Abstract :
[en] Frost accretion in heat exchangers of refrigeration systems is a major issue involv- ing energy consumption penalty in the heating or cooling devices. Numerous studies try to suppress, or at least delay, this frost formation. Among them, the use of (super-)hydrophobic coatings has shown encouraging results. The present thesis aims at investigating one step further the use of such materials in a heat pump evaporator.
The first step is to understand the physical phenomena involved in such process. A first classical heat pump has been built and has been highly instrumented, especially around the evaporator. At this stage, the surface of the fins and tubes evaporator is classical aluminum. It allows to get a very fine understanding of the frost problematic in these devices. In parallel another test rig is set up, to understand how frost may behave on superhydrophobic surfaces on elementary geometries, such as flat plates. Beside the the physical phenomena observation, those test benches allow to build a experimental data set.
Based on observations, new simulation models are implemented. A segment-by- segment discretization is envisaged for the modeling of the evaporator. It allows to get independent frost layers on each tubes, as observed during the experimental campaign. An originality of this model is to account for the fin thermal conductivity, which is determinant in the frost distribution through the device. For numerical robustness, the model is implemented as a dynamic one. A modeling work is also conducted at the surface scale. It accounts for major parameters impacting the macroscopic scale, such as contact angles or roughness. The major objective of this model is to predict the nucleation time (defined here as the time necessary, for a given surface, to be fully covered by frost nuclei).
The next step is, still separately, to compare the measurement results to the model predictions. At the evaporatior scale, the results are compared for the refrigerant side and for the air side in dry, wet or frosted conditions. This allows to successfully val- idate the model and clearly underline the effect of the fin thermal conductivity. At the surface scale, the same task is conducted, leading to the validation of the model.
As models can now be trusted, the ultimate step of the work is to merge them to predict the performance of heat exchanger in frost conditions, with different surface characteristics. The main result found there is that a real frost delay can be observed compared to regular surfaces, only if the hydrophobic level is sufficiently high. Slightly hydrophobic materials do not have any significant impact while superhydrophobic ones are game changers.
