CFD analysis of heat exchangers can be useful as a lot of these items were manufactured years ago: before computer flow modeling became effective. These units stay in service for many years and upon refits or maintenance turndowns can be tweeked to improve performance.
A primary concern in CFD analysis of heat exchangers is the fanatacism of the modeler in reproducing the geometry accurately. Often computerised CAD models are not available and what one has to work with are sets of large format blueprints. From these the tube bundle outlines can be scanned and the 3-D model generated. In doing these CFD models we take great care in making sure that baffles and channels are correctly reproduced. Naturally this is important as the flow finds all sorts of ways to bypass the bundle and flow around the outside of the shell, thereby short-circuiting the exchanger duty.
Another important parameter is matching the process results, obtained from a general purpose process modeling package like Hysim or Prosys to the allotted duties in the cfd model sections. Baffles are placed so as to insure the flow takes a certain path thru and along the bundle and the process generated sections need to correctly correspond with the geometry sections in the cfd model.
Also condensation can be modeled as well as the accumulation of non-condensibles. Given the capital cost of shell and tube heat exchangers, the relatively low cost of cfd modeling of them, and their initial design from a pre-cfd period, it is usually beneficial to do cfd models of shell and tube exchangers if one wishes to squeeze some additional capacity out of them.
We really need to put some more pictures up but we prefer not to illustrate clients’ designs even when after 40 years the exchanger designs are hardly unique. Thus additional illustrations are awaiting a “generic marketing” cfd model of a shell and tube exchanger.