Degradation Modeling And Reliability Assessment Of Silicon-Containing Foam Sandwich Structure Based On Bayesian Method

Silicon-containing foam sandwich structure is an important structure of long-storage products.Revealing its performance degradation behaviors and accurately evaluating its storage reliability is of great significance to guarantee reliability and safety of longstorage products and mission success.However,functional material and silicon foam material,as the critical components of silicon-containing foam sandwich structure,have less failure data in the storage stage.Moreover,due to the restrictions of structure and assembly,it is impossible to monitor the performance of materials directly or via disassembling the silicon-containing foam sandwich structure.Therefore,it is difficult to know the performance evolution of functional material and silicon foam material directly in the storage stage.Additionally,the functional material and silicon foam are affected by common environmental stresses,leading to the dependency between their degradation processes.Therefore,how to accurately characterize the performance degradation processes of the two materials and effectively characterize the dependency between performance degradation processes of the two materials,is very crucial to accurately evaluating the reliability of silicon-containing foam sandwich structure.In this dissertation,we aim at characterizing the performance degradation behavior of functional material and silicon foam based on accelerated degradation data in the context of the lack of material failure and degradation data,and evaluating the reliability of silicon-containing foam sandwich structure with or without dependency between material degradation processes.The primary research contributions and innovative outcomes are summarized as follows:(1)Construction of the performance degradation models of the functional material and silicon foam based on their accelerated degradation data.The degradation processes of materials are characterized by accelerated degradation model,which is derived from Gamma process and acceleration model.The model parameters are estimated by Bayesian method based on the accelerated degradation data.The deviation information criterion is used to identify the best model among candidates.Finally,the performance degradation law of materials is well-characterized.The numerical example shows that the 95%confidence interval of the degradation path of the elastic modulus fits the accelerated degradation data well.(2)Development of a reliability assessment method based on the preloading degradation data with considering the resource constraints and the difference of degradation law between accelerated degradation test and storage condition.Firstly,based on the finite element simulation model of silicon-containing foam sandwich structure,the model that characterizes the relation between material performance(elastic modulus)and system performance(preload)is constructed by a regression method.Secondly,the likelihood function of the preloading degradation data is constructed by the Monte Carlo method.The Bayesian method is used to estimate the parameters of material degradation models.Lastly,the reliability of the silicon-containing foam sandwich structure is computed by the Monte Carlo method.The results from the numerical example show that the proposed method can accurately estimate the unknown parameters.(3)Development of a reliability assessment method for silicon-containing foam sandwich structure with dependency between degradation processes of functional material and silicon foam.Firstly,Gamma process is used to characterize the degradation of materials and the linear model is used to characterize the effect of preloading on degradation parameters of materials.Based on the relation between material performance and system performance,the degradation-rate-interaction model of functional material and silicon foam is constructed.Secondly,the performance degradation behavior of materials with dependency is characterized approximately by assuming the preload being constant in each unit time.The likelihood function of the preloading degradation data is constructed by the Monte Carlo method.The Bayesian method is used to estimate material degradation parameters.Lastly,the reliability of the silicon-containing foam sandwich structure is computed by the Monte Carlo method.The results from the numerical example show that the reliability of silicon-containing foam sandwich structure is underestimated without considering dependency between degradation processes of functional material and silicon foam.