Alter exisitng heat exchanger conditions 1

[cal6404]

Hi,

I've been asked to examine the theoretical output conditions of an existing HEX if the cold side input flow and temperature were to change.

I've simply used a temperuate gun on the HEX while it's been running to get the stream temperatures. The flows are recorded by flow meters. It only has one running condition at the moment.

Is it valid to assume constant UA? Altering the level of turbulence will surely change this significantly.

Thanks!

 

[Caloooomi]

If you want it to be a very rough estimate then yeah, Q = UA LMTD and go for it. While you can assume that A will stay constant, U will change as your cold stream heat transfer coefficient will change greatly.

 

[ione]

The surface of the heat exchanger is a geometrical bounday and consequently A will be constant. The overall heat transfer coefficient U, on the other hand, is a function of the temperature difference and the velocity of the fluid streams (not to quote the type of fluids, which in your case doesn't matter), so changing the flow and temperature of one side of the HE will alter the heat transfer coefficient.

 

[cal6404]

Thanks for getting back to me.

I guess my real question is, if all stream flow rates and temperatures are known, but ALL details of the HEX are not (tube count, heat transfer area etc), then single stream entering has its temperature and flow varied, is it possible to calculate the new output stream temperatures? How is the new UA value calculated?

If you assume temperature of a stream entering changes but flow rate doesn't, the output streams are strait forward to calculate since UA does not change.

I've just assumed constant UA in my calcs. Gives a reasonable result. Just curious how one would perform this task when constant UA isn't assumed.

Cheers!

 

[ione]

In order to calculate U you must know (rating procedure of HE) or make assumptions on (design procedure of HE) on geometry and on heat transfer area A.

 

[Drexl]

UAF = const is a reasonable assumption in many cases where changes are not so big and highest accuracy is not required. However the next step would be to adjust U for changes in flow velocity, assuming this is your biggest change. If you try calculating the heat transfer coefficient on both sides to see which one is limiting, then write that as k(w), heat transfer coefficient as a function of flow velocity you get a good feel for how U scale with w. In many cases it will be something like U ~ w^0.8.

 

[cal6404]

Thanks!

Makes sense. Let me confirm:

1) 1/U = 1/h1 + 1/h2 + R
2) If stream 1 flow rate changes, then only h1 will change which directly related to velocity change of fluid and remaining values used to describe h1 remain constant (d, Cp, j, etc).
Hence, U = f(velocity)
3) Get HEX temperautres and flows at 2 conditions and find U = f(velocity) relationship
4) Calculate new theoretical output conditions

Yeah this confused me since I know absolutely no details of the HEX other than the input/output flows and temperatures.

Cheers

 

[sshep]

Check out the GPSA manual which has affinity rules for your case, assuming there is no phase change. They even work a rerate problem similar to what you describe.

Best wishes,
Sshep

 

[MortenA]

Constant UA can only be assumed if the flow rates are (close to) constant.

The relation is that Nu is directly proportional to h and Nu=a*(RE^b)*Pr(c)

where Re is flow dependant (and large temperature changes may also cause Re and Pr to change since there is a viscosity dependency

I would caution aginst assuming too much here.

 

[Krazylion]

Using the NTU-(e) method will give a good insight.

Good luck

 

[cal6404]

How can I use the NTU method if I don't know the heat transfer area?

 

[amustha1407]

On such condition where there are too many unknown parameter, I would suggest to plot the UA for various time first. If you have sufficent historical data with fluctuation, you will get understanding on how different deltaT affect the efficiency.

If you think the correlation obtained above is valid, you can simply extrapolate to get the UA to be used.

 

[Krazylion]

Sorry for the late reply,

The general NTU method is as briefly as follow:

you determine the effectiveness of the HE as: e=q/q max
where q max= (C) min *(Thot,in- Tcold,in) where C min = mCp minimimum, it can be the one of the cold fluid or the hot fluid

Besides, NTU = UA / C min

Then once you calculate the e, depending on the type of HE you can calculate the NTU, based on the equations reported on the literature.

However according to your problem I think that is enough for you to determine the e for your HE, which does not change if the only modification to your HE are the T of the fluids, according to: q= e*(Cmin)*(Thot,in-Tcold,in)

Then you can use the regular eq: q= (mCp)(Tout,hot-Tin,hot) , same for the cold fluid.

Hope this helps.

Regards