In selecting the recommended method for this handbook, the considerations summarized in Section 248 indicated clearly that, of all the methods surveyed, the Bell-Delaware method is in principle the most suitable one at the present. It fulfills the following requirements.

1. It is a method that can be handled readily by hand-held calculators, progressing in a stepwise manner and making use of graphical solutions of auxiliary variables. It is considered essential that the recommended method may be used in this way for educational purposes and other reasons discussed later. Even if complex computer programs are available, there is — and always will be — a need to calculate the performance of an exchanger by hand methods. Only if calculations are performed manually will the engineer develop a "feel" for the design process as compared to the impersonal "black box" calculations of a computer program.
   
   
2. The recommended method can be, however, readily programmed, possibly in parts, into even small programmable calculators. This is probably the most effective way of using the method, freeing the engineer from the time-consuming details of calculations and yet retaining the closeness of individual judgment.
   
   
3. The method is based on the principles of Tinker’s flow distribution model, and thus is much superior to "integral" methods, such as Kern’s or Donohue’s. It was selected in preference to the original Tinker method, which did not have the benefit of the Delaware data. Major overhaul of Tinker’s method would be necessary to achieve the same accuracy (Palen and Taborek, 1969).

It should be clear, however, that this nonreiterative method cannot compete for accuracy with the complex stream analysis-type methods, which require relatively fast computers for their solution. Nevertheless, for a "well-designed" exchanger without any extremes, the results are within respectable limits of accuracy (see additional comments later in this section).

A. Method structure

The method uses ideal tube bank j_i and f_i factors and then corrects directly the resulting α_i and Δp_i for deviations caused by the various split streams. The j_i and f_i factors are defined as follows: