Heat exchanger question

[Cagliostro]

Hi,

I'm modeling a heat exchanger for a customer. The values that they gave me don't make sense based on what I know about heat exchangers.

The heat exchanger is a single shell steam heater. It is now in service as a liquid/liquid heat exchanger.

T hot in=330
T hot out=240

T cold in=200
T cold out=260

As you can see the outlet temperatures of the liquid streams cross. How is this possible for a single shell design? I always thought that for the outlet temperatures to cross that you would need a 2-shell HX.

 

[TrevorP]

Counter-current exchanger
Hot in / Cold out - 330 / 260 = 70°C dT (Hot end)
Hot out /Cold in - 240 / 200 = 40°C dT (Cold end)

Draw a length/temperature diagram and you will see no cross-over with counter-current flow.

My question would be why such a large temperature drop on the steam side? Either there is a large pressure drop across the exchanger (unlikely) or the steam is superheated, in which case you will only have a very short length at the end at 330°C, and most of the rest of the exchanger will be at 240°C. Then you really will find it a challenge to achieve 260°C on the cold stream!

 

[Cagliostro]

Trevor,

It was origninally a steam heater they converted it into a liquid-liquid heat exchanger.

I had understood it that T cold out was not suppose to be hotter than T hot out.

F the crossflow correction factor is quite low, actually off the chart that I am using (single shell/2-pass).

 

[Montemayor]

Cagliostro:

What do you mean by “modeling a heat exchanger for a customer”? Are you “rating” the capability of that existing and operating exchanger? If so, in what kind of service and in what orientation? What is the purpose of your “modeling”? In the absence of your use of correct units, I have to assume you are dealing with degrees Fahrenheit and that all your heat transfer is sensible only and doesn’t involve any heat of vaporization.

Please furnish the complete TEMA classification for the exchanger in question. If it is not a conventional TEMA design, then detail out the physical characteristics of the unit. As you should be aware, the physical configuration of any shell and tube exchanger is very important and sets your “F” factor.

Two fallacies you have been exposed to are corrected below:

1) It is very possible to achieve a temperature cross in a single heat exchanger shell.
2) It is very possible for T cold out to be hotter than T hot out. It all depends on the relative temperatures and the configuration employed.

Your description of the configuration is misleading. “Single shell/2-pass” doesn’t mean anything of importance without your designation as to the 2-passes; are the 2-passes on the tube side or on the shell side? --See what I mean about the TEMA classification? For example, you could be applying an F TEMA shell, and this would give you 2-passes on the shell side.

Contrary to what you state, the existing liquid temperatures don’t seem to cross. They seem perfectly normal – depending on the exchanger’s configuration.


Art Montemayor
Spring, TX

 

[TBP]

I won't say which one, but I work with a heat exchange manufacturer, and I have their selection software. Using the temps provided, and assuming some flows, I can get possible selections to come up for either water to water, or steam to water service. For steam service, it's not really very hard to get some subcooling on the condensate.

As a side note, you very much do NOT want the steam to have any amount of superheat. Superheat is sensible heat, and you have to get past that to get to the latent heat. And the change of state from steam to condensate is where all of the action is.

 

[25362]

As Art Montemayor says, it is possible to have a cross of temperatures in a one-pass shell unit. The configuration, the Cp of the fluids, their relative flow rates, an over-surfaced unit, are all factors that may contribute to the crossing of end temperatures.

One can imagine that if the unit was previously a horizontal steam heater with steam on the shell side, the shell baffle cuts may be vertical. If this is the case the shell-side fluid would be moving from one end to the other in practically a "horizontal unmixed" path .

Thus the first tube pass would appear to work in a kind of parallel flow while the other would approach a countercurrent pattern which could induce a temperature cross.

 

[tickle]

Please define modelling of the heat exchanger - I would have thought HE a specialist would be required

 

[Cagliostro]

Art,

First, thank you for clearing up my misconceptions on HX's.

The configuration is 1 shell, 2 tube-pass and the temperatures are in Fahrenheit.

'Modeling' I guess is a non-descript term. The flows and temperatures of the streams to the heat exchanger will be changing. I need to know how this is going to effect the output temperatures in the future.

From the current data I had Q transfered, A, and the dTlm. I wanted to back calculate U from Q=UAFdTlm. Once I had U I could itteratively calculate the output temperatures of the heat exchanger at the new flow rates/temps.

When I try to use the charts to calculate F(from the single shell/two tube pass chart) F does not fall within the chart (The charts I have stop after a F of less than .5). That is why I thought there was possibly something wrong with the original temperatures that I was given.

Nh=(Tcout - Tcin)/(Thin - Tcin)=(260-200)/(330-200)~0.46
Z= (Thin - Thout)/(Tcout - Tcin)= (330-240)/(260-200)~1.5

 

[TrevorP]

I suggest that you try to get access to as much original design info. as possible. Get out the drawings, the original design calcs (for the new liquid duty), and the physical properties of the two streams. This is an essential starting point for your model. You will then be able to see the exchanger configuration, including baffles etc., as described by Art Montemayor and 25362 above.

Are the temperatures quoted design or operating temperatures? If they are operating temperatures, then it obviously does work, even at a correction factor of below 0.5. A correction factor below 0.5 is generally considered to be poor design, but that's not to say that it won't work. With existing equipment the question generally asked is - "Will it work?" and "Is it fit for purpose?", not "Is it the optimal design?"

You have assumed in your description above that the 'U' value will not change. If the flowrates vary significantly, this may not hold true. That is why you need design info. and physical properties.

 

[StoneCold]

Caglistro
I am sure that I will get slammed for this by the experts but here it goes. This is down and dirty and will give you a very conservative answer that may say the exchanger is too small for the new rates and you will have to dig deeper into the problem as TrevorP mentioned. But if you are just changing the flowrates of the same fluids you can do this. Take Q=UAFdTlm and group UAF into one term. Solve the existing equation for the existing conditions. Apply the UAF to the new flowrates. If both flows are going up then U should be increasing as well as TrevorP suggested but this would give you a first crack at will it work.

Regards
StoneCold

 

[tickle]

If the flows are changing then the individual tube and shell side heat transfer coefficients are going to change.

Now the tubeside coefficient can be found relatively easily from Nu=0.0023 Re^0.8 Pr^0.33. The shell side coefficient is slightly more difficult, but there are approximations that can be used for cooling water with a superficial cross flow velocity across the tubes ho would be 4000 W/m2.K. Add your fouling and wall coefficients and viola a design for the HE. I am assuming that you are aware of calculating U from hio and ho etc.

 

[turbA]

Hi everyone!

I am having a similar problem with the analysis of a heat exchanger. I need to determine the 'U' value for an operating heat exchanger. The temperatures are as follows (in degrees C): Th,i=99.5; Th,o=54; Tc,i=46.7; Tc,o=78.9. My Nh and Z values as defined by Cagliostro are ~0.61 and ~1.4 respectively. These values do not intersect on the correction factor charts, which begin at F=0.5. Can anyone point me towards a chart whose F values start at a lower value, or another method for determining 'F'?

Thank You!

 

[25362]

To Cagliostro: Reading the graph in Ludwig's Vol 3, for one shell-pass, I found a correction factor of ~ 0.5.

To turbA: You are right. Your temperatures would correspond to a factor ~0.5, but for a two shell-pass unit. Have you considered the possibility of the exchanger having a longitudinal baffle ?

Ludwig credits TEMA for giving permission to use their graphs.

 

[turbA]

Unfortunately the heat exchanger is a single shell two tube passes unit. Hence my dilemma!! I'll check up on the possibility of a baffle, but I don't think that is the case... The hot fluid flows through the shell side and is initially steam, which is then condensed and further cooled. I'm not sure how this affects the LMTD calcs, which generally consider temps only but not whether or not there is a state change...