Velocity in Sharp Elbow & Erosion 1

[MikePAMechEng]

I'm working on a heat exchanger fabricated from 90/10 CuNi (likely changing to 70/30) that includes a sharp elbow in the piping leading to/from the core. The elbow is essentially one pipe tee'd into another, like a branch flow coming into the run of a tee, only there is no run. There isn't room to put in a radiused elbow. We will be doubling the water flow rate through this piping to increase the thermal capacity of the HX. I'm looking for any reference to velocity flow or any experimental data on this type of fitting to help estimate erosion rates. I've found some things for short radius elbows, but nothing for a mitered elbow or this specific type of geometry.

Right now I plan to base the acceptable flow on the allowable wall shear stress for CuNi and assume a 60% increased wall shear due to the elbow. The 60% comes from experimental data showing increased wall shear in tubing entrances. The flow around this tight elbow will also have some contracted area with higher velocity and wall shear, but I wanted to find better data than just using the tube entrance value. Short radius elbows seem to have (max velocity)/(average velocity) ratios of ~1.6-2, at least in what I've found.

Has anyone looked at this before or seen any good references?

Thanks

 

[LittleInch]

Can you draw what you mean by a "sharp elbow". your description sounds like a blind ended tee?

Having difficulty believing a short radius elbow can't be used?

Or a tee with a blank end.

I've seen what are called "cushion tees" or target before now available in lots of different materials.

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

 

[MikePAMechEng]

Yes the geometry is like a blind ended tee or a target tee. The flow is both towards the "target" or towards the run depending on if it's the inlet or outlet side.

The joint is made by connecting two pipes with a saddle weld and capping the target end.

There really is no room, at least not for an elbow I've ever seen. The pipe is a .75 rod that is drilled out, and there is not room to move it up or increase the diameter to 1/2" pipe. It's sandwiched between the core return tubes and a mounting plate.

 

[EdStainless]

Move from CuNi to a 6%Mo superaustenitic or a superferritic stainless. I have seen both AL-6XN and Sea-Cure tested in seawater at >100ft/sec jet impingement.
Even with 70/30 7ft/sec will cause erosion.
Maybe you should be changing the alloy in the tubes also.
Moving to high alloy stainless you would go to 1/2 of the wall used in 90/10, or even less.
It more than makes up for the lower TC of the tube material. In the case of Sea-Cure your thermal performance would increase significantly.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[racookpe1978]

If you're only making one of these, machine a special elbow. ANY curve will improve flow significantly. Yes, it is more expensive. But years of service and pump costs override that cost of one special fitting.

 

[MikePAMechEng]

The velocity in the core tubes is <4 fps, so I'm not concerned about those. The average velocity in these other pieces will be 13.7 fps with infrequent flow up to 15.2 fps. But that is with the current ID. I own to increase the ID to drop the velocity to 12 fps.

One thing working for me is that it's fresh water, but I'm using the alowable shear values for sea water since it's hard to find fresh water data for CuNi.

I thought welding/brazing stainless steel to CuNi was discouraged due to iron in the steel causing poor weld life.

 

[EdStainless]

Who on earth designed a HX with 4fps in the tubes? That is below the design values for brass, much less CuNi.
No wonder you are trying to raise the flows.
CuNi has been brazed to 6%Mo superaustenitic stainless with high reliability.
Or you could use a Ni alloy such as 625 for the elbow to better resist erosion.
Even in fresh water 12fps is unadvised in CuNi.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[MJCronin]

EdStainless is spot-on .... as always

I agree that your best path forward is an upgrade of materials that are more erosion resistant.

With a tube velocity of < 4fps, I suspect that this is a repurposed HX .... Is that true ??

At < 4fps, I don't believe that you will even achieve achieve turbulent flow.

MJCronin
Sr. Process Engineer

 

[MikePAMechEng]

Thanks for all the responses so far. The funny part is, 4fps IS the faster flow in the tubes. This is a HX that was designed over 20 years ago. I think that since it worked, they called it done. Now we need more heat out but don't want to drastically change the design.

I have thought about using 625, but people are resistant to the cost. We may have to go that route.

As far as 12 fps being too high, The ASM materials handbook has 10-12 fps in tubes and these pieces are thicker walled than the tubes. Some mil specs use similar values. Is there a reference that you've found that shows this is too high, or a factor to go from sea water to fresh water?

 

[EdStainless]

Just basing my comments on what I have seen on the inlet end of condenser tubes with 90/10 in fresh water.
The tubes looked fine at 8fps, but there was clear inlet end erosion in less than 10 years of service.
Who cares is this fitting is $1,000? If you gets your system running reliable then it is well worth it.
Make it from 625, braze it in, and be done with it.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube