pressure of the system and equipment), then the first PSV also protects the downstream equipment and system. The remaining PSVs would be sized looking at any other possible overpressure causes, fire being a typical one. You might also want to consider sizing the PSV on the wellhead separator for the full incoming flow giving you in effect a spare PSV to allow the upstream PSV to be serviced if required with the plant on-line.

Where in your flare system are you above the critical velocity? By critical velocity do you mean sonic velocity?

 

[hesanchez]

According to that, I will size the PSV before the Well Separator at full capacity with antoher valve as spare (with the same set point), and I will size the rest of the valves downstream considering the other probable contingencies.
I mislead the term critical velocity, I meant erosional velocity. I have velocities higher than erosional velocity at the outlet of the relief valve, is it recommended to install a piping expansion just after the relief valve? or velocities above erosional are acceptable taking into account that a relief load is not expected in normal operation and probably it won't be usual?
Thank you in advance

 

[TD2K]

How are you estimating the erosional velocity?

I don't tend to worry about erosion of piping in gas service and especially in flare systems where the period of maximum velocity is very short.

 

[hesanchez]

I am estimating erosional velocity as reported in API 14E,
Ve = C/(PM)^0.5
Setting C as 125, recommended for intermittent service.
Then, in this case (relief), it is recommendable to ignore the velocity parameter and take into account only the back pressure to size the outlet pipe of the relief valve?

 

[TD2K]

I thought that might be the formula you were using.

No, I would not use that velocity to size a relief header. Velocities in a relief header are usually several hundred ft/sec. 70% of the sonic velocity is a common limit I've seen in various client specs subject of course to the actual backpressure at the various relief valves and their limits. For example, if you have a conventional PSV set at 150 psig, the built-up backpressure can't exceed 15 psi even though you might be way under your limiting % of sonic.

 

[hesanchez]

Ok, thank you for your help, it's been really usefull for my design.