13 replies to this topic

[Blaghness]

Hi,

 

I'm a little stuck on this issue and wondered if someone could help.

 

I want to design a shell and tube heat exchanger that takes a cool liquid, and heats it up to boiling temperature, then boils it. There is also a small bleed stream outlet. The hot side is also liquor so the temperature will change accordingly.

 

In this situation we have two functions, heating the liquor, then the evaporation. For the first part the LMDT is larger, and with the second part it's smaller. So I believe it's not valid to use the LMDT over the entire head exchanger. Is this correct?

 

How should one go about calculating the LMDT in this situation? Would it be a matter of calculating the sum of two separate areas for the separate functions (heating and evaporating) and summing them together for the total area?

 

The basic concept is a shell and tube with multiple passes.

 

Thank you!

Blaghness

[Bobby Strain]

The solution depends on the the fluids. Maybe some details will get a (better) response.

 

Bobby

[Blaghness]

Hi,

 

Sure, they are water based liquids with some dissolved solids in them, but for the purposes of this they could be assumed to be water.

 

Thanks.

Blaghness

[latexman]

Do you know what type S&T HX you have in mind?  That would help.

 

I'm not sure a shell and tube with multiple passes is ideal for this.  So far, I'm visualizing a kettle-type reboiler or a bayonet heat exchanger bolted onto a tank/pressure vessel.  The boiling will provide mixing of the cool feed into the bulk liquid.  That's based on my past experience though.  Others may think differently.

[katmar]

I would not try to perform both of these duties in a single exchanger, I would use a pressurized horizontal multi-pass unit for the heating duty and then drop the pressure (i.e. flash) via a control valve into a single pass vertical unit to vaporize the liquid.  And I would not accept a design from anyone else unless they could show me a reference plant.

[breizh]

Hi ,

Why don't you consider an evaporator to perform the task ?

https://www.lcicorp....ss-applications

Breizh 

[Blaghness]

Hi,
 
Sorry for the delay in replying.
 
I'm obviously not a heat exchange specialist, just someone who is trying to challenge our organization, so let me please clarify. The concept is a kettle reboiler, the basis design is a NKN design. Also, in our industry this is a fairly standard design, but I am not particularly well versed in the sizing, thus why I am here :-). The reason to not use evaporation is that this is creating clean steam from a dirty source - there is an existing evaporator solution also.
 
In reality the reason I am here is to see if there is a better way to do this as per @Katmar's comment, I believe it's perhaps not the most efficient way and there is a better way - your comment is very interesting, I am a bit confused about the function of the second single stage though - wouldn't it be enough to exchange the heat in a tube and shell exchange then send the heated liquor to a flash system, or is this what you mean?
 

I would not try to perform both of these duties in a single exchanger, I would use a pressurized horizontal multi-pass unit for the heating duty and then drop the pressure (i.e. flash) via a control valve into a single pass vertical unit to vaporize the liquid. And I would not accept a design from anyone else unless they could show me a reference plant.

 
Even so, I would like to start with the basis of design to form the argument for a better solution, so it would be great if I could get an answer for the original question, but if I am still not providing enough information please let me know and I can provide more details.
 
Thank you.
Blaghness

Edited by Blaghness, 19 October 2021 - 05:56 AM.

[Bobby Strain]

It would help if you supplied the flow and cold temperature. You need a free source for the heating media to justify steam recovery from what appears to be a waste stream. This is the deciding factor.

 

Bobby

[Blaghness]

This is a minimum viable product at this point, so there is not a set temperature. If we use typical though it would be:

 

Hot side - 150 deg C

Cold side - 70 deg C

Steam pressure: 0.5 bar

 

Both sides are water based solutions with properties very close to water.

 

The real question I have is about the LMDT though, the validation activity is something I can and have started to do myself.

[latexman]

T_steam @ 0.5 barg = 111-112^oC

 

With both sides using water properties, for the sensible heat LMTD the profiles may look like:

 

70.00  --->  111.5

108.5  <---  150

 

delta T = 41.5^oC = LMTD since both are equal.  Beware the temperature cross.

 

Note - relative mass flow rates were not given.  I assumed equal mass flow rates, which may not be correct.  Does this help?

[Pilesar]

Using only the temperature difference at the ends of the exchanger is not adequate since there will be a non-linear heat curve internal to the exchanger. The exchanger designer (whether human or computer) will need to know the temperature for both streams at points internal to the exchanger and also the duty required to get those streams to reach those points. The properties of the streams (heat capacity, thermal conductivity, surface tension, etc) important to heat transfer will also change as the streams pass through the exchanger. A complete heat curve is required to properly design your heat exchanger, Optimal design for single phase heat transfer is different from the optimal design for phase change. Compromised designs which try to do more than one service aren't always the best choice.

  Your responses suggest you are not close to ready to design this heat exchanger. If you supply a heat exchanger vendor with enough information for the heat curve, they might propose a design.

[Blaghness]

T_steam @ 0.5 barg = 111-112^oC

 

With both sides using water properties, for the sensible heat LMTD the profiles may look like:

 

70.00  --->  111.5

108.5  <---  150

 

delta T = 41.5^oC = LMTD since both are equal.  Beware the temperature cross.

 

Note - relative mass flow rates were not given.  I assumed equal mass flow rates, which may not be correct.  Does this help?

It would not be equal flow rates, it would basically we a reboiler system meaning with a 5-10% bleed off stream.

 

In this case though the heating from 70 deg C to 111.5 deg C has a high driving force increasing the LMDT, but in reality the majority of the heat is transferred at 111.5 deg C which is a much lower driving force. Wouldn't one then want to separate into two phases - the heating and the evaporation phase and use the LMDT for the respective areas? -- This is the real question. Using the LMDT as calculated above would potentially undersize the heat exchanger, or am I misunderstanding?

 

Thank you.

 

Best Regards,

Blaghness

 

Relative flows would be say 100 kg/s for the hot stream and under 10 kg/s for the clean liquor side.

[Blaghness]

Using only the temperature difference at the ends of the exchanger is not adequate since there will be a non-linear heat curve internal to the exchanger. The exchanger designer (whether human or computer) will need to know the temperature for both streams at points internal to the exchanger and also the duty required to get those streams to reach those points. The properties of the streams (heat capacity, thermal conductivity, surface tension, etc) important to heat transfer will also change as the streams pass through the exchanger. A complete heat curve is required to properly design your heat exchanger, Optimal design for single phase heat transfer is different from the optimal design for phase change. Compromised designs which try to do more than one service aren't always the best choice.

  Your responses suggest you are not close to ready to design this heat exchanger. If you supply a heat exchanger vendor with enough information for the heat curve, they might propose a design.

OK this is what I thought. At this phase I am trying to prove a concept rather than actually design a heat exchanger, so I would think a simplified curve like my last reply would be sufficient for the concept phase. I personally would never be designing the final design of the exchanger but instead would, as you say, rely on the experts for the final design.

 

The problem I am actually trying to solve is whether we can change a two phase heat exchanger to a single phase and reduce investment cost and maintenance costs. The heat exchanger I am describing in this post is the existing solution, which is a "proven solution" in our industry that I am trying to evaluate if we can replace with a new system. In order to do this I need a basis of design, thus why I am posting. The proposed solution, being a one phase heat exchanger, is much simpler to design.

[latexman]

I was just throwing an idea up against the wall in my post #10 to elicit more information.  It was never meant as design information.

 

If you are trying to design one HX for the total duty of both phases, I suggest to construct and study a graph of delta T vs % of total heat duty curve to get an idea of the effective LMTD to use.