Research On Hybrid Ice Detection Technique For Aircraft Engine Air Intake

During the flight of the aircraft,the icing of the engine air intake will occur,which will reduce the flow cross-sectional area of the air intake and affect the normal operation of the engine,and will affect the flight safety and cause safety accidents.Aircraft engine air intake icing detection technology can detect the icing condition of the air intake,realize the identification of different states of the air intake environment and the measurement of ice thickness.However,due to the characteristics of large wind speed and large temperature change in the air intake environment,there are false alarm and low accuracy of ice thickness measurement in the typical icing sensor.In order to improve the performance of aircraft air intake icing detection system in terms of state identification and ice thickness measurement,this paper proposes an engine air intake icing detection method based on complex impedance icing sensor and multi-spectral icing sensor.Research and experimental work on hybrid icing detection technology has been carried out.Firstly,the working principles of complex impedance icing sensor and multispectral icing sensor are studied,the theoretical models of complex impedance and multispectral icing sensor are established,and two kinds of icing sensors are simulated and analyzed.The variation trend of the interpole capacitance,the variation trend of the ray trajectory of the transmitting fiber under different ice thickness conditions,and the variation curve of the luminous flux of the receiving fiber under different wavelengths are studied.Secondly,on the basis of theoretical analysis,a hybrid icing detection system is designed,including a complex impedance icing sensor subsystem,a multispectral icing sensor subsystem and a data processing platform for data acquisition,analysis and visualization.Further,based on the machine learning algorithm,the verification experiments and result analysis of the complex impedance and multispectral icing sensors were carried out.Among them,the state recognition accuracy of the complex impedance icing sensor is the highest,which is 96.0%,and the root mean square error of ice thickness measurement of the multispectral icing sensor is the smallest,which is0.1525 mm.Finally,based on the above experimental results of complex impedance and multispectral icing sensors,the complex icing sensor data and multispectral icing sensor data are fused at feature level to verify and analysis the hybrid icing detection system.The experimental results show that the fusion system has a maximum increase of 11.8% in the state recognition accuracy,and a maximum reduction of 55.3% in the root mean square error of ice thickness measurement,which verifies the advantages of the hybrid icing detection system.