Karlanmanın ısı transferi üzerindeki etkisi: Deneysel inceleme ve sayısal model geliştirme

Water molecules freeze when the temperature drops below the triple point. This phase change process results in the formation of ice crystals. With air filling the gaps between the growing ice columns, together this porous medium known as "frost". The accumulation of frost creates a thermal barrier across a surface due to the low thermal conductivity of the frost layer. This phenomenon is commonly observed in heating, ventilation, and air-cooling applications, such as evaporators in household refrigerators, air source heat pumps, and aircraft precoolers. Therefore, it is crucial to investigate the dynamic heat transfer rates across surfaces operating under frosting conditions to assess their performance and develop defrosting strategies. In this thesis study, the effect of frosting on heat transfer is investigated experimentally, and the gathered data is utilized in a numerical frosting model development. A large number of experiments are performed by altering the parameters of free stream air velocity, relative humidity, and cold plate temperature. Frosting impact on heat transfer rate is measured through a high-precision thin film heat flux sensor. Accordingly, heat flow is significantly obstructed due to frosting in time. The rate of decrease in the heat flux is not found to be constant over time but shows a decreasing trend. Additionally, an enhancement in the heat transfer is reported during the early stage of frost formation under certain environmental conditions, due to the increase in effective heat transfer area induced by ice crystals and the corresponding increase in latent heat generation. The densification of the frost layer and its effect on the total thermal resistance is discussed. Based on the convective heat and mass transfer balance over the frost layer, frost surface temperature was predicted. A correlation for estimating the heat transfer decrease during frosting is proposed via non-dimensional numbers. The gathered empirical data is utilized to establish a robust numerical model. An Eulerian-Eulerian multiphase approach is followed to model humid air and frost phases separately. Heat flux during a dynamic frosting process is also calculated during the frosting simulations, and validation of the proposed model is done over the experimental heat transfer measurements. Frost accumulation on the cold surface is modeled with an empirical mass source term. Model constants are tuned in a systematic way using experimental heat flux and frost thickness data. A velocity dependent model constant is introduced into the mass source term. The heat flux rise observed experimentally at the initial stages of the frosting could be captured with the use of the velocity dependent model constant and the addition of this term considerably improves the accuracy of the frost model at the initial stages of frosting. The developed numerical model is tested with three different frost thermal conductivity models. Using the thermal conductivity of solid ice for the frost thermal conductivity results in the most accurate prediction at the early stages of the frost growth process indicating a rather column-wise vertical growth of ice crystals with very low lateral branching. However, the overprediction of the numerical heat flux with the thermal conductivity of solid ice points out a decrease in the thermal conductivity of the newly added frost layers indicating a more pronounced lateral branching of ice crystals within the frost layer. The effect of the diffusion coefficient of the water vapor in humid air on frosting is also investigated. An artificial increase in the diffusion coefficient improves the accuracy of the heat flux prediction of the model at the initial stages of frosting which might indicate an eddy-driven enhanced mixing in the boundary layer which might not be captured in the laminar flow model. Finally, the developed numerical model is also tested on another scenario where surface temperature is set differently.