Page 338 - Mechanical Engineers' Handbook (Volume 4)
P. 338
5 Use of Computers in Thermal Design of Process Heat Exchangers 327
heat transfer rates and pressure drops under the new conditions. This mode of calculation
can also be used to monitor apparent fouling resistance increase on operating units in order
to better schedule maintenance.
5.5 Program Quality and Selection
All heat exchanger programs are not created equal. Heat exchange is not yet an exact science
and all of the heat transfer coefficients and friction factors used in calculation are from
correlations with empirically determined constants. Therefore, the data base used for corre-
lation development is important.
Methods Source
The methods used for the program should be available and documented in a readable form.
Good methods will be based on theoretically derived equation forms that either are limited
in range or automatically achieve theoretically justified limits. ‘‘Black box’’ methods, for
which this may not be true, should be avoided.
Data Base
Good programs are also backed by a sizable data bank covering the range of conditions of
interest as well as demonstrated successes in predicting field performance. No non-tested
methods, including so-called rigorous incremental methods, should be accepted without some
data-based support.
Suitability
Completely general programs that apply to all geometries and process conditions and fulfill
the above data base requirements probably will not exist for sometime. The program manual
should list recommended ranges of applicability. When in doubt, consult the supplier.
5.6 Determining and Organizing Input Data
As of this writing, available programs still require a large number of input data decisions to
be made by the user. The quality of the answers obtained is crucially dependent on the
quality of these input decisions.
Process Data
The basis for the calculation is the heat duty, which usually comes from the process flow
sheet. There must, of course, be a heat balance between the hot and cold sides of the
exchanger. The temperature profiles are much more significant to a good design than are the
heat transfer coefficients. Only in rare cases are these straight lines. For multicomponent
phase-change cases, the condensing or vaporization curves should be calculated by a good
process simulator program containing state-of-the-art vapor–liquid equilibria methods. Most
good heat exchanger programs will accept these curves as the basis for the heat-transfer
calculations.
It is important to specify realistic pressure drop limitations, since the heat-transfer co-
efficient and the fouling rate are functions of the velocity, which is a function of the available
pressure drop. For phase change, too much pressure drop can mean a significant loss in
available temperature difference and one rule of thumb suggests a limit of 10% of the
operating pressure. For liquid flow, erosion velocity often is the limiting factor, and this is
usually taken to be in the range of 7–10 ft/sec tubeside or 3–5 ft/sec shellside. Velocities