Research On Reliability Evaluation Of Automotive Rubber Materials Based On Accelerated Testing And Random Process Models

Rubber is a type of renewable materials with excellent mechanical performance.Compared with metal materials,it has the advantages of high elasticity,low modulus,low hardness,and high damping.With the development of production technologies and material science,rubbers have been widely used for the development of many high-reliability and long-life products,such as machinery,automobiles,aerospace and electronics.Due to its functions of insulation,sealing,connection,transmission and vibration isolation in mechanical and electronic equipment,rubber items play an important role in modern industry and engineering systems.However,temperature,ultraviolet rays,oil pollution,ozone and other types of environmental stress can inevitably result in performance degradation of rubber-based products,which could not meet the needs of normal working conditions.Under this circumstance,the establishment of reliability assessment models for rubber-based products can accurately and quickly evaluate degradation levels,thus providing scientific guidance for the development and replacement of rubber products.This article takes automotive rubber materials as research object.In order to shorten testing time and save money,accelerated testing is introduced to accelerate degradation processes of these rubbers.The relationship between macro and micro performance indicators of aging rubbers and degradation time,as well as accelerated stress is revealed to understand aging mechanisms of rubber-based products.Then,the Weibull distribution is used to fit pseudo life failure values of compression set rate of these aged rubbers.Considering that many products tend to degenerate with several performance indicators,stochastic processes and Copula functions are then employed to establish bivariate reliability models for aging automotive rubbers.More details are presented as follows:In the first chapter,recent research conditions of rubber products on vehicles,accelerated degradation methods,failure mechanism consistency testing methods,as well as degradation modeling methods,are presented.To be specific,macro and micro evaluation indicators that are commonly used in the research of rubber materials’ degradation are compared.Then,recent status of reliability research of vibration isolation rubbers,tire rubbers and sealing rubbers are discussed from different application scenarios.According to the way of applying accelerated stress,constantstress and step-stress accelerated degradation testing,as well as their optimization methods in reliability modeling,are discussed.Since the consistency of degradation mechanisms is the prerequisite for accelerated testing technology to maintain its effectiveness,advantages and disadvantages of commonly used degradation mechanism consistency testing methods and their respective applicable scenarios are presented.The development of machine learning algorithms,statistics and Copula functions have enriched reliability modeling methods,and therefore this article also focuses on comparing degradation modeling methods in reliability engineering.The second part introduces some basic concepts in reliability engineering,such as reliability,failure rate,median lifetime,and average lifetime,etc.and then elaborates degradation models and acceleration models that will be used for degradation modeling of automotive rubber in this article.Theoretical basis of Copula functions and parameter estimation methods have also been discussed.Based on the constant acceleration factor principle,parameters’ acceleration relationships of the Wiener,Gamma and Inverse Gaussian process are derived.These works can provide a “bridge”for the extrapolation of products’ reliability from accelerated conditions to use conditions.The third chapter mainly studies the relationships between macro and micro degradation indicators of aging rubbers and their aging mechanisms using different experiments.First of all,dumbbell-shaped rubber samples are designed according to national standards,and then placed in aging boxes at different temperatures for aging testing.The electronic tensile testing machine is adopted to carry out tensile testing on rubber samples with different aging degrees to obtain stress and strain data at fracture.Combining with scanning electron microscopy experiments,the relationships between rubbers’ mechanical properties and accelerated temperature stresses are further discussed.In the fourth chapter,the dynamic curve method is employed to fit the degradation data of compression rates of O-type sealing rubbers under 50℃,60℃,70℃ and 80℃.When the pseudofailure life data of the rubber samples are obtained by the above-mentioned curves and pre-defined degradation thresholds,the Weibull distribution is adopted to establish a statistical analysis model for reliability assessment of these rubbers.Meanwhile,the consistency of accelerated degradation failure mechanisms of these samples is verified based on the constant acceleration factor principle.Finally,the equivalent temperature method is used for evaluating of rubbers’ reliability in natural environment.In the last chapter,a bivariate degradation modeling method based on stochastic processes and Copula functions are developed for reliability assessment of vehicle sealing rubbers,in which two types of accelerated stress(temperature and ultraviolet light)and two performance indicators are considered.Specifically,different random models are used to establish marginal degradation processes of microscopic failure performance indicators of aging rubbers.Considering that each degradation process of rubber products is caused by two types of accelerated stresses,the Eyring relationship is selected as the accelerated model in marginal degradation modeling.Then,the MCMC algorithm is used to estimate unknown parameters in marginal degradation models with the selection of best marginal model through BIC criterion.Moreover,four different Copula functions are used to connect the best marginal models selected above,and then BIC criterion is reuse to choose the best Copula function.Finally,a multivariate degradation model considering two types of accelerated stress is developed for the reliability assessment of vehicle sealing rubbers.The correctness of the established model is verified by the degradation testing data of automotive sealing rubbers under natural environment designed by National Institute of Standard Technology.