Geometric characterization of aircraft icing

This thesis focuses on the characterisation of the internal and external structures and the synthetic reconstruction of aircraft ice accretion from small supercooled droplets on a cowl-lip model. Icing encounters provide challenging scenarios for civil aircraft. Not only the performance but also the safety of the flight is dramatically affected by such encounters. Particularly interesting is the accretion of glaze ice, where water film dynamics plays an important role in the accretion process. The study of the characteristics of this type of accretion focused on two main characteristics: porosity and roughness. A better knowledge of these features allowed to produce synthetic geometries, which has the same influence on the airflow as the original ice surface. Glaze ice encounters were reproduced in the Icing Wind Tunnel of the Institute of Fluid Mechanics of the TU Braunschweig. The porosity and roughness of the ice produced during these encounters were characterised using micro-computed tomography. Ice samples and silicone moulds of the accretion surface were scanned using this technique. Such measurements allowed to quantify the porosity of the ice accretion and to characterize the roughness on the surface. The results of this work show that the porosity of glaze ice accretion lies below 1%, which makes it impossible to find channels connecting the pores. This is an important result because it means the airflow does not penetrate the ice layer. The other feature affecting the airflow is the roughness of the accretion. It was observed that increasing the water droplet size increases the roughness of the ice accretion. Longer exposures to icing conditions tend to reduce the roughness of the ice surface. In glaze ice formations, the water creates a distinctive footprint when freezing on the ice's surface. This results in a surface exhibiting wavy characteristics. It is proposed to summarize all the roughness characteristics of an ice layer in 17 parameters. With these parameters, synthetic ice layers were produced using two different methods. The first method (SynthPS) uses only the information from the spatial distribution of height elements. The second method (SynthHPD)included both the spatial distribution and the height distribution of the roughness elements. The flow in the boundary layer over the real and the synthetic ice roughness was simulated. The comparison of numerical outcomes shows that synthetic surfaces represent the average flow behavior well. However, a more detailed analysis of the Reynolds stresses within the boundary layer reveals limitations in replicating the flow features when using the SynthPS method. The flow in the boundary layer at downstream locations of the real and the synthetic accretions was measured. The synthetic ice layer reproduces well the velocity profile of the boundary layer behind the ice accretion. Differences between the two methods are observed when analysing the turbulence profile and the energy spectra inside and outside of the boundary layer. Considering both spatial and height distribution (method SynthHPD) to generate synthetic surfaces ensures a better reproduction of turbulence and energy content on the flow.