A detection method and detection system for large-area icing of aircraft

Abstract

The invention belongs to the technical field of aircraft ground icing detection, and discloses a large-area icing detection method and a detection system for an aircraft. The collected multi-angle images of the aircraft surface are preprocessed and labeled to obtain a training data set. According to different label types, the obtained training data set is divided into ice type training data set and ice accretion coordinate training data set; according to external environmental conditions and For the icing points distributed in each area, use the ice accretion coordinate training data set to train the icing coordinate network model; train the ice type recognition network model according to the divided ice type training data set, and output the divided icing pictures of each area of ​​the aircraft fuselage to be detected. According to the ice accretion ice type, the formula for calculating the icing intensity grade is established, and the amount of deicing fluid is calculated according to the icing intensity grade. In the aspect of the calculation of the ice accretion intensity level in the present invention, the accuracy of the traditional calculation formula of the ice accretion intensity level is improved by introducing the external environmental variables obtained by the sensor into the calculation formula of the ice accretion intensity level.

Description

technical field [0001] The invention belongs to the technical field of aircraft ground icing detection, in particular to a large-area icing detection method for an aircraft and a large-area icing detection system for an aircraft. Background technique [0002] In cold weather, because the moisture in the air condenses on the surface of the aircraft, it will cause the problem of aircraft icing. The impact of aircraft icing on flight safety is multi-faceted, mainly reflected in the following aspects: 1) Under icing conditions, The maximum lift coefficient of the aircraft and the slope of the lift curve decrease, and the drag and critical stall speed increase; 2) Wing icing can lead to a reduction in the airfoil stall angle of attack, and empennage icing can lead to tail stalls, especially under flaps In the wash flow; 3) asymmetrical icing may cause additional rolling moment; 4) icing on the tail of the aircraft may produce greater viscous force and increase the difficulty of c...

AI Technical Summary

Problems solved by technology

[0007] (1) Existing studies have seldom studied the icing area and freezing point distribution on the aircraft surface, but focus on the ice type and thickness of the icing, less consideration is given to the icing area, and cannot comprehensively Reflects the objective state of the aircraft when a large area of ​​ice is formed

[0008] (2) In the prior art, the same network model is usually used for the fitting output of various identification functions of ice accretion, and the fitting effect of each individual index is not good

[0009] (3) At present, images are used to identify disease thickness or sensors measure indirect variables to determine icing ice types and icing intensity levels, but there are few studies that combine the two with multiple evaluation indicators, which cannot fully reflect the icing status of the aircraft surface. It is easy to cause a large deviation in the amount of deicing fluid used

[0010] The difficulty of solving the above problems is: for different recognition functions, it is necessary to design multiple neural network models with different emphases, and it is necessary to fully investigate the core difficulties and data characteristics of the task, and the design is difficult

Images