swing check valve and surge

[laura80]

Hello everybody

I need some help about two questions:

1) Is posible for a swing check valve to close before the flow reverses because of the dropp of fluid velocity? I think it is not posible, but some of my collegues think it is.

2) I have the following instalation:

Pump(centrifugal) - swing check valve - security valve

If the security valve closes, and the pump is still running, ¿is the check valve going to close? I think, yes, is going to close when the flow reverses.

Thank you in advance for your help.

 

[laura80]

Sorry, I have to clarify that it is not a security valve. It is SDV valve.

 

[MikeHalloran]

Okay, now you can clarify what an SDV valve is. Please.


Mike Halloran
Pembroke Pines, FL, USA

 

[laura80]

Emergency shutdown valve

 

[SNORGY]

Presumably you will not have a rapid-closing ESD, since that is inviting the hammer that the context of your question appears to draw concern towards with respect to the swing check. What I would expect would happen is that as the flow slowly pinches off as the ESD closes, the centrifugal will back up towards the left on its characteristic, and the flapper position will already be starting to drop towards closed; how much depends on a number of things such as the ESD closure time, the geometry of the swing check, the mass of its flapper, and the shape of the pump's characteristic curve. In any event, I don't think the flow will reverse, so much as stop altogether; the ensuing transient will find its way back to the check valve and the pump. Depending on how much pipe is involved, the check valve might already be closed before any reverse flow or pressure transient arrives.

Regards,

SNORGY.

 

[stanier]

Laura80,

ARD Thorley Fluid Transients in Pipeline Systems and technical papers by Delft Laboratories would be the first point for your research.

The swing check valve will close when the flow slows based upon the mass of the disc, friction in the bearings, counterweight mass, dashpot characteristics, fluid density, losses in valve body versus force from the flowing liquid (ie velocity). Easy way to find out is a physical test. Hard way is to model it.

The other way is to model this in Impulse, Flowmaster, Hytran . Kypipe or other siftware package with the check valve ignored. Check out the fluid reverse velocity vs decelaration. Check with the valve manufacturer to see how this compares with test data (if they have it!).

Thorley uses a non dimensional method to analyse check valve behaviour. he also describes swing check valves as basically agricultural devices. Most manufacturers copied a design of decades a ago and would not have a clue as to the hydraulic behaviour of their valves. They make shapes; generally blue ones.Look for one that has used Delft Laboratories or Utah State to test their product,

"Sharing knowledge is the way to immortality"
His Holiness the Dalai Lama.

http://waterhammer.hopout.com.au/

 

[gerhardl]

The best way to answer is in my opinion to recommend the best possible installation: two valves (one checkvalve and one emergency shutdown valve) that can work independently of each other, without the one or the other disturbing the function of the second.

You need:

a) To avoid slamming by pressure peaks:

A checkvalve that closes when delta p is zero, or more correctly spring operated closing, marginally before delta p is zero. Best choice: non-slam soft closing nozzle check valve. Or, second choice, but depending on size and pressure class: disc (not swing) check valve with spring. Both valvetypes from top quality manufacturors.

Another choice might be tilting disc with counterweight, with hydraulic piston damping.

This third alternative will require exact flow data (all possible situations!) at all times to dimension the solution. Same is valid for the two others, but with larger limits for accuracy.

b) An emergency shutdown valve with regulation (setting) possibillities for the closing down time stroke, preferably if large size: fast for (say) first 70-80%, slower for rest.

Dimensioning is necessary both for size and closing time (both for one and two-step closing) to prevent hammerblow pressurepeaks backpressure.

The whole piping construction need to be considered for correct dimensioning and placement of alll components.

Direct answer to your questions:

1. Yes, but only if dimensioned correct, mounted correct and best with spring assistance, and only under certain hydraulic conditions (flow amount) and limits. In practice it will more likely have a very good chance of closing with a 'bang'.

You will often see swing check valves damaged with pin and hinge or sealing problems. This is caused by one or more of following causes: bad/weak construction, wrong dimensioning or mounting (placement) causing sudden closing creating shockwaves and repeating hammerblow closing (or 'gulping'-opening closing rapidly ande repeatedly- if too large for flow or throtteled flow).

2. Yes, but see point 1.

Good luck!

 

[gustorf]

From my point of view the SDV valve is only used for isolating the flow.
Typically the isolation valve is closed before the pump is stopped, and opened after pump restart.
The check valve shall only prevent reverse flow in cases the isolation valve is open and the flow decelerates.

I have to point out that under any decelerated flow any check valve of whatever design is closed AFTER flow reversal.
Even nozzle check valves dont close BEFORE flow reversal.
This is a term widely used in manufacturers literature as sales advantage, but physically impossible.
The question is - How soon after reversal is the check valve closed:
The sooner, the less risk of slamming.

Key figures for soon closing after reversal flow:
1) Short disk travel way
2) Low moving mass (light weight disk)
3) Strong springs supporting the closing action

In your case the check valve will close very slowly when the isolation will be closed.

 

[gerhardl]

Good points gustorf, and I agree of course!

( To clarify own posting: I did not say the nozzle check valve closed before the flow reversed. The valve will close as stated when the delta P across the valve is only fractions above zero. This will happen when the flow across the valve is zero.)

 

[laura80]

Thank you very much for your answers

To summ up the anwer to my questions are:

Conslusion of the second question:

-The check valve is going to close when the SDV totally close, even though the pump is runnin at nominal r.p.m.

To my first question I am not very sure of the answer, is controversial...

a) it depends of manufacturers and many others factors (Stanier)
b) always is going to close affter flow reverses (Gustorf)
c) is going to close when the flow across the valve is zero (Gerhardl)

In my simulation I have considered an "ideal" check valve that suddenly closes when the flow is zero, in 0 seconds.

Maybe this is not very conservative...but I don't have any way to do it with the information I handle.

Thank you very much for your help