The Civil Aviation Transport Aircraft Risk Assessment And Controllable Factors Analysis Method Study

In recent years, with the rapid development of China’s civil aviation industry and improvement of people’s living standard, the civil aviation safety problem attracts more and more attention. However, there are a lot of elements of civil aviation safety: Hardware reasons, also software reason; necessary factors, also occasional factors; the management system reasons, factitious factor causes. As a result, these factors evolved into hidden danger in commercial fights. As the main carriers of civil aviation transportation, the management mode, operation mode and staff quality, etc, of airlines directly or indirectly dominates the development of airline safety. The study mainly discusses the essence of several main factors which threat airlines security and how to construct the analysis methods to reasonably strengthen these factors, so as to avoid safety risk effectively. This study mainly includes two aspects: anomaly detection and approach analysis.This study mainly includes two aspects: anomaly detection and analysis of controllable factors, which is divided into five chapters that are the introduction, anomaly detection and approach study of civil transport aircraft, anomaly detection based on fleet aging, anomaly detection based on flight data, anomaly detection based on human factors.First of all, the thesis develops a systematic process of anomaly detection based on existing anomaly detection methods, which includes the framework of the risk assessment process and comprehensive risk assessment process.Every step of the process was carefully described, includes both qualitative methods and quantitative methods.Secondly, the thesis develops anomaly detection from three aspects, including fleeting aging,j it take on the risk analysis from three aspects separately of the fleet of aging, the flight data and human factors.Anomaly detection based on fleet aging: The feasibility of applying Weibull analysis to validate fleeting wearing and aging was proved by evaluating datasets of civilian flights. Anomaly detection based on flight data: this part intrudes Gauss mixture model to conduct clustering analysis. The main goal is to combine cluster analysis with this model, forming a cluster-analysis calculation method based on mixed Gauss model. This thesis realizes the verification of feasibility of the applying Gauss mixture model to cluster-analysis.Anomaly detection based on human factors: it is highly difficult to evaluate the importance of the characteristics of pilot’s behavior during flight, thus this part introduces method of principal component analysis to find out main characteristics that pilot behaviors. This thesis developes "bridge logic gate" management model for pilots, to assist the airline to manage pilots and the entire flight process.