Water velocity in Shell and tube HE

[wilsoncheung]

Hello,everyone!

As you may have designed chiller or used chiller. Do you know the ordinary water velocity in shell and tube heat exchanger. I have read a lot of books, mostly suggest it to be about 2m/s.It's said by ASHRAE that no damage will happen to the copper tube even the water velocity exceeds 3m/s. I wanna know your experience from your design or your application, especially in flooded chiller.

Your any comment will be appreciated!

Wilson Cheung

[Josip]

Hi, wilsoncheung :)



Hello,everyone!

As you may have designed chiller or used chiller. Do you know the ordinary water velocity in shell and tube heat exchanger. I have read a lot of books, mostly suggest it to be about 2m/s.It's said by ASHRAE that no damage will happen to the copper tube even the water velocity exceeds 3m/s. I wanna know your experience from your design or your application, especially in flooded chiller.

Your any comment will be appreciated!

Wilson Cheung

minimal velocity is 0,5 m/s and preferred is 1 - 1,4 m/s, but this required speed is opposite proportional to the most common request for a minimum pressure loss across the heat exchanger!

see this ... a lot to read&learn ;)

http://www.neatbook.org/w/wolverine%20tube%20heat%20transfer%20data%20book.php

Best regards, Josip :)

[wilsoncheung]

Thanks, Josip!
Yeah,minimal velocity is specified to prevent fouling.But is it the preferred velocity is 1-1.4m/s? Low velocity will have low Reynold number,and low Reynold number will be in laminar flow. So in flood chiller, the preferred is more 2m/s,I just wonder what's the maximal velocity should be.

[chillerman2006]

Hello,everyone!

As you may have designed chiller or used chiller. Do you know the ordinary water velocity in shell and tube heat exchanger. I have read a lot of books, mostly suggest it to be about 2m/s.It's said by ASHRAE that no damage will happen to the copper tube even the water velocity exceeds 3m/s. I wanna know your experience from your design or your application, especially in flooded chiller.

Your any comment will be appreciated!

Wilson Cheung

Hi wilsoncheung

Shell and tube systems I have commisioned were all

3 metre's per second

Chillerman :D

[US Iceman]

The allowable fluid velocity for the tubes is an attempt to limit erosion of the tubes without having too much tube side fouling. Copper being relatively soft can use velocities up to 8 feet per second (as an estimate). Stainless steel tubes can accommodate a slightly higher velocity.

You need to stay under these maximum velocities however, you can be constrained by the actual heat exchanger configuration for velocity versus allowable pressure loss.

[wilsoncheung]

Hi wilsoncheung

Shell and tube systems I have commisioned were all

3 metre's per second

Chillerman :D

_____________

Really? All 3m/s? So how about the water pressure drop, I'm sure it will relatively higher.

And what do you base on to design this velocity?

Thank you!

[Entropie]

Minimum and maximum flow rates are always related to the engineered datas of the heat exchanger (design flow rate). The minumum value of 1 m/s for heat exchangers is with regards to the laminar flow and increased fouling level. The excessive flow rate as already said, can cause erosion. In dx coolers for example excessive flow will disort the pass baffles and hence damage the tubes as well.

regards,
Entropie

[valterb]

Dont go down the minimal. Laminar flow is dangerous (no exchange)...
usualy on shell and tube i attempt to acheve 2,5 m/s more/less.
On plate heat exchanger, you can go slowly, you dont have laminar flow because for design of plate.

[wilsoncheung]

Thank you valterb and Entropie!

[TXiceman]

Many engineers will place limits on water flow without regard tot he operating parameters.

Typically the upper flow velocity is set at 8 to 9 fps (feet per second). Lower limit is usuallu 3 fps. Some will try to limit the flow velocity to 5 to 8 fps.

The lower velocity will not creat an errosion problem, but will create problems with laminar flow. Laminar flow equates to poor heat transfer. You need the flow to be in the turbilent regiem for better heat transfer.

By going too high on the velocity, you can get into errosion problems at the tube enterance area. Certian fluid can experience reduced heat transfer ratees if you take the velocity too high.

The rules change when you go from a shell and tube design to a plate design. Also on a DX where the cooled fluid in on the shell.

Ken

[desA]

I agree with the range of 1 - 1.5 m/s as being a typical tubeside design norm.

To achieve the desired Reynolds number, the designer plays with tube inner diameter, number of tubes, & pass arrangements.

Once heat-transfer has been met, then the pressure drop is determined - hopefully it is under the allowable value. If not, iterate on tube dia etc.

[smpsmp45]

You check with the ACME catalogues & they give out very good details on this

[wilsoncheung]

We have talk so much about the velocity? Do you mean the tube side velocity?

What about the shell side velocity in DX evaporator? There should be velocity limit to prevent vibration, right?

[desA]

Yes, there is a typical working range. It depends a lot on the type of baffle arrangement used.

You can find more information on the recommended velocity range for the baffle arrangement you're using, in industry design codes. Could be some interesting research - will have to search a bit. Enjoy. :)

[wilsoncheung]

Thanks, desA!

I have read books of baffle arrangement, but there are mostly about optimum design not the critical design. What I want to know is about the maximum or the minimum velocity on shell side.

Anyway, it's very kind of you to give me this tips.

[sterl]

Shell Side water velocity has (3) limiting bits of geometry:

At initial entry & exit: You will have the connected pipe velocity redistributing to wrap its way around the minimum number of "facing" tubes and spreading end to end. Result is recirc eddies which will cause erosion and sometimes some strange Heat Xfer. Worst cases involve impingement by anything solid coming down the pipe "blasting" tubes.

Recirculated treated water usually limited to 3.2 M/sec at entry; well water lower.

At Partition Plate Cuts: The velocity again goes high and by general pattern, goes from perpendicular to parallel and back to perpendicular. All this accelerating and momentum changes makes for good heat transfer but this second site for recirculation and erosion. TEMA and ASME III both have indicating tables based on shell diameter compared to baffle cut.

Induced Vibration: This to do with tube wall thickness and stiffness as well as partition baffle spacing and clearance baffle plate to tube. For a DX chiller its usually not an issue because the tubes need to tolerate the standby pressure of the refrigerant and the refrigerant side heat transfer is often worse than the water side, thus controlling in terms of optimizing the HEX. ...But it still happens. Best references are ASME III, Pettigrew & Gorman, and the studies by Au Yang et al.