Full-scale runback ice: accretion and aerodynamic study

The runback ice phenomenon is well-known for anti-icing or de-icing systems when the system is not evaporating 100% of the water impinging the surface. The water runs back to the point where the added heat no longer raises the surface temperature above freezing. The water freezes behind this limit. No runback ice is tolerated for some flight configurations, but not for all. Then for o.-design cases, some runback ice may grow on the wings surface. However, data from full-scale realistic runback ice is not very well-known by aircraft manufacturers and they are not sure what thickness is allowed before the e.ect of the ice on the flow becomes too adverse. To better understand full-scale high-fidelity runback ice growth and how it can be simulated with simplistic shapes, test campaigns and CFD studies were undertaken. First of all, tests in the Cranfield icing tunnel were performed. In this work, full-scale runback ice shapes were grown on a model with a full-scale leading edge equiped with an electrical heating system. An innovative moulding and casting technique has been introduced which allowed the production of 3D planarised full-scale realistic runback ice castings. In parallel to the icing tunnel tests, a mass and energy balance has been computed on Excel. This energy and mass balance can predict the heat and mass fluxes involved in the runback ice accretion mechanism. Following this, aerodynamic tests of the ice castings were lead in one of the low speed wind tunnels at Cranfield University. The aerodynamics of simplistic shapes such as geometrical shapes or ballotini layers were also studied. The e.ects of the ice castings on the flow were compared to the e.ects of the simplistic shapes. The tests were done on a flat surface and not on an airfoil due to technical complications. The boundary layer displacement thickness was the parameter used to quantify the e.ect of the shapes on the flow. 2D CFD simulations were performed as a support to the testing but as well to compare with the experimental data. The CFD simulations were for steady or unsteady flow. It has been possible to grow full-scale ice shapes in a relatively small icing tunnel. The shapes have been successfuly moulded and cast using silicone and plaster mixed with polymer. A catalogue of runback ice shapes for different liquid water content, heat inputs and positions along the chord has been recorded. Following the wind tunnel tests, it has been possible to find a relationship between the real ice and the simplistic shapes. Thin runback ice shapes (4 mm) has a similar e.ect on the flow as a layer of 1 mm ballotini. It was found that thicker ice shapes, of the order of 1 cm, is equivalent to a rectangle with rounded corner, associated with 1mm ballotini. The triangle shape which is usually used to simulate runback ice by the aircraft manufacturers, was found to be the most aerodynamically penalising simplisitc shape that has been investigated in this PhD project. It was found that rounded corners greatly improve the representativeness of the simplistic shapes, such as triangle or rectangle.