Preventing water hammer from damaging pumps and pipes

[Charli K. MatthewsKeymaster]

WHAT IS WATER HAMMER?

Water hammer occurs when the flowrate of fluid in the pipe changes rapidly. It is also known as “surge flow”. It can cause very high pressures in pipes, very high forces on pipe supports, and even sudden reversals of flow. It can cause burst pipes, damaged supports and pipe racks, and leakage at joints.

Water hammer can occur for any fluid, in any pipe, but the severity varies depending on the detailed conditions of the fluid and the pipe. It usually occurs in liquids, but it can occur in gases. It can cause pipes to burst and structures to collapse.

This article will describe the conditions most likely to lead to water hammer problems and the issues that pump and pipe designers and operators can face. It also outlines some of the ways to resolve the problems.

Read more: http://empoweringpumps.com/preventing-water-hammer-from-damaging-pumps-and-pipes/

[famiriParticipant]

Dear Charli
I faced a problem regarding water hammer . I have designed a pump station with a closed loop working on 7 barg at the pump outlet. When ever one the two working pump stops , a suddenly back pressure comes from working pump towards sopped pump and suddenly slammed the swing check valve , considered on the pumps outlet. So what can I do to solve this problem.
Best regards
famiri

[Charli K. MatthewsKeymaster]

famiri,

Thanks for the question… Here is the response I received:

On the face of it his problem is a water hammer one, but not the type associated with long pipes. As one pump stops (and I presume it’s a centrifugal type), reverse flow can flow through it once the impellor slows down enough, which can be rapidly if its moment of inertia is small, which it probably is. With a significant pressure rise over the pumps, there’s a big driving force for reverse flow, so reverse flow can build up rapidly. He put non-return valves on the outlets, which is necessary for a parallel pair of pumps. However he used swing check valves, which are the ideal way to create water hammer – the reverse flow builds up significantly before it drives the valve shut, when the slam stops the flow suddenly, hence the spike pressure and probably noise in the system. There’s a chance that repeatedly doing this will lead to a mechanical failure.

How does he fix it? He has two alternatives:
1. Use motors with much higher moment of inertia to prevent the reverse flowrate building up much before the valve closes. Ideally fit them with flywheels. However these are not readily available. The only one I installed we had to make specially for the job.
2. Use a better non-return valve. There are valves made specially to counteract this problem. They are spring-assisted or fluid padded and close gently before any significant reverse flowrate can build up, hence the problem is avoided. The detailed design isn’t simple, but manufacturers will help select he right one. Suggestions: Goodwin (http://www.checkvalves.co.uk/en/products/nozzle-check) , or Val-matic http://www.valmatic.com/checkvalves.html

[Randal FermanParticipant]

Famiri,
Goodwin’s two alternatives are precisely the ones that I thought of when I read the about water hammer problem. Ignoring the manifold piping in the pump station itself, if the system is basically a single conveyance pipeline from point A to point B, then it is relatively easy to perform a time-step water hammer transient analysis using a spreadsheet. You will need the pump’s rotational inertia data and a quadrant curve from the manufacturer, or as a first cut, you could use one of the curves provided in Stepanoff or other pump texts. The pressure surge/water hammer affecting the swing check valve might be the valve itself slamming shut, the sudden change of flow when one pump shuts down, or a combination of both.
I would not attempt any changes to the system without first doing at least an elementary water hammer analysis. Model the system and possible ‘fixes’ and decide if a higher level water hammer analysis is appropriate.
Good luck,
Randal

[Author]

Posts