Reliability Evaluation Of Rubber For Vehicles By Accelerated Test Method And Intelligent Algorithm

Rubber is an environmentally friendly synthetic material with excellent performance,low cost,and relatively mature production processes.This material can also be recycled and reused.Therefore,rubber is widely used in vibration isolation,sealing,and insulation of mechanical systems.Fatigue and aging phenomena are inevitable in rubber parts due to factors,such as heat,liquid,moisture,ozone,salt spray,and radiation in the use environment.These phenomena result in mechanical performance degradation,which causes the difficulty of the material in meeting the requirements of use.A fast and accurate assessment of reliable durability of rubber materials,especially in terms of aging and fatigue,can enable prediction of the mechanical behavior of rubber parts at the design stage.Thus,a theoretical basis can be provided for the reliability optimization design of rubber parts,and the formulation of replacement maintenance cycles can be reasonable.This research aims to evaluate the reliability of automotive natural rubber.Accelerated test methods are introduced,and the application of artificial intelligence method in rubber reliability data analysis is also discussed.Accelerated test methods and macro and micro combined technologies are integrated in the study.A comprehensive research is conducted by the identification of acceleration factors,modeling of aging constitutive relations,evaluation of natural environmental aging,modeling considering dispersion,and application of artificial intelligence methods in the prediction of rubber fatigue life.(1)Research on accelerated test and identification of accelerated factors.On the basis of the principle of time–temperature equivalent translation method,sample points of the acceleration test are translated to reference stress to obtain the acceleration factors relative to the normal stress,and nonlinear fitting of the degradation trajectory is performed on all the sample points after the translation.A modified particle swarm optimization is introduced to identify the acceleration factor for minimizing the average relative dispersion coefficient of the fitting results.This method effectively solves the disadvantages,such as insufficient accuracy and low efficiency,of the traditional acceleration factor recognition method.On the basis of the proposed acceleration factor identification method combined with the pretest of constant thermal stress-accelerated aging data of rubber,the step-up and step-down stressaccelerated tests are designed to verify that the proposed test arrangement can meet the expected degradation trajectory requirements.Analysis of the measured step-up(-down)stress-accelerated test data confirms that the step-up(-down)stress-accelerated tests can improve the evaluation efficiency of rubber aging.(2)Research on rubber aging evaluation index and aging micromechanism.Aging tests of dumbbell-shaped rubber test pieces of different hardness at different temperatures for different times are performed in a free state to obtain samples of different aging degrees.Stress and strain data,elongation at break,and tensile strength are measured on an electronic tensile test bench.Tensile modulus and elongation at break are used to verify the consistency of the acceleration mechanism by Ahagon diagram.The results show that the elongation at break of the rubber sample obeys Arrhenius' law,and the regularity of tensile strength is poor.A Peck–Yeoh model is then proposed to describe the effects of temperature,hardness,and aging time on the constitutive relationship.After scanning electron microscope analysis of samples with different aging degrees,the micromechanism of rubber aging is explained by combining surface morphology changing and thermogravimetric analysis.(3)Reliability evaluation of aged rubber in natural environment.Pseudo-life method is used to obtain the pseudo-life distribution of samples under different temperature aging conditions in consideration of the difference in the decay trajectories of different testing pieces.Weibull distribution is then introduced to establish a reliability model of rubber life.The concept of temperature amplitude variation coefficient is proposed to solve the problem of reliability assessment in variable temperature,which greatly improves the reliability assessment efficiency of rubber aging under natural environment.(4)Modeling of rubber aging considering dispersion.The influence of hardness dispersion on the aging life of rubber is considered by establishing the degradation trajectory equations of different initial hardness compounds.The results show that the decay rate in the degradation trajectory equation is related to temperature and hardness in the accelerated model.The normal distribution model is used to fit the initial hardness on the basis of the statistical analysis of the initial hardness of rubber.In addition,the Monte Carlo method is introduced to simulate the decay trajectory of rubber samples with an initial hardness that obeys the normal distribution at room temperature.The kernel density modeling method is used to obtain the probability distribution curve of pseudo failure life.Three typical random process models,namely,the Wiener process,the Gamma process,and the inverse Gaussian process,are introduced in the research of rubber aging modeling considering trajectory dispersion.The parameters of the model are identified using the Bayesian method,with which the trajectory in room temperature considering the rubber dispersive is obtained,and the pseudo-lifetime probability distribution is also calculated.(5)Application of artificial intelligence method in rubber fatigue life prediction.A support vector machine model optimized by a modified gravitational search algorithm is proposed to predict the high-temperature fatigue life of rubber under a limited sample size.It is used to train the high-temperature fatigue data of rubber under the influence of multiple factors,and the results are compared with those of the BP neural network model to verify the accuracy of the proposed model.A random forest model is used to establish a rubber fatigue life model under constant amplitude stress in consideration of the influence of strain amplitude,mean strain,and strain ratio.The fatigue life under variable amplitude stress is thus predicted in combination with nonlinear fatigue damage theory,and the accuracy of the proposed method is verified by comparing the predicted values with the experimental results.In summary,this study aims to conduct aging and fatigue life prediction of natural rubber for automobiles.It focuses on reliability data processing,life prediction,dispersive influence analysis,and high-temperature and variable-amplitude fatigue life prediction by using accelerated testing combined with intelligent algorithms.The results of the study further improve the theoretical system of reliability evaluation for rubber materials.Thus,the engineering practice of accelerated test methods is enriched,and the application of intelligent algorithms in reliability prediction of automotive rubber parts is expanded.Accordingly,a solid foundation is provided for the design optimization and life extension of rubber parts.