10 replies to this topic

[clarenceyue]

Hi Majster:

The "L" in the Seider-Tate equation refers to the overall length if I am not wrong.

[Andres Valencia Michaud]

Well, including the hairpins, every heat transfer equation is meant for the overall lenght. If still in doubt I recommend you to read a simple book like Donald Kern's Heat Transfer Processes.
Greetings.-

[breizh]

Hi ,
An other correlation for laminar flow :

http://www.wlv.com/p...abook/ch2_4.pdf
with L = total length

Hope it helps

Breizh

[David Southall]

Hi

I am trying to estimate if existing double pipe heat exchanger is suitable for other duty than originaly intended.Due to very viscous liquid in inner pipe, there is laminar flow. For calculation of heat transfer coefficient in laminar flow, I used Sieder Tate equation:

Nu=1,86*(Re*Pr*(D/L))^(1/3)*(eta/eta_wall)^0,14

Heat exchanger is constructed from 6 metres long straight tubes in overall length of 120 metres.

I have folloving dilema: should I use L=6 m or L=120 m ?

Best regards,

Majster

Dear Majster,

To add to the responses that you have received so far, the length you should use is 120 metres. The ratio (D/L) helps quantify the enhancement to the heat transfer coefficient that occurs due to the entrance effect.

At the entrance to a duct, the flow is not developed, enhancing the heat transfer coefficient, but then develops with length along the duct, with the enhancement tending towards zero as the length tends to infinity, The length to use is therefore the length over which the flow develops.

The exception in your example would be if there were breaks or interruptions to the flow between tube lengths, such that the flow is disrupted and can be assumed to be undeveloped at the entrance to the next tube section. In that example, the length to use in your equation would be the length between breaks, 6 metres.

Hope this is helpful.

Best regards,


David.

[riven]

While estimating the heat transfer coefficient is important I would first get an idea of the pressure drop.

You state that this unit was used in another service and that the current service is viscous. If the previous service was not viscous then your design is unlikely to be acceptable for viscous fluids. I would ascertain this first before getting into the more complicated heat transfer requirements.

[David Southall]

Hello everyone,

thank you very much for your answers, especially yours David. I forgot to mention, that heat exchanger is built from 6 m long segments connected together with elbows. Does elbow in laminar represent enough distorsion to the flow to assume undeveloped flow at entrance to next 6 m long section?

Best regards,

Majster

Hi again,

I read article on link Breizh sent. All my doubts are gone. There is written, that L is length of single pass so 6 m in my example (breizh you did not read carefully-L is not total length). Article also states, that in case of U tube L is length of straight portion of tube from tubesheet to the tangent point of the bend. In the bend there is secondary flow, which disrupts main flow, so it is not developed as it enters next straight portion of tube.

Thank you all again especially you Breizh.

Best Regards,

Majster

Dear Majster,

Firstly, you're welcome - I'm glad I could help. It looks like you are well on the way with justifying the length you need to use.

One other way to look at this problem, particularly where you want to quickly estimate the performance of the existing exchanger, would be to 'bound' the problem. We can anticipate that the shortest developed length you could have would be the length between elbows, 6m. The longest could be 120m. If you are using a spreadsheet or similar to estimate the performance, it should be straight-forward to evaluate Nusselt numbers under both assumptions.

This would produce the two extremes that the answer could take. Sometimes this type of approach is close enough provide the answer you need, and can be quicker than trying to resolve or justify which value of L to use.

It does seem reasonable for there to be some degree of flow disruption around an elbow - certainly there are methods to evaluate the enhancement due to the change in direction (I think Shah or Joshi may have published a paper on this). This effect can be ignored to provide a pessimistic Nusselt, or included to provide an optimistic Nusselt, or both to bound the problem.

Best regards,

David.