Efficiency of overdesigned heat exchangers

[Iradah]

Hello all,

I am building a simulation model for our LPG plant. I was assuming 15 oF as delta T approach for the heat exchangers, but then I realized that these heat exchangers are now handling around 30% of the original design flow. So are 15 oF still valid, I think that the heat exchangers should now be less efficient being operating away from the design conditions, can I know your views and suggestions. Many thanks.

 

[macmet]

More flow would mean higher velocities and more turbulence in the heat exchanger, which is more efficient.

What type of heat exchangers do you have? If you're referring to plate and frame HX you may be able to remove some of the plates which would help.

 

[Iradah]

I have shell and tube heat exchangers, and the flow now is only 30% of what it was originally, so I think the heat transfer coefficient should be lower now, and for that when I build my simulation I should not assume that at the end of the heat exchanger the difference between the shell T and tube T is only 10 to 15 oF. I am assuming that is between 25 to 30 oF.

 

[MintJulep]

It depends on what you think that "efficiency" means as it relates to your model, and what, if any controls are in the heat exchanger and what the purpose of your model is.

 

[racookpe1978]

I think you are misunderstanding the purposes and role of your "model" and your "real world" - the plant itself!

Your "model" (the equations and the constants and measurements and areas and flows and deltaT that goes INTO the equations) is NOT the real plant. Here, and in your original question, you seem to be trying to force the real world to become the efficiencies (the flows and deltaT) that your model has. Reverse that!

in your plant, measure the inlet and outlet temperature of both liquids on both sides of your heat exchanger. Get your real world, actual flows of both liquids going through the heat exchanger. (These will include the 30% value on one of the sides you mentioned above, right?)

Now, go back to to the "model" (the program or spreadsheet (I've seen both used)) and adjust the "constants" of efficiency and heat transfer coefficient and wall thicknesses - area should not change from the original design, unless you've removed plates as mentioned above) until the "model" results are the same as the "real world" results.

Then you can begin thinking about changing flow (get a variable speed pump to save power costs?) or change area (pull plates if you have a plate-type HX) or change inlet temperature or whatever.

 

[Iradah]

Thank you all for your support. I agree with you, I was building a simulation on a program similar to HYSIS called VMG Sim.
some of the real temperatures at the plant are not available, so I was assuming in my model a 15 oF delta T, however, I realized that the end result is not matching actual conditions. that is why I thought that the heat transfer is not very efficient with overdesigned heat exchanger, lower velocities and lower heat transfer coefficient.

 

[MintJulep]

Again, it depends on what "efficiency" means to you.

The purpose of a heat exchange is either to make a cold fluid hot, or make a hot fluid cold.

There are only 5 heat flows of any possible interest:

Hot in
Hot out
Cold in
Cold out
Shell losses

Whether shell losses are important to you depends on what you are trying to do, and how the exchanger is arranged.

 

[racookpe1978]

If the "thermowell" temperatures are not available for your plant HX, make them available: Get a IR thermometer (even Home Depot or Lowes has them (US), you can also order on-line. Pull 3 inches of insulation on the pipe leaving the HX, or aim the IR thermometer at an exposed flange or pipe.

Write down the results, then go back to your computer model.

 

[MisterE2]

Odd question, there could be an increase in effectiveness of the heat exchange with more time for the fluid to pass through, but you could be losing a bit of heat to convection/conduction around the system, are you looking in to insulating the system's hot components for the lower flow requirement?