5 replies to this topic

[Bodhisatya]

In Air Separation units we generally encounter multistream exchangers where numerous streams (hot and Cold) exchange heat for bringing down Air temperature to its liquefication temp.

 

While running a simulation in ASPEN HYSYS, performance page registers an LMTD. 

 

I understand that there won't be a single LMTD, might be a weighted one.But how do HYSYS calculate that.

 

 

Requesting forum members to help me in calculating the same.

 

Bodhisatya

[Bobby Strain]

And what will you do with this information? I believe that only a few companies who make these exchangers know how to design them. And they likely allow significant overdesign due to uncertainties. If someone can provide a reference on this subject, it will be interesting, if only from an academic view.

 

Bobby

[Pilesar]

If you were to model the exchanger as a large number of exchangers in series, then LMTD can be calculated for each of the 'subsection exchangers'. It is more useful to calculate a Weighted LMTD from the combination of these subsection exchangers because the stream temperature curves are not linear within the exchanger. If you were to try to calculate this yourself, plot the composite hot stream and cold stream temperature curves and divide the plot into multiple sections of equal area. Calculate the LMTD for each section and average them. For your multi-stream exchanger, the important constraint is 'what is the closest internal temperature approach practical?' This minimum internal temperature approach (MITA) will pinch the exchanger performance. If your operating MITA is far from the design MITA, then you definitely have some optimization opportunity.

[Bodhisatya]

 For your multi-stream exchanger, the important constraint is 'what is the closest internal temperature approach practical?' This minimum internal temperature approach (MITA) will pinch the exchanger performance. If your operating MITA is far from the design MITA, then you definitely have some optimization opportunity.

 

Dear Mr Pliesar

 

Thanks for contributing.It is the DTmin that we state during initial simulation and also we quote the same to the vendor, as you rightly pointed that dtMin remains a very important constraint.

 

I have plotted the Composite curves, but somehow getting an erratic result while calculating LMTD by the method prescribed by you.Can you please at your end and let me know the result.HYSYS has given me a value of 5 K.

 

PFA the excel.

 

Bodhisatya

Attached Files

•  CC.xls   27KB   27 downloads

[PingPong]

I am sure that Hysys is capable of calculating the correct LMTD.

 

What LMTD you have calculated from those composite curves is not clear to anybody. In fact it is not even clear how you calculated the composite curves to begin with. Moreover there seems to be very little temperature difference at the cold end (on the left), which may be the cause of the rather small LMTD of 5 K that Hysys calculated.

[Pilesar]

The curve you supplied shows temperature vs heat flow. Can you get composite curves for temperature vs exchanger length or exchanger area? The software will determine LMTD by analyzing many 'zones' internal to the exchanger. Each zone will have different duty and a different contribution to the overall LMTD. The pinched area at the cold end looks like a small area on the graph because there is little heat flow at the pinch. This graphed region might show much larger when plotted with temperatures on the Y-axis and exchanger length on the X-axis because pinched areas require a higher proportional share of exchanger area for the same duty.