Analysis and research of aviation safety risk based on PCA and MEST algorithm

With the continuous development of technology, the relevant technology for flight data recording is constantly updated. QAR overrun data can be used for aviation safety management and flight training data support. Based on the QAR data of a civil airliner in a certain flight segment, this paper aims to conduct dimensionality reduction processing for each index of the QAR data in this flight segment by using principal component analysis method, and analyze the importance of each flight parameter to aircraft safety. Based on the flight data after dimensionality reduction, the Multivariate multivariate algorithm was adopted to Multivariate State Estimation Technique (MSET). Based on the comparison between the aircraft safety parameters actually monitored in the flight segment data and the aircraft health data during normal flight, the alarm threshold of deviation values was set based on the Laida principle. The normal navigation parameters of the aircraft at this moment are calculated and estimated. If the threshold value is exceeded, the automatic alarm is set up.