Research On Probabilistic Damage Tolerance Analysis System And Key Technologies

The object of damage tolerance evaluation is to ensure that aircraft structures can satisfythe residual strength requirements without major damage and harmful deformation. Itspurpose is to ensure the safety of aircraft structures in the inspection interval at high reliabilityrequests. The conclusion of traditional deterministic damage tolerance analysis is qualitative,in which the definition of "high reliability" is not very clear. Based on the reasons above, thefixed value method to evaluate the crack propagation life of structures has its limitations.Probabilistic damage tolerance technology can be more accurate and more quantitative toevaluate the reliability of residual strength, the reliability of specified crack propagation lifeand the crack propagation life of specified reliability. Probabilistic damage tolerance analysishas become an important direction of engineering technology as well as an important part ofthe field of solid mechanics. As for the random variables considered in probabilistic damagetolerance and the main contents and methods of probabilistic damage tolerance analysis stilllack consensus in academia, it has not been widely accepted by the theoretical system, and thekey technology in analysis is also imperfect. So research of probabilistic damage toleranceanalysis system and key technology has not only academic value but also practical value inengineering.This dissertation consists of two parts. The first part is about the research on theoreticalsystem and key technology of probability damage tolerance analysis. The analysis processand the common analysis method of probability damage tolerance are proposed, the randomfactors considered in probabilistic damage tolerance analysis studied and the statisticalanalysis method given by the random variables analyzed. Considering the residual strengthanalysis, crack growth analysis and reliability analysis as the main content of probabilitydamage tolerance analysis, the analysis system of probability damage tolerance is thereforeestablished.The p-a-σ curve is proposed to show corresponding allowable stress values underdifferent crack size at specified reliability. Based on the probabilistic characteristics of thestructure failure critical condition, the probability distribution function expression of thecritical crack size has been obtained. The stochastic models of crack propagation based uponrandom process and random variable respectively are proposed. According to the modelproposed above, an analysis method to get crack propagation life under specified criticalcrack size and the flaw size distribution under given propagating time are given. For the core content of probability damage tolerance (reliability analysis), three categories method ofreliability analysis are proposed: sampling method, probability space transformation methodand the response surface method, where the random sampling method and the probabilityspace transformation method are traditional methods in probability damage tolerance analysis,but the response surface method is first applied in the damage tolerance analysis. Based on thelinear elastic principles of failure of structures containing cracks, the response surfacefunction and random Hermite polynomial response surface function is put forward， suitablyto analyze the reliability of structures containing cracks under given life. The response surfacefunction proposed in this dissertation needs sampling only for the initial crack size and thecrack propagation rate parameter, which reduces the computational complexity.The second part of this article studies some new methods and new content of theprobability damage tolerance analysis.Firstly, interval analysis is promoted to put forward a new non-probabilistic reliabilitymodel, which is suitable for the systems with different distributions of the random variablesand applied to the crack propagation life analysis. Monte-Carlo stochastic extended finiteelement method proposed based on Neumann is used for fracture reliability analysis ofcracked structure.The probabilistic damage tolerance analysis methods of multiple crack structures are alsostudied. In view of whether there is a mutual influence between cracks, three kinds of crackstructure are studied: independent multiple crack structure, multiple sites damagestructure(MSD)(including crack interaction) and multiple local multiple sites damagestructure. According to the features of these three kinds of structure, the probabilistic modelsof damage tolerance analysis are established and practical analysis methods are givenrespectively.On the basis of sensitivity proposed by Millwater, the sensitivity of distributionparameters for the random variables in probabilistic damage tolerance analysis is analyzed inthis dissertation. The finite difference method and kernel function method of sensitivityanalysis are also given and applied to analysis of MSD structures. Secondly, considering theinfluence of variation range of the random variables to the failure probability, a newsensitivity index of failure probability to a random variable distribution is defined and itsanalysis method is determined.