Abstract:

Icing has long been described as one of the most serious meteorological hazards in aviation activities because it affects aircraft's flight performance negatively or makes it completely unfit for flight. Structural icing on the aircraft occurs with water vapor in the air at freezing point temperatures or especially super-cooled water droplets hit the surface of the aircraft, causing the flight direction of the aircraft, such as the edges of the aircraft wings, the engine and its exterior surfaces. The two main factors for the formation of structural ice in an aircraft are temperature and the amount of liquid water in the air. To occur ice on the plane, the airplane must be flying in an environment with water droplets, and the ambient temperature should be 0 °C or lower. According to the Federal Aviation Administration (FAA), the risk of icing is shown in four different classes: trace, light, moderate and severe. Ice accumulating in the wings and on the outer surface of the aircraft disturbs the aerodynamic structure of the wing, resulting in reduced lift, increased drag, and increased total weight of the aircraft. This situation leads to rise in the minimum stall speed of the aircraft. In this study, three-input and one-output parameters affecting ice formation are used for estimating the risk of iciness by artificial neural networks. Input parameters of the proposed system are temperature, dew point and relative humidity. Output parameter is degree of icing. It has been determined that the above given parameters have an effect on icing. The output of the Artificial Neural Networks system, which will be designed with the effect of these, will give us the degree of icing.

Keywords: