Vacuum Breaker Flow

[kryanl]

kryanl (Mechanical) Apr 22, 2005
Hello. My problem is with two rather large vessels, A and B. Vessel B (rated for nearly a full vacuum) is kept at a normal vacuum of about 12” w.c. We are installing vessel A (rated for 20” w.c.) about 100’ away from B, and will connect the two together with a duct. Now, vessel B is known to have a fluctuating vacuum, so it is my mission to size a vacuum breaker for vessel A. I know with PSV’s, they recommend the pressure set point not be less than 10% less than the rated pressure, so I think for our vacuum breaker, sizing it to open at 18” w.c. should be ok.

My problem is trying to figure out what the design flow should be. Initially, I calculated what the airflow should be if somehow both outlet pumps (usually only one operates, the second is installed spare) are running, and the vessel drain is open. This loss of liquid would need to be replaced by the flow through the breaker, correct? But, if I consider if the vacuum in Vessel B drops, then how much flow is requried? Do I try to determine what theoretical airflow can travel through our 100’ long duct at a differential pressure of 18”? That seems like it would be pretty high, and I am not sure I see why the breaker would be required to pass that much. Plus, since I don't know what breaker I am going to pick yet, I don't know it's orifice size, therefore can not determine what the pressure drop through it would be.

I usually tend to try to look into things too much and am assuming I am doing the same here. Can anyone set me straight? Thanks for any and all help.

 

[Montemayor]

kryanl:

I admire your importance to detail in project engineering of this sort. After 45 years in the business, I’ve found that engineers have made past costly mistakes because of a disinterest in engineering details that often come back to bite them. My experience has shown me that taking a lackadaisical attitude in project work leads to gross mistakes and serious accidents. The most important things to keep in mind in this application are the latest, accurate tank vacuum rating and the maximum rate of vacuum draw imposed on the tanks. These factors lead to identifying the two critical items you must identify:

1. The correct, safe partial vacuum set pressure for the vacuum breaker;
2. The correct, safe capacity rating for the vacuum breaker.

Do not take the real, existing vacuum rating on your tanks lightly. It is a very simple matter to “suck in” (collapse) any atmospheric vessel or storage tank – regardless of size. I know of some engineers who presently still walk around with the horror of watching their tanks collapse in front of their eyes simply because they didn’t take care of engineering business by checking out all the details in the tank design. I strongly recommend you obtain the services of an experienced mechanical engineer who can do the tank pressure/vacuum rating on your vessels – that is, if you haven’t had it done already. This is very important. Many atmospheric vessels are usually 20+ years in age and have suffered corrosion or mis-application in the actual or prior service. Do not take anybody’s word or opinion on the actual tank rating. Rely on a serious, methodical and documented rating.

You state that “they recommend the pressure set point not be less than 10% less than the rated pressure”. I don’t know who “they” are, but the set point on any PSV (or tank vent) should be the MAWP (or MAWV) of the vessel. For the pressure case, you can set the PSV at a lower pressure than the MAWP. For the vacuum case, you can set the PSV at a higher pressure than the MAWV. Do not get confused with “vacuum”, partial vacuum, or absolute pressure. This is a common mistake when dealing with so-called “vacuum” designs. That’s why the safest unit to deal with is absolute pressure.

You should identify all the credible and possible pressure and vacuum scenarios. What you describe as the pump-out scenario (for the vacuum case) is only one – the simplest one. The fluctuating vacuum in vessel B is another scenario and must be identified as to source and capacity – just like the pump-out scenario. Another scenario is one that has been contemplated in the past on other applications: a broken level gauge glass or a broken or ruptured tank nozzle – with subsequent spill-out. When dealing with chemical or hydrocarbon tanks I have had to evaluate and size vacuum capacities in the event steam is used to clean out the vessel and a rain shower suddenly drops on the vessel and causes a vacuum due to steam condensation. These may not be scenarios that apply to you; my point is that you should study all your possible and credible scenarios – not only the pump out and vessel B (which you haven’t quantified). The bad part about vessel B is that if it is connected to vessel A and you are breaking the vacuum at vessel A while vessel B is the source of a greater vacuum, then you must de-rate the PSV (over-size it) in order to overcome the resistance to flow that exists in the atmospheric air getting through to vessel B through the 100 ft of piping (& fittings). Calculating flow resistance for vacuum service is an art rather than a science because of the many unknowns. You’re going to have to be very conservative in the line size between the two vessels. I would rather place a separate, independent PSV on vessel B also. I don’t believe in economizing on such a relatively low-cost item such as a PSV. If you have ever lost a tank or vessel by “sucking it in”, you’ll agree with me on this. You haven’t been specific by giving us fluid identities, capacities, sizes of vessels, pipes, or flows so I can’t give any specifics also. My experience shows that the price of an independent PSV on vessel B would be the wisest and most practical way to resolve your valid concern.

I hope this experience helps.

 

[kryanl]

Thank you Montemayor for your reply. You had very good insight. As far as the rating of the tanks, A is a new tank we are having constructed to a rating of 20", so I am pretty comfortable in that. B is an existing vessel which was indeed re-rated last year. It had been originally designed for a full vacuum, and has lost some of that, but is still in good shape for it's new application. The 'they' I referred to is ASME and/or API. In the past, we've always specified pressure safety valves for, set points of, say, 90 psig, when the rating of the vessel is 100 psig. The extra 10 percent is to counteract chatter in the valves. Of course, we do have to consider normal operating pressure, making sure it doesn't apprach 90 psig.
Putting a vacuum breaker on B is an interesting idea. I would assume it would be set at much lower than B's rating, to protect A. The only thing about that, is that I had kind of thought that the piping in between the two tanks would reduce the vacuum reaching A, creating a sort of buffer, if you will. If we added a breaker to B, it would pop more often than if we put it on A.
And, again, as far as the capacity of the breaker, would I need to add all scenarios' required flows, or do I take the largest of the bunch?

Thanks again for your input.

 

[Montemayor]

kryanl:

What I recommend you evaluate is placing a separate, independent PSV on vessel B and isolating this vessel from vessel A – that is, if your process can accept this type of operation. And what I mean by isolating the tank is that no liquid or vapor lines are connected between the two tanks.

The reason I recommend this (if possible) is mainly to protect tank A (the weakest link of the two). Also, when you consider the labor in installing a larger vapor balance line between both tanks, you’ll find that buying another PSV for the new, tank B set at the rated vacuum for that tank will be less costly. You would keep the current PSV on tank A, protecting that tank. I suspect you might be able to do this because the only reason I normally join two tanks together via liquid and vapor lines is due to the need to have them act as mutual surge or capacitance providers. In your case, with tank B piped up differently from tank A (as evidenced by the fact that it undergoes totally different and deeper partial vacuum conditions), you are “handcuffed” by the higher partial vacuum rating of tank A. And your tank A will always be subject to any lower vacuum PSV setting on tank B as long as both are inter-connected. You cannot protect two inter-connected tanks from different pressure scenarios independently – this applies to pressure as well as vacuum. The moment you have inter-connected the tanks, you “have bought the ranch” – i.e., you are tied to the rating of the weakest link in the connection and you must protect at that level. The fact that you conceivably could have a deeper vacuum imposed on one tank, while it is connected to a weaker tank, means that you must set your PSVs to protect the weakest one – whether you place a PSV on each tank or not.

Additionally, you cannot rely on the PSV to regulate the flow rate of the incoming air to control a vacuum. You either break the internal vacuum with sufficient incoming air or you don’t. The PSV is not designed to act as a vacuum regulator. It merely allows the entrance of the designed air flow to break the vacuum at the rate of the maximum credible vacuum case. And it does this in step-wise fashion – not in a differential, controlled manner such as a control valve.

Please treat my recommendations as general, since we don’t know the details of your process and the scope of your project. What I perceive may be different than from what you are doing – or plan to do. But generally, tanks are protected against partial vacuum failure much in the manner I have described. You can verify or confirm this by going to the ProtectoSeal Website:

http://www.protectoseal.com/ventingmenu.html

These people can provide you with information as to their recommended setpoint for the PSV in order to ensure that your tank does not exceed the MAWV rating for the tank in question. If your PSV supplier is another manufacturer, then contact them for their recommendations. What you are doing on this project up to now is all positive and will contribute to your safe and secure tank operation in the future.

Good luck and safe operations.

 

[kryanl]

Montemayor,

I understand your recommendations, however, you were right, our process won't allow us to separate the two vessels. We want to form a vacuum on tank A, but there is no way to do that. (We didn't want to add a fan or steam ejector, or any other vacuum forming device to the tank) So, the only way to put A under a vacuum is by connecting it to B. If a vacuum breaker on A were to open, allowing atmosphere into the line, with the airflow obviously going to Tank B, it will not adversely affect any operations. These vacuum are basically operating as collection vents.

What my question now is, is how can I correctly estimate what the airflow from atmosphere through the line will be? I read you can use Darcy-Weisbach equation for airflow in pipe, but when I worked through it backwards (starting with the head loss of 18", and calculating for velocity in the pipe, and then figuring flow based on velocity and pipe diameter), I come up with a ridiculously low number. I have from the Crane technical paper, a chart which gives you head loss per 100' of pipe for air. When I plug in my calculated airflow and try to find a general head loss to check my work, I come up with zero loss per 100'. I have checked several times, and I am pretty sure that I am converting away from the 100 psi that is shown in the actual chart. Can you tell me, is there a better method to figure what the airflow through a pipe will be, when all you know is the pressure loss and size of pipe; without knowing a velocity or Reynolds number?

Thanks again.

 

[Montemayor]

The basic equation is the Darcy-Weisbach relationship. However, you are dealing with compressible flow - as opposed to non-compressible (liquid) flow. With compressible flow you are subject to "choked" (sonic) flow and if that occurs, you have constant mass flow rate. Additionally, the Darcy-Weisbach equation is valid (or accurate) for compressible flow only when you stay within 10% of the initial pressure as the pressure drop. If you are breaking your vacuum with atmospheric air at 14.69 psia, that means you have an approximate limitation of 1.45 psi as pressure drop. But you should first check to see if you are achieving "choked" flow. I don't know your vacuum level, but I suspect you are.

If you are using Crane TP #410, you are using the correct guidelines to calculate the correct pipe diameter. Read carefully page 1-9, "Principles of Compressed Flow in Pipe" and also page 3-3, "Limitations of Darcy Formula".

I hope this helps.