Research On Reliability Sensitivity Method Of Structural Mechanisms With Time-dependent Problems

Sensitivity analysis is one of the important analytical methods to determine the failure sources of structures and mechanisms,and is widely used in practical engineering.Since the structural mechanism with time-varying problems generally involves stochastic process,the sensitivity analysis is more complicated and difficult to implement in practical engineering problems,especially when engineering problems require time-consuming finite element software to solve,the computational cost will increase greatly.At the same time,how to combine the finite element software to analyze the sensitivity of engineering problems involving stochastic process is of great significance to the application of sensitivity analysis in practical engineering problems.This paper takes the structure involving random processes and motion mechanism as the research object,the reliability and reliability sensitivity of the structure under random excitation and the motion mechanism are studied based on the dynamic strength formula of the first passage method and the Possion's assumption-based first-passage method,respectively.Simultaneously,sensitivity analysis-based dimension reduction optimization strategy is proposed by combining sensitivity analysis with traditional optimization method,and is applied to the constraint support optimization of a certain complex high-dimensional aviation hydraulic pipeline.The main research contents are briefly described as follows.(1)Studied the reliability sensitivity analysis method of random structures under random excitation involving finite element software analysis.Based on the dynamic strength reliability formula of the first-passage method,combined with ANSYS finite element software,the reliability of a complex aeronautical hydraulic pipeline system under random excitation is studied,and a non-probability global sensitivity index is proposed.Secondly,aiming at the maximum stress response,the non-probabilistic sensitivity analysis of the clamp support position of the aeronautical hydraulic pipeline system is carried out,and the support positions with no or weak influence on the objective function are filtered out according to the sensitivity results.Finally,the optimization scale of the aeronautical hydraulic pipeline system is reduced by considering only the remaining design variables that have a large influence on the objective function in the optimization analysis,so as to improve the pertinence and purposed of the optimization.On this basis,a dimension reduction optimization strategy based on sensitivity analysis is proposed and applied to the sensitivity analysis of the mean time between failures(MTBF)of the aeronautical hydraulic pipeline system,and the design variables that have an important influence on MTBF are determined.Based on this,the MTBF of the aeronautical hydraulic pipeline system is optimized and improved.(2)Studied the application of the weight-point methods in the sensitivity analysis of the random structure involving random excitation.There are three most widely used weight-point estimation methods,namely three-point estimation method and Gaussian integral-based weight-point estimation method and sparse grid method.Firstly,the basic principles of the three kinds of weight-point estimation methods are introduced.On this basis,the application of weight-point estimation method in two kinds of moment-independent importance measures is studied,and the core idea and specific implementation steps of calculating the two moment-independent importance measures using the weight-point estimation method are given.Secondly,the application of weight-point estimation method in variance-based sensitivity analysis is studied and proposed a loop-nest sensitivity analysis method based on weight-point estimation method,and then,applied it to the sensitivity analysis of the structural parameters of key parts of aeronautical hydraulic pipeline system under the stochastic excitation,and the structural parameters that have important influence on its key parts are obtained.Combining the weight-point estimation method with the multiplicative dimensional reduction method(H-DRM)-based sensitivity analysis method,the sensitivity of the clamp support coordinates on the aeronautical hydraulic pipeline system is studied.The method greatly reduces the computational cost and provides powerful theoretical support for the variance-based sensitivity analysis of complex engineering problems.(3)Studied the application of the Possion's assumption-based first-passage method in the reliability sensitivity analysis of motion mechanism.Taking the motion mechanism as the research object,the local reliability sensitivity and variance-based global reliability sensitivity of the motion mechanism are studied based on the first-passage method of Poisson's assumption.Firstly,the first-order Taylor expansion of the non-linear performance function of the motion mechanism is expanded at the most probable point(MPP).Combined with the PHI2 method,the partial derivative of the motion mechanism reliability to the distribution parameters of random variables is derived,and the local reliability sensitivity analysis method of the motion mechanism is proposed.At the same time,the global reliability sensitivity analysis method of the motion mechanism is also derived based on the theory of analysis of variance.(4)Studied the approximate analytical solution of the local reliability sensitivity of the motion mechanism by the Poisson's assumption-based first-passage method,and derived the failure probability-based moment-independent reliability sensitivity index of the motion mechanism.Firstly,the MPP of the corresponding performance function at discrete time points is obtained by the sequential quadratic programming(SQP)optimization method.Combined with the approximate analytical solution of the crossing rate,the approximate analytical solution of the local reliability sensitivity index of the motion mechanism is directly obtained by solving the partial derivative of reliability with respect to the distribution parameters of the design variables.At the same time,the failure probability-based moment-independent global reliability sensitivity index of the motion mechanism is derived based on the concept of failure probability-based moment-independent importance measure.(5)Studied the hybrid reliability analysis of the motion mechanism under random uncertainty and interval uncertainty,and proposed the three basic types of hybrid uncertainty.The hybrid uncertainty of random variables and interval variables is the first category.The hybrid uncertainty of random variables with interval distribution parameters is the second category.The hybrid uncertainty of random variables with interval distribution parameters and interval variables is the third category.According to the Poisson's assumption-based first-passage method,combined with the two-loop iterative optimization method,the hybrid reliability analysis of motion mechanism under the above three basic hybrid uncertainties is carried out respectively,and proposed the unified hybrid reliability calculation method of the motion mechanism under three kinds of hybrid uncertainties.