Tank Implosion due to insufficient vent diameter 4

[tmengineer]

Hello,

A vessel has imploded due to having an insufficient vent diameter and I have been tasked with calculating the new appropriate diameter.

The original vent was 4", 20m long, with 3 elbows, had a fan which wasn't used in the pipe and increased in diameter to 5" after 5m.

The tank volume is 13.0 m^3 and was emptied with a flow rate of 900 l/min.

I started off by calculating the pressure drop in the pipe by using Bernoulli's equation and accounting for major and minor losses but the pressure drop I calculated was negative and insignificant (-39 Pa).

Either I have made a mistake in my calculations or my approach has been incorrect. I used incompressible flow models as the flow velocity was small (1.91 m/s) and so the Mach number < 0.1.

Is it a calculation error or is the method wrong?

Any feedback would be greatly appreciated!

 

[bimr]

If a closed tank was heated somehow from 20 degrees C to 80 degrees C, the tank pressure would increase 155.6 mm Hg above atmospheric.

 

[katmar]

It is these rapid heating or cooling events that destroy tanks. The vent sizes required are large because the heat transfer from a spray of hot ot cold water to the surrounding vapor space is very rapid. The API 2000 standard covers tank vents and the sizes it recommends will be very much bigger than the vents you have - which seem to be matched to infill and outflow rates only.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[tmengineer]

bimr,

Yes when using the ideal gas law for a constant volume (closed system) and dT from 20 degrees C to 80 degrees C the overpressure is 207.4 mbar (155.6 mm Hg).

And also, when modelling the tank as a closed system with a pure water at 80 degrees C pool at the bottom the vapour pressure is 478 mbar (355.1 mm Hg).

Which of these do you think is more accurate? 478 mbar is obviously the worse case.

Katmar,

Thank you! I had always assumed that the heat transfer between the incoming fluid and vessel gas was instantaneous but I wasn't sure if that was a valid assumption. I have API 2000 5th edition from 1998 (it's the only one that was free to download) is there a particular section that you think is relevant? Just to save me from reading the whole thing

The spreadsheet that I made gives a maximum overpressure for a 5" vent of 84 mbar (63 mm Hg), for a 6" vent 54 mbar (40 mm Hg). Do these seem reasonable? The spreadsheet removes the closed system assumption by accounting for the outflow of gas due to the overpressure.

I still haven't figured out a way to determine what is the maximum overpressure that the tank can take. Zdas04 suggested 0.5 kPa (5mbar) but this is far below the pressures that my spreadsheet is giving out.

 

[katmar]

I'm not sufficiently familiar with API 2000 to be able to recommend which bits to read and which to ignore. I seem to remember a discussion on Eng-Tips somewhere regarding a recent revision that made it even more conservative than it was before. A search here for discussions regarding API 2000 should be useful.

Any calculations of the overpressure will have to make some assumptions regarding the rate of heat transfer. I think that is the real value of API 2000 - they have done the research to know what works and what doesn't.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[EmmanuelTop]

I think you should read the entire API 2000 - not necessarily because some paragraphs are less (or more) related to your particular case, but rather to get a full insight of inbreathing/outbreathing requirements and all the factors that come into play for determining the vent size. Once you get a complete picture of all the processes which may occur during transient conditions, doing the calculations afterwards is the easier part.

As other contributors to this topic have said already, there are multiple factors which contribute to the venting requirements - apart from liquid inflow/outflow. Heat transfer can be a complete game changer, and this is evident in your particular case. API 2000 provides the best and the most detailed analysis of all possible scenarios. I think you won't regret reading it entirely, since it will come back to you on many occasions during your professional life.

I am attaching a handy API 2000 spreadsheet to this post, developed by one of the veteran members at ChE Forums and available for download at the parent website.

Dejan IVANOVIC
Process Engineer, MSChE

 

[bimr]

Although API 2000 is a great publication, I an not sure that API 2000 is really relevant to a scenario with a wooden tank that is just 2-3 m diameter and 3 meter tall. API deals with weather, flashing, conflagration, petroleum, etc., none of which are present here.

The OP recently stated that "that the tank failed due to the overpressure caused by the hot rinsing rather than a vacuum from draining" but neglected to say at what flow rate (or volume) the high rinsing is done. There is a temperature of 80 Degrees C.

To resolve this problem, the flow rate and volume of rinsing needs to be understood.

Since the purpose for the rinsing is disinfection, the temperature of the tank is probably above 60 Degrees C. Is the tank full at this point?

Based on everything the OP has stated to this point, the pressure drop for air flow through a 4-Inch vent is very low and adequate.

It also seems that the OP should also have some concern for condensation in the vent pipe since it likely to fall downward into the product. One would think that was the purpose of the fan. To prevent condensation from forming in the vent pipe.

 

[LittleInch]

The other factor to consider here is that many of the equations and systems used are designed for steady state and over the long term things settle down. The real situation though is a rapid transient change and the instantaneous flow rates and over pressure can be 4 or 5 times what the steady state flow is, but even if this lasts for only 2 or 3 seconds, the damage is done. Although low, the mass of air in the vent has an incremental effect when being asked to accelerate from nothing to several m/sec in 1 or 2 seconds.

tmengineer - you really need to see one of these operations in action to see how they do it, what sort of start flow of hot water they use, the start temperature of the vessel, the temperature and pressure of the hot water, the amount of free space, the orientation of the vent pipes etc before you can start any serious transient assessment.

They may be able to solve it simply by introducing a small flow of hot water for 1 or 2 minutes before turning it on flat out.

As someone said on a different thread - don't trust what an operator says he does, do you think they're going to admit they did something bad? or maybe no one has told them what the correct procedure is and why.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

If at some time you intend on going for a backup PVSV, there are a few things to note - there must be sufficient head room on design overpressure and underpressure to allow for

a) control (+/-) pressure range ( ie. normal operating press range that gives you room to carry out your normal operations) - these should be within the capacity limits of a well designed normal operating vent line

and

b)the setpoint of the vacuum and overpressure PVSV - industry practice on low pressure tanks is setpoint is at 80% of tank design pressure ( for either (+) or (-) settings, with typical accumulation of 20%.

Others have mentioned the prudence in using industry wide guidelines for deriving the capacities for these PVSVs. If the vapor space in these tank is fouling, 2x100% PVSV units with a slide valve / plate for each set may be better if you want to minimise downtime for cleaning and recalibrating these offline.

It may be a good idea to check if the newer SS tanks meet these requirements.

 

[tmengineer]

Hello!

Sorry for the late reply - I haven't forgotten about this I've just been busy doing some other work for the last few days. I'll get back to you properly next week once I've had a chance to look at this again.

Thank you so much for your contributions so far!

 

[Duwe6]

I'll bet a full round of drinks that this was a collapsing vapor event. When steam collapses back to water, it happens so fast that you will outrun the Conservation Vent capacity by enough that the roof or shell will be sucked in. And when the engineer shows up to investigate, all the operators swear that they didn't purge that tank out with steam. Have hat one shell 'beer can; collapse , and about 6 sucked in roofs in the past decade. All these small tanks had functioning Cons Vents -- 3NPS on 15-ft tanks. Only mechanism with enough energy, that fast, is a steam collapse/condensation.

And none of the [union] operators did Anything Wrong; Honest! After the tank cools off, there is no latent evidence, except the massive energy it took to happen while vented.

 

[tmengineer]

Duwe6,

At 80 degrees C and 1 atm the saturated vapour pressure of water is 481 mbar (calculated with the Buck equation - similar to Antoine but simpler), and the water vapour pressure is 478 mbar.

The temperature which the saturated vapour pressure is 478 mbar is 79.82 degrees C (again, calculated using the Buck equation). Does this mean that as the tank cools from 80 to 79.8 deg C there will be no decrease in pressure, but then as the T decreases from there the pressure will drop due to condensing vapour?

Or is it the case that the difference in the two pressures is due to errors in the Buck equation and I've just calculated the saturated vapour pressure by two different methods?

This is different from the failure mode I had been looking at (overpressure) - but it seems plausible.

LittleInch,

Yes I completely agree. This is a transient problem that cannot be assessed with steady state equations. I've tried not to fall into that trap but it's possible that I have.

Bimr,

Volumetric flow rate of hot rinse is 15 m^3/hr, tank is empty with a gas volume of 86.4 m^3

 

[bellowsmfg]

A lot of good things covered here. Utilizing API 2000 is a great one. The major thing that stands out for me in addition to what is discussed here is the 20m length of vent pipe (in addition to your other pressure drop cause by extra elbows and fan). Vents on tanks should be short as possible. A Vent on a tank should be short, oversized and with minimal elbows to prevent rainwater entering. The overall purpose is to avoid collapse during draining or lifting the roof off it - which happens from over-pressure.

Conor

Bellows Manufacturing and Research, Inc.

https://bellowsmfg.com

 

[KvdW]

I'm with Duwe6 on the possible cause. Vent calculation for collapsing vapor has been addressed in thread391-329825 Vent Sizing

 

[tickle]

API 2000 was updated in the last few years and I believe that the tank venting calculations were updated. I would suggest that you make the effort to get the latest version.