water hammer

pimpa · July 14, 2014, 15:36

Hello everyone,

i crunched through related threats which was educative already. While at it, a question formed:

how deal with a negative pressure surge (suction) behind/downstream of the pump when it suddenly get's switched off (on purpose or by power failure). What should the inlet and outlet BC's of the domain (tube run) be for that? I'd think a steady state should establish the field (v, p) and a subsequent transient resolve the supply failure w the suction:

- massflow(over time) at the inlet (pump exit)
- pressure at the (remote) pipe run outlet ?

What would I model: adaquate mesh & time step for given wave speed.
What do I want to get out of it ? I aim at the pressure over time at the inlet, and along the pipe run.

Advice welcome.

ghorrocks · July 14, 2014, 18:19

Before starting this you need to do some basic estimates. First of all - is the time scale of the pump rotation speed change faster, slower or the same as the fluid time scale.
* If the pump time scale is faster than the fluid time scale you can probably assume the pump stops instantly and your simulation can be modelled as a transient fluid simulation only, with no pump motion. This is a simple simulation.
* If the pump time scale is slower than the fluid time scale then you can probably assume that the flow stops through a series of steady state conditions at a series of pump speeds. This is a simple simulation.
* If the pump time scale is similar to the fluid time scale then it probably is coupled and you will need to consider a full simulation of the fluid and pump together, probably coupled with the pump speed. This is a far more complex simulation.

Which of these three options is relevant to you?

pimpa · July 15, 2014, 12:56

Hello Glenn,

what do you intent by the fluid time scale ?

ghorrocks · July 15, 2014, 18:46

Time scale means an approximate measure of the amount of time something will take. In turbulence this could be the eddy turn over time, in normal flow this could be the convection time from the inlet to the outlet, or for a pump rotor it could be the amount of time it takes for the rotor speed to double is speed.

For instance if the fluid time scale (time taken for the fluid to go from the inlet to the outlet) is 1 second but the rotor time scale (time taken to double the rotor speed) is 1 hour then you know the fluid time scales are far faster than the rotor speed, and this can mean you can make simplifying assumptions.

pimpa · July 16, 2014, 07:34

Hello Glenn,

thank you for clarification ! Time scale for fluid passing from inlet to outlet, about 100 secs. Time for pumps to shut down or valves to close, some second. Time for the wave to travel the pipe run a fraction of a sec.

Any further hints appreciated.

July 14, 2014, 15:36	water hammer	#1
pimpa New Member Join Date: Mar 2009 Posts: 9 Rep Power: 17	Hello everyone, i crunched through related threats which was educative already. While at it, a question formed: how deal with a negative pressure surge (suction) behind/downstream of the pump when it suddenly get's switched off (on purpose or by power failure). What should the inlet and outlet BC's of the domain (tube run) be for that? I'd think a steady state should establish the field (v, p) and a subsequent transient resolve the supply failure w the suction: - massflow(over time) at the inlet (pump exit) - pressure at the (remote) pipe run outlet ? What would I model: adaquate mesh & time step for given wave speed. What do I want to get out of it ? I aim at the pressure over time at the inlet, and along the pipe run. Advice welcome.

July 16, 2014, 07:34		#5
pimpa New Member Join Date: Mar 2009 Posts: 9 Rep Power: 17	Hello Glenn, thank you for clarification ! Time scale for fluid passing from inlet to outlet, about 100 secs. Time for pumps to shut down or valves to close, some second. Time for the wave to travel the pipe run a fraction of a sec. Any further hints appreciated. Last edited by pimpa; July 17, 2014 at 08:40.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Mass imbalance problem in multiphase water and steam CFX case	Antech	CFX	1	October 26, 2020 04:03
Getting the pressure waves in Fluent for a water hammer problem	teopetre	FLUENT	2	August 6, 2013 19:24
Water subcooled boiling	Attesz	CFX	7	January 5, 2013 03:32
Terrible Mistake In Fluid Dynamics History	Abhi	Main CFD Forum	12	July 8, 2002 09:11
uptodate water distribution network	fredius,magige,tanzanian,(e.a)	Main CFD Forum	0	January 27, 2002 07:10

July 14, 2014, 18:19		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Before starting this you need to do some basic estimates. First of all - is the time scale of the pump rotation speed change faster, slower or the same as the fluid time scale. * If the pump time scale is faster than the fluid time scale you can probably assume the pump stops instantly and your simulation can be modelled as a transient fluid simulation only, with no pump motion. This is a simple simulation. * If the pump time scale is slower than the fluid time scale then you can probably assume that the flow stops through a series of steady state conditions at a series of pump speeds. This is a simple simulation. * If the pump time scale is similar to the fluid time scale then it probably is coupled and you will need to consider a full simulation of the fluid and pump together, probably coupled with the pump speed. This is a far more complex simulation. Which of these three options is relevant to you?

July 15, 2014, 12:56		#3
pimpa New Member Join Date: Mar 2009 Posts: 9 Rep Power: 17	Hello Glenn, what do you intent by the fluid time scale ?

July 15, 2014, 18:46		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Time scale means an approximate measure of the amount of time something will take. In turbulence this could be the eddy turn over time, in normal flow this could be the convection time from the inlet to the outlet, or for a pump rotor it could be the amount of time it takes for the rotor speed to double is speed. For instance if the fluid time scale (time taken for the fluid to go from the inlet to the outlet) is 1 second but the rotor time scale (time taken to double the rotor speed) is 1 hour then you know the fluid time scales are far faster than the rotor speed, and this can mean you can make simplifying assumptions.