Flow Turbulence 1

[jkate]

Hello:

A client is complaining that there is excessive turbulence inside a filter as the water enters the vessel and is asking what to do about it. Attached is a drawing they provided of the piping arrangement. My first observation was two things. It seems they are reducing the pipe size very close to the filter upstream of the last elbow before it enters the filter. This is increasing the water velocity and causing turbulence before the water enters the filter. Also, the elbow before entering the filter is very close to the filter inlet not allowing the water flow to smooth out before the water enters the filter. Are my observations wrong? Is there anything else in the discharge piping that can be improved?

Thank you.

 

[1503-44]

You could try that.

You might also check the pump flowrate against the rated capacity of the filter. It wouldn't be the first time someone tried to operate in "go to fast" mode. Speaking of which, what is the flow rate and the pressure drop in the filter?


Is that a valve that can be partially closed? How much does it have to be closed before the turbulence becomes acceptable?

 

[jkate]

Thanks for your reply 1503-44.

The optimum flowrate of the filter is 2100 gpm. The headloss is about 26ft.

 

[bimr]

Have not witnessed a problem with the filter piping before. Many times one will find that there is an elbow on the inlet to a filter which doesn't effect the filter operation.

You might check with the filter manufacturer. It is typical that the filter manufacturer furnishes some type of crude baffle inside of the filter on the inlet.

Is this a new operation? Is the filter designed for the flow?

 

[Rputvin]

How are they measuring turbulence? How much is "excessive"?

Are they complaining about noise, ripped up filter media, uneven filtration over the vessel, etc, or are they actually complaining about turbulence?

Edit:
DN 300 is close to 12" pipe. That can carry 3500gpm easily. For a vessel rated to 2100gpm I would expect 10"/DN250 (closer to 2500gpm pipe) piping to it if there is no flow control valving in place.

Check the pump, I bet there's a mismatch in there and their duty point / rated flow is a heck of a lot lower than their actual flow.

 

[jkate]

bimr & Rputvin:

Inside the filter is a nest of hundreds of tubes that hang down from an upper plate. Each tube is coated with a filter media. There is a baffle on the inside of the filter that intended to redirect the flow through the center of the filter theoretically spreading the flow move evenly up through the tube nest. This is a system that has been in operation for a while. I do not know if this turbulence is something new or if it's been ongoing. They are complaining that the turbulence causes the water and filter tubes to move around making the precoating of the tubes take longer.

 

[LittleInch]

"The optimum flowrate of the filter is 2100 gpm. " Great. What is the flowrate actually being experienced / capable of by the pump? you seem to be very shy about saying what this is despite being asked twice.

And don't give me the rated duty point of the pump - what flow is it actually doing?

If you want to try something then maybe insert a flow conditioner plate between the faces of the flange - like this

https://mattech.cz/en/produkty/deskovy-zankeruv-usmernovac-proudeni-264/

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[bimr]

Perhaps it is occurring because of a quick startup or shutdown. How long are the filter runs?

 

[jkate]

LittleInch, not being shy regarding the flowrate. I asked the client to provide. Still waiting for an answer. I was thinking about straightening vanes, but the perforated flow conditioner is a great idea. Thanks.

Bimr, filter runs are typically 6 hours.

 

[bimr]

Perhaps a slow startup and shutdown of the filter may help.

 

[LittleInch]

"Still waiting for an answer"

I think that tells you all you need to know about the client. It's the very first thing you would look at it provide data.

Those pages are great though.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[HTURKAK]

You are probably wright .. you may check Reynolds number .(Re=ρ*V*D/μ )

You see that Re increases linearly with diameter, velocity and diameter ..and reverse proportional with viscosity. If Re ≥ 4000, the flow is turbulent.

Could be two decades ago, we encountered a case at the inlet piping of a booster PS and cracks developed at the inlet piping .We have used plate baffles ( paralel plates located inside the bend and inlet pipe ).. and divide the flow multi zone . It has worked.

If you provide real flow , ID of pipes, and existing pipe supports , you may get better responds..