**4. Discussion**

The results yielded by the present model and by HTRI Xchanger Suite for the air-to-water heat exchanger from the literature have shown that even though an apparatus can be decomposed into parts for which correlations or calculation procedures for heat transfer coefficients and pressure losses may exist, successfully applying them may not be straightforward. The main reason is that such procedures require local fluid and material properties, and these generally depend on the temperature and pressure, that is, quantities which the designer is trying to calculate. This is where the present model steps in. The data have also highlighted the facts that the accuracy of any model depends to a large degree on the quality of equations utilized for the calculation of the various coefficients and that further research in this area is, therefore, necessary. Additionally, one can draw the conclusion that using averaged values for the entire tube side and shell side can lead to notable differences. In the case of the outlet temperatures in this air-to-water heat exchanger, it was up to ca. ±7%.

As for the heat recovery hot water boiler, the most important stream parameters are listed in Table 7. It can be seen that the tube-side outlet temperature and pressure drop were better predicted by HTRI Xchanger Suite, while the accuracy of the shell-side outlet temperature and pressure drop predictions was higher in the case of the present model. Overall, the accuracy of the model was deemed acceptable concerning its intended purpose.


**Table 7.** Summary of the main results obtained using the two discussed approaches alongside the data provided by the operator of the boiler.

<sup>1</sup> Tube side (water); <sup>2</sup> Shell side (flue gas); <sup>3</sup> Estimated (not measured during the guarantee test).

A significant advantage of the model over the industry-standard design package offered by HTRI (or other commonly used heat transfer equipment design packages) lies in the fact that it provides detailed data on the tube-side fluid flow and temperature distributions. This information can be of great value when trying to prevent certain types of operating problems (e.g., excessive thermal loading of the tubes in the bundle and the subsequent mechanical failures). Another advantage is that in the case of cross-flow tube bundles (the primary target application of the present model), HTRI Xchanger Suite assumes that mixing of tube streams occurs after each pass, even if this may not actually be true. Such a simplification may increase convergence, but it also may diminish local effects and, therefore, introduce errors into the data.
