Improving the prediction of the thermohydraulic performance of secondary surfaces and its application in heat recovery processes

Abstract

Accurate prediction of heat transfer and friction performance is a key issue in the design of heat exchangers. This is particularly important in technologies that use secondary surfaces such as the plate and fin exchanger, where the number of surfaces available for design is large. Sets of experimental thermohydraulic data are available for most surfaces, but its use in sizing methodologies requires the development of generalised correlations. Some of the available correlations cover a limited Reynolds number interval as fitting these expressions to the whole range, causes prediction errors to increase. This work looks at the development of simple generalised correlations for surfaces such as louvered, rectangular, and triangular fins. The predictions extend from the laminar to the turbulent regime. Compared to experimental data, the average absolute errors of predictions for louvered surfaces are between 4% and 12% for Colburn and between 3% and 8% for the friction factor. For rectangular and triangular surfaces, the error is 5% on both parameters. Use of these correlations in design gives maximum volume prediction errors of 4.78%, while the use of other published generalised correlations results in volume errors in the order of 91.85%.