Inlet line sizing for non-governing PSV sizing cases 3

[MarkESmith]

Hi all,

What is the generally accepted procedure for dealing with the following scenario:

I am checking adequacy of a currently installed relief valve on an existing separator vessel, for reverse water flow from a new vessel to be hooked up downstream. I have calculated that the required relief load from the reverse flow case requires an orifice area of ~200 mm2, vs. the installed ~1200 mm2 orifice. So the exisiting PSV itself is fine, however there is a potential problem with the PSV inlet piping: API 520 Pt II (S.4.2.2) requires that the inlet line is sized such that line losses are less than 3% of PSV set pressure at the rated capacity of the PSV. Using the rated capacity [(1200/200)*req'd = ~600% of required liquid water relief flow] the pressure drop is ~6% of PSV set pressure. The inlet piping is, however, fine for the governing case (gas breakthrough).

Is this a problem? The discrepancy stems from the fact that the PSV (and presumably the PSV inlet piping) is sized for a gas breakthrough case, and I am checking the adequacy for a liquid water relief load.

Any help greatly appreciated.

 

[don1980]

Mark....you're asking a very good (i.e. tough) question that we sometimes sweep under the rug.

There's no requirement to size the pipe for all the potential relief scenarios, but it's definately something for the relief designer to consider. Personally, I seldom do it. It adds time and cost to the installations. I've heard a few folks say they do, but I think the more common practice is to base the pipe size on the controlling scenario.

It's ultimately a risk management question. If there's a lesser scenario with a relatively high probability of occurring, and the consequences of failure are high, then I'd say you ought to consider sizing the pipe so that the inlet losses are satisfied for that scenario in addition to the controlling scenario. I suggest assessing each case separately rather than giving a blanket suggestion that this should be done for every installation.

The fact is, the cost impact of this can be high, and I'm not aware of any past accidents caused by failure to consider the pressure losses for these non-controlling scenarios. If you incorporate this into your routine work process, you'll likely find yourself installing a lot more remote-sensing pilot PSVs.

 

[Latexman]

If the risks are too great not to be rigorous, the way to do it is design and modify the installation to handle the controlling vapor scenario, controlling liquid scenario, and controlling two-phase scenario.

Good luck,
Latexman

 

[Chance17]

I have seen dozens of cases where the greatest inlet pressure drop contingency is different than the greatest orifice area contingency.
This typically occurs with large difference in molecular weight.
For example compare a reboiler steam leak contingency vs a oily reboiler excess heat contingency.
My view is that all possible overpressure contingencies should be evaluated rigorously.

Liquid relief valves have a "mine-field" of issues (flashing, certified vs non-certified, derating factors)
and I am reluctant to simplify comments on a public forum.
But I would recommend finding adminstrative means to use the 200mm2 case to evaluate the contingency.
Liquid release do not have the "pop" action of vapor venting.
Consequently, the valve will only open as much to relieve the actual input flow.
The valve may open 100% briefly, but this is steady state flow only if the valve is too small.
In your case you can't sustain steady state 1,200mm2 flow.
Why worry about the 1,200mm2 case inlet pressure drop?

 

[MarkESmith]

Thank you all for your comments.

We are proceeding by requesting the client to confirm that they are happy with the 3% limit to be breached in this non-governing case. My reasoning, along the lines of Chance17, is that the valve is extremely unlikely to ever see the rated flow of water at relieving pressure from the reverse flow case. As Chance17 has pointed out, there is simply not this level of inventory available to reach the rated flow. In any case, I believe (with my very limited experience and knowledge) a pressure drop of 6% will not give backpressure on the vessel sufficient to overpressure it in a hurry.

 

[delf]

Excessive pressure drops may occurs, ex. the rated folw of liquid relief valve may be siginficantly higher than the required flow, cause a very high inlet line pressure drops. In such cases, there are some methods that could be doned, such as increasing the valve blowdown. But it need to be considered carefully in your case.