Evaluation of the liquid channel influence on heat transfer performance in a plate-fin heat exchanger by means of three-dimensional CFD simulations

Abstract

Paper presented to the 10th International Conference on Heat Transfer, Fluid Mechanics and Thermodynamics, Florida, 14-16 July 2014.

Compact Heat Exchangers (CHEs) play an important role in a wide range of applications, e.g. in windmill gear units, machine tools, mobile hydraulic systems and so on. For each field, specific requirements are needed (i.e. heat transfer rate, mass, size and pressure drop), which are achieved also through different types of configuration for the fin geometry in both the oil and the air channel. By means of CFD simulations of a small part of the CHE core, it is possible to know how a certain coupling of air and oil channels performs. However, when the number of configurations to be analysed is consistent, it is preferable to choose the smallest meaningful computational domain, in order to reduce computational resources while keeping all important physical phenomena. The purpose of this paper is to demonstrate that, despite oil flow can lead to convective heat transfer coefficient several times higher with respect to air flow, a change in the fin geometry affects significantly the “conjugate heat transfer” and the CHE performance. Besides, it is presented a simplified model for the heat exchanger element where the oil channel conductive and convective heat exchange are modelled by using a fixed temperature boundary condition and an effective thermal resistance. 3-D simulations were carried out considering a fixed fin geometry for the air channel and five different fin geometries for the oil channel. Two cases of operative conditions were taken into account. Furthermore, for each coupling of channels, the simplified model was applied. Results demonstrate that a change in the fin geometry for the oil channel affects the overall heat transfer, and this influence is greater or smaller depending on the operative conditions. Secondly, the reduced model is shown to yield results with a reduction in accuracy that can not be neglected.