Study On The Reliability Data Screening And Imputation Of Civil Aircraft

Reliability is an inherent characteristic of the civil aircraft and an important factor of flight safety, synthetic effectiveness and life cycle cost. Therefore, the quality of the data, which is collected in the survey, influences the civil aircraft performance monitoring and new type plane developing, it is of great importance to conduct examination for reliability data quality.First, the dissertation makes an analysis on reliability data source and the features of civil aircrafts, and works out some data pre-solve methods such as data sample selection, data exchange and data reduction, in order to inspect the data quality.Focus on its outlier problem during the process of data collected, the dissertation researches on the mechanism in which outlier occurs and the guideline by which reliability data filters; also analyzes on the outlier inspection methods for fulfilling different distribution and compares between its pros and cons. Refer to the data items of high association degree, the dissertation not only sums up the outlier testing using regression methods, but also analyzes on their pros and cons by means of example testing, so as to confirm suitable outlier inspection method of reliability data.Then, aimed at data deficiency during the process of data collecting, the dissertation studies on the mechanism in which deficient reliability data occurs, puts forward filling-up guideline of reliability data and sums up the solving methods of deficient data; meanwhile, it investigates on the applicability of those methods, and explicitly demonstrates on MCMC method of multiple imputation and gives out the application examples.Based on the aforementioned research, the dissertation finally develops the filtering system of reliability outlier and filling-up systems of deficient data, plus describes its whole design process and introduces its developing frame and function module.