Pressure surge(water hammer issue)

[atayto]

Can anybody help me with issue descibed below.

During introducing of seawater from 14barg pressurized system to another system which at this time is at 1barg, damage to downstream equpment was observed. Our suspect is that damage is created by water hammer. I found a lot of formulas to calculate water hammer effect if discharge valve is closed. But in my case is backwards, isolation valve is opened. My quiestion is can I use the same formulas. If not can anybody suggest right formula for this.

 

[MortenA]

Knot know what the formular you refer to is about thats hard to say - however:

1) Surge arises from abrupt change of velocity of a fluid
2) Cavitation (column separation) may make things worse - these often occur at the same time

When you have a flowing stream and closes a downstream valve you have a classic water hammer - this can be estmitated by use of the Joukowsky equation. Check this article at wikipedia - look OK:

http://en.wikipedia.org/wiki/Water_hammer

It could also occur if you have a dry pipe that you prime - the downstream end terminates in a valve. This valve will have a very high capacity for air compared to water. So once the water hits the valve it will slow down considerable and you have a water hammer - even though you don't close a valve.

2) can arise eg. from tripping a pump or closing an upstream valve. But other circumstances must also apply - so that the pressure increase sufficiently for the cavity to collapse.

Best regards

Morten

 

[SEP87]

As I understand it, water hammer occurs due to the shock wave created when a valve is shut quickly causing a pressure surge back up the pipe.
I haven't heard of water hammer working the other way, but it may be possible, in which case (this may be stupid question) can't you just open the isolation valve slowly? Therefore introducing the high pressure water at a much slower rate.

Or use a pressure regulator therefore actively controlling the pressure rather than just opening up a valve?

 

[willhammett]

Water hammer can also occur from pulsations is flow, this is especially true at higher velocities. The sudden drop in pressure shouldn't have it creating water hammer, especially since the equillibrium pressure is 1barg.

 

[21121956]

Hello everybody:

In a closed conduit, when in normal operation a valve or wicket gate is open, the flow of water is increased, energy is supplied to set the water in motion, with a consequent reduction or drop in the hydrostatic pressure.

atayto (and specially SEP87), in order to get a better idea regarding this subject, you can read the book "Waterhammer Analysis" by John Parmakian.

 

[21121956]

Hello everybody:

Here you can see the effects of a sudden opening of a valve or gate in a hydroelectric power plant.

 

[BigInch]

Obviously introducing 14 bar pressure into a 1 bar system will give you a pressure wave eminating from the 14 bar system. That pressure wave will accelerate the fluid in the 1 bar system to immediate movement, the first molecule hits the next molecule of fluid, etc. etc. until it hits a wall and starts building pressure, at which point the wave will be reflected back to the source increasing pressure in the 1 bar system, that effect continuing until the 1 bar system pipe is expanded as much as is required to hold the volume of fluid injected into it and a new equilibrium pressure is reached.

The equations calculating pressur increase based on valve closure, OR OPENING, time are not based on a F = M A concept as above, but on Bernoulli's change in pressure is inversly proportional to change in velocity relationship. V^2/2g. That assumes that the initial velocity or final velocity is constant whereas with F=MA, V = A dt. You'd be better off figuring the Acceleration on the fluid column with an instant introduction of 14 bar on one end, and the Force in which that would result. Its more like hitting the 1 bar system pipe on one end with a real hammer with a total force of 14 bar * the cross sectional area of the pipe, or something like a 22,000 lb smack on the end of a 12" pipe. Ya. You might expect a bit of damage from that.

The valve formula would essentially divide that blow (or something like it) into an impulse lasting for the valve closing time, say 22 seconds, which is only 1000 lbs a second or so. Maybe the same 22 K lb hit through a 2 meter thick hard rubber mat. Much different than a hammerlike hit of 22K lbs right on the "bean".

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[BigInch]

Calculate the head equivalent of 14 bar. Then I think it would be like starting with the velocity equivalent of the 14 bar head (14 bar head * 2 * g)^0.5 = V, and then use a valve closure time of just about zero.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[SEP87]

In a closed conduit, when in normal operation a valve or wicket gate is open, the flow of water is increased, energy is supplied to set the water in motion, with a consequent reduction or drop in the hydrostatic pressure.

Thank you for your valuable input.

atayto (and specially SEP87), in order to get a better idea regarding this subject, you can read the book "Waterhammer Analysis" by John Parmakian.

I will endeavour to read said book, to ensure I can provide equally valuable information as yourself in future. Perhaps I could even attempt to answer the question posed. I would, at least, be able to direct the person asking the question to some useful information.

 

[atayto]

Thanks gents for good advises. My main question was an i Use Joukowsky formula or not. As i understand this formula describes the condition where valve on pipe outlet is suddenly closed. My case is slightly different. Imagine 8" vertical 100m long line with heat exchanger at the top and gate valve on the bottom. This line is full of water. when gate valve(14 barg behind of gate valve) is open we have exchanger rupture. I used juokowsky formula and calculated that if the valve is opened in 4 sec then there is no problem. This gate valve is rusty and hard to operate it is impossible to open it quick but we did a trial and found that first 30 % can be opened in 10 sec. It looks like we open it slower than Joukovsky suggested but still have water hammer. Any comments?

 

[vpl]

atayto

When you open the "gate valve on the bottom" is there another valve on the other side of the heat exchanger that is also being opened?

If you have a system that is full of water (which is relatively incompressible) and you open a valve and try to inject more water at a higher pressure, the water has to go somewhere. If there's nowhere for the water to go to, the pressure in the system will go up until it's at equilibrium (like BigInch said). It sounds like the equilibrium pressure was higher than the heat exchanger (which was the weakest point) could take -- so the heat exchanger broke.

A general rule that I like to keep in mind with water is "what goes in, must come out. If you're adding water to a full system, then you need to have somewhere for the water to go -- an open exit valve, a relief valve -- or a leak in the piping somewhere where you don't want it.

Patricia Lougheed

******

Please see FAQ731-376: Eng-Tips.com Forum Policies for tips on how to make the best use of the Eng-Tips Forums.

 

[BigInch]

OK so we will use a 4 second time step. In 4 seconds a pressure wave travelling in water will move almost 4,000 m, so everything within that distance will be rapidly increasing pressure. You must somehow release all of that pressure build up within that 4000 m before the next 4 second period expires (by commencing flow downstream in that segment at a sufficient velocity such that no further compression is possible) otherwise the pressure in that 4000 m segment will continue to increase.

If you only have a system 100 m long, most of that pressure increase will happen in 100 m /1000 m/s, or about 2/10_th of a second.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus

 

[atayto]

BigInch and vpl
Thanks for comments. Could you please also advise how can i prevent this issue fro reoccuring. Will be bucking pin PSV be appropriate for this. I looked through option of installation of surge vessel(similar design to pulsation damper) but size required is became enormous. I am currently looking through option of replacing gate valve with equal persentage globe valve.Could you please advise what is the most reliable method in your opinion?

Thanks

 

[BigInch]

I don't know your design code, so I can't say for sure what you need, but you obviously can't leave the 14 bar system connected to a 1 bar system for a lot of reasons; your pressure pulse problem being just one of them. Your 1 bar piping needs to be protected from the full pressure of that 14 bar source.
I think it would be appropriate to place a downstream pressure control valve (set to the 1 bar system's MAOP) between the 14 bar source and the 1 bar system. I would also put an automatic closure valve just downstream of that PCV set to close at the 1 bar system's MAOP +5% or so, and a relief valve downstream of that set to MAOP + 10%.

Let your acquaintances be many, but your advisors one in a thousand’ ... Book of Ecclesiasticus