Calculating Venting Capacity for Open Vent 1

RJB32482,

Essentially, you will evaluate the capacity of the vent in a similar fashion to a PSV discharge tailpipe - ie evaluate the hydraulic losses through the vent at your design vent flow and check the backpressure does not exceed 110% of the tank design pressure (assuming an API 620 tank - other standards may vary). You didn't say whether a PV valve is involved, but the principle is the same with or without. In plain vents you commonly find yourself having to estimate the pressure drop across a flame arrestor.

Good guidance is contained in API 521, 620 and 2000.

It's simply a fluid flow calculation with air at atmospheric pressure going into the 12" pipe and fittings and expanding into your tank to replace the volumetric flow of your scenario.

Good luck,
Latexman

On a re-read, you're interest is in in-breating. Much the same except tank vacuum mustn't fall below design gauge vacuum.

API 2000 design doesn't cover vacuum developed from condensation following steam out. In-breathing flows in this scenario can be very high and require rigorous evaluation on a case by case basis. Special operating precedures and/or temporary tank insulation are often required.

RJB32482:

The need that you describe is real and existing - especially when you know steam-outs will be employed on a storage tank. I'm not only familiar with this situation on a first-hand basis, but I've been there, at plant site, and witnessed the results of a tank maintenance contractor imposing a steam-out procedure on a storage tank and subsequently watching the tank crumble ("suck-in") like tissue paper when a brief (but fatal) rain shower fell on the tank. It's not a pretty sight to see.

Subsequent engineering audits proved that the tank vent employed to vent the steam was the wrong one (& definitely too small). The way we proved this was the same way I always size venting requirements for this type of scenario: size the required vent to replace the condensing steam with substitute atmospheric air without butting up against a sonic flow condition across the supplied vent opening. I developed a spread sheet calculation method to design the required vent and I've used it on a variety of steam-out design scenarios. The last time I did it a major oil company approved of the method upon their review. So you can also do similarly.

Art:

With reards to condensation rate:

You say that the collaps was caused by the fact that it rained - i.e. a faster heat transfer because the surface was now wet and supplied with fairly cold rain continiously (as far as i understood your post). How did you calculate the heat transfer rate and thus the condensation rate under these circumstances? If conensation is assumed to happen "instantaneously" the any vent will be too small

Best regards

Morten

Morten:

This phenomena of a rain shower causing a tank “suck-in” while it is under a steam-out condition is not strange for us in East Texas. You will have to believe me when I tell you that it has happened all too often in the past and will probably continue until people understand the basics involved. I’m not the only one who has personally witnessed this event take place.

The heat transfer rate is calculated by applying the vertical falling film condensation theory of Nusselt (I believe) and what I came up with was an estimated film coefficient of 500 Btu/hr-ft2-oF. I believe that you will find that for a condensing system, an overall heat transfer coefficient of 250 to 700 Btu/hr-ft2-oF is considered as very credible. From this basic start, I calculate the total heat transferred and the rate of steam condensation inside the tank. I can do this because I can easily calculated the cooling surface of the tank and the latent heat of condensation for steam at atmospheric pressure.

Knowing the rate of steam condensation, I also can calculate the rate of volume displaced by the steam because I know the steam’s specific volume before it condensed. I assume the amount of net volume reduced (steam volume – condensate volume) is essentially the same as the steam volume since the ratio of condensate to steam volume is very small.

The maximum possible air velocity entering the tank through the tank’s vacuum relief nozzle is the sonic velocity of air:

Vs = (kgRT)^0.5 = (kg144PV)^0.5

Where,
vs = Sonic velocity of air, ft/sec
k = Ratio of specific heats for air
g = acceleration of gravity, 32.2 ft/sec2
P = Absolute pressure, psia
V = Specific volume of air, ft3/lb

The required nozzle cross-sectional area = steam volume displacement (ft3/sec)/ft2(maximum air velocity)

The recommended nozzle area is one size larger or more.

Art: I'm not in doubt I was mostly interested in how to estimate the increase in heat tranfer! The i can follow quite easily

Best regards

Morten

Perry's 7th has a range of U = 400 to 1000 Btu/hr/ft^2/F on 11-25 for steam on the shell and boiler feed water in the tubes of a tubular exchanger. It includes 0.0005 dirt factor. In Perry's 5th it's on 10-44.

Film coefficients for steam and organic vapors calculated by classical Nusselt theory are generally known and concluded to usually be conservatively low. This could put your tank at risk to be sucked in. Dukler theory is the accepted method for condensate films and falling films. See Perry's 7th on 5-20 or 5th on 10-21.

Good luck,
Latexman

Morten:

I wasn't under the opinion that you doubted; rather, I offered the detailed explanation so that it might answer what others have always questioned me about when this subject has come up. The basis for replacing the volume of condensed steam is as I explain it. One has to come up with a rate of the condensation and the theory (because that's all we can conjure up) is Nusselt's classical work on falling film condensation on vertical plates. There may be other theories, but the basic premise I rely on is that the method should be conservative (because of all the contingent uncertainties involved). The fact, as you state, that the condensation will not be "instantaneous" (although it seems to be) but really differential in character, helps in making the estimated nozzle size conservative. Also, note that I use the "next size or larger".

The important point is to not cause a choke effect on the incoming air that is trying to relieve the undesired vacuum within the tank. From my previous experiences with other engineers - most of them young - this basic point is left not understood and consequently unresolved in picking the proper relieveing nozzle size. Thus, my tendency to explain in detail. I hope you agree.

One trap I see that has not been brought out clearly yet is that when a heat transfer coefficient is said to be conservative, we end up buying a heat exchanger that is somewhat too big. Nobody complains because there is extra capacity until fouling reaches a point where the extra capacity is used up. Even then since the time between cleanings is more than planned, everybody pats themselves on the back.

However in this case, if we use the same conservative heat transfer coefficients, the tank is at risk to be sucked in because the steam collapses faster than calculated. Based on this, a healthy safety factor should be applied.

I think a conservative upper limit on U would be to use the heat transfer coefficient of just the vessel wall as follows:

U = thermal conductivity/thickness

This may result in a large relief, but it does give you a real upper limit to consider.

My company advises us to design the tank for full vacuum if it is to be steamed out, which is a sign that our past experience at sizing a relief for this scenario has not been good.

Good luck,
Latexman

Monteymayor,

I have a follow up to your recomended method for sizing a vacuum relief vent. In your method you look to make sure that the vent size is large enough so that the velocity through the pipe is less than the sonic velocity for required in-flow of air. Then you add one pipe size.

My question with this method is that don't you need to take in account for the pressure drop through the vent nozzle? If you have a tank that is rated for 4" wc it would it not be able to pull less air through the same sized nozzle than a tank rated for 30". I use an orifice area equation in crane for this.

Best Regards

Sorry I should have edited better. The sentence should say 'If you have a tank that is rated for 4" wc would it not be able to pull less air through the same sized nozzle than a tank rated for 30" wc.'

Diborane:

You are correct in that you should take into account the fact that there will be nozzle resistance and you should design accordingly using the relationships such as found in the Crane Tech Paper #410. I have not followed through with all the details involved in arriving at the specific nozzle size because I was more interested in laying out the logical steps employed to arrive not only at a reasonable condensing rate, but also a practical, subsonic flow through the nozzle in order to ensure that the tank can be relieved safely. Thank you for taking the logic and the thinking all the way through to its completion.

Bringing this back to the top of this thread.

Two things:

-I read an old engineering relief manual from our plant and they had a certain relationship for the "steam" out condition:

CFH=(3*V*A*(tc-ta))/L

CFH= inbreathing required (ft^3/hr)
V= specific volume of steam (ft^3/lb)
A= surface area of the tank (ft^2)
tc= steam condensing temperature (F)
tf= ambient temperature (F)
L= latent heat of condensation (Btu/lb)

This only takes the heat loss from the tank as 3 BTU/hr ft^2 F. I believe this is WAY too low. Any thoughts (better to use the 500 as mentioned in previous posts).

-also on the vent, why do you need to avoid choked flow into the vent? At choked flow, you know the velocity is at its sonic speed (as posted earlier how to calculate it, say its V). If your vent is large enough so that V(ft/s)*A(ft^2)=Q is over the required venting rate, you are still good even at choked flow. Correct?

Thanks.

I'm a bit disturbed is seeing references to "sonic flow" while discussing the venting of an atmospheric tank. To get sonic flow, you would need a ratio of the upstream to downstream pressures of (say) 0.5. We're talking about avoiding a vacuum situation in an atmospheric tank. So the upstream pressure is atmospheric or 14.7 psia. Downstream pressure is the pressure inside the tank. It would need to fall to about half of atmospheric or 7.4 psia or less before a sonic flow situation would develop. Most atmospheric storage tanks I've heard of would collapse well before that point. So please, stay with subsonic flow when talking about atmospheric storage tanks.
Doug