Pipe bend radius and erosion due to turbulence

[Hareti]

Hello all,

I have some sea water cooled heat exchangers being installed and afew of the bends in the stainless steel piping are extremely tight. I have brought this up with the engineers but I am making no headway with them. I have done a fair bit of searching on the internet but mostly get information on the effects of bend radii on the material of the pipe, not the flow of fluid in the pipe.

Is there a rule of thumb that will give a guideline for the radius of a bend, given the fluid velocity in the pipe, to reduce turbulence at the bend?

If anyone could point me in the right direction to where I could read up on the subject that would be great.

Thanks,
Gareth

 

[BigInch]

The most important parameter is velocity. Clean seawater shouldn't have excessive erosion, however if you are paranoid about turbulence, keep the velocity where you have Reynold's numbers below the transition zone.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[BigInch]

Guess I should also mention your heat txf rate may go down.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[SJones]

And noting that stainless steel is more resistant to erosion than carbon steel.

Steve Jones
Materials & Corrosion Engineer

http://www.pdo.co.om/pdoweb/

 

[JohnBreen]

The problem that has gotten international attention is bursting of piping components (elbows, Tee's, and straight pipe near changes in direction) in power plants. But as my colleagues above have pointed out, this is caused by erosion-corrosion in high pressure, high velocity carbon steel piping (e.g., boiler feedwater piping systems). In these cases, the E-C occurs due to turbulent flow that will scour away the protective oxides of iron that normally form inside the piping components. Stainless steel will not have this problem.

Regards, John.

 

[SJones]

API RP 571, sub-clause 4.2.14

Steve Jones
Materials & Corrosion Engineer

http://www.pdo.co.om/pdoweb/

 

[Hareti]

Thanks for the info.

From the replies it seems that erosion corrosion will only be an issue if the sea water is very high pressure and velocity. Also the fact that stainless doesn't rely on the protective oxides of carbon steel reduces the problem further.

What I don't understand is BigInch's comment on Reynold's number. Even if the Reynold's number for a particular flow predicts laminar flow in a straight pipe, surely that flow could become turbulent at a bend, then return to laminar flow after the bend. The calculation for Reynold's number doesn't take into account the shape of the pipe.

Gareth

 

[BigInch]

You're right, but a lot less turbulence develops at the bends if the straight flow is laminar. You're thinking all the water particles hit the pipe and bounce off, but actually you tend to get more of a velocity decrease at the shorter radii and increase at the longer radii rather than turbulence.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[Hareti]

I had a go at calculating the velocity of water that would give a Reynold's number of 2000 just to get some idea of what it would be. Pipe diameter 50mm, and fresh water at 20 degrees C to start with.

density p = 1000 kg/m3
viscosity u = 0.001 PaS
diameter D = 0.05 m

v = (R * u)/(D * p)
= (2000 * 0.001)/(0.05 * 1000)
= 0.04 m/S

4 cm per second? That's so slow, I was expecting the transition to turbulence at a much higher velocity. This means that fluid in all the pipes in the system are well within the turbulent region.

There is something that bothers me about the Reynold's number calculation and it's probably just evidence that I'm in over my head, but I'd like to understand more about it. According to the formula, an increase in pipe diameter increases the Reynold's number for a given velocity. Now the turbulence is caused by the effect of the pipe walls on the fluid as it passes by, so surely a larger pipe means that a lower proportion of the volume of fluid is affected by the pipe walls as it moves, so there should be less turbulence. Why does a larger pipe diameter make it more turbulent?

 

[BigInch]

Re=[ρ]*V*D/[μ] increases linearly with diameter, at the same time velocity (holding flowrate Q constant) equals V=(Q/Area of a pipe) is decreasing by the square of the radius -> net decrease in Re with increasing diameter.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[gerhardl]

OK- You have a heat exchanger, SS for seawater, and are asking for optimum seawater velocity in the piping to avoid corrosion caused by turbulence in the thight bends.

If this is within 'reasonable normal velocities' for similar exchangers, I do, as others, not belive this to be the most critical issue.

More critical for designing will be the quality and origin of the seawater. Seawater in the Red Sea will not be equal to North Sea, Mexican Gulf or brackish water - different properties and temperature can alter the chemical and corrosive properties considerably. The higher (outgoing) temperature more crtical than the entering.


Every North Sea offshore material engineer will know this, and oil companies offshore often do not allow SS for seawater handling at all, even if this is commonly used in shipbuilding. If SS is selected, material quality will be most important, and have to be of proven for seawater from the area and similar applications.

Also consider in the practical use that all sand and shell rests should to be filtered from the seawater, abrasion from this do probaly do more for the bends tah Reynolds at normal velocities, and think about inside growth in exchanger pipes and cleaning methodes for pipe inner wall.

Outside cleaning?

If you want to use pigging for the exchanger the sharp bends may limit your choice of pig type.

Also to be considered: how critical the application is and the required lifetime length. If critical you might want to exchange the material specification to exotic.

 

[crvlad]

Use a long radius bend with a Boron infused internal surface and a conditioning vane on the inlet of the bend.