Institutions Reliability Analysis Method

The mechanism reliability problems of aerocraft are common concern, and the reliability of some special mechanisms is critical for the safety of aerocraft. But, it is lack of systematic and professinal research focusing on this field. So, the analytical method of mechanism moving reliability is fully studied in this thesis.In chapter 2, the basic priciples of mechanism moving reliability are discussed systematically. At first, the reliability analysis methods of kinematical accuracy and dynamical accuracy of mechanism are summed up. And then, we discuss the relationship between joint-gap and effective length model, and propose the continuously contacted model. The relationship between wear and mechanism kinematical accuracy is analysed resumptively. At last, from the point of view of reliability analysis, the problem of dead-point seizure of mechanism is discussed, and the reliability analysis method of dead-point seizure of mechanism is presented.In chapter 3, simulation algorithms of mechanism moving reliability analysis are systematicallty introduced, and a brief analysis is made on essential, applicability, advantages and disadvantages of some popular reliability analysis methods. And then, a discussion on the way of introducing neural networks into complex mechanism moving reliability analysis is made. By checking computation results of several numerical examples, validating the applicability and reliability and making a good foundation for the forward reseach. At last, the reliability analysis of the elevator mechanism system of some model aerocraft is provided to show the applicability of neural networks method. It has been proved that this method can overcome the limitation of response surface method, and achieved high efficiency and good accuracy.In practical engineering, it is usually to perform an extreme small-scalesample test if sample is a large-scale complicated extreme expensive structural system. So, the research focusing on estimation method for extreme small-scale sample test is made in chapter 4. Firstly, we discuss the Bootstrap method that was presented to estimate small samples test, and point out shortcoming of the Bootstrap method. And then, some methods are present to improve the Bootstrap method. At last, virtual expanding estimation method and synthesis probability distribution estimation method are presented to estimate extreme small-scale sample test.