Characterization and Identification Methods of Failure Mechanism Variation for Rubber Elements under Accelerated Storage Testing
Keyword:accelerated storage testing，characterization and identification methods，failure mechanism variation，likelihood ratio test，molecular dynamics simulation，Optimal design，physics of failure，rubber elements
Accelerated storage testing is an effective means of storage life prediction for equipments.The fundamental premise for deriving valid conclusions is that the failure mechanism under accelerated stress should be consistent.Characterizing and identifying the failure mechanism variation are of great importance to realize reasonable life decision and life extension.Due to complex variation process of failure mechanism,traditional accelerated storage testing identifies the variation based only on engineering experience.But in essence,it can hardly gurantee consistent mechanism and support the storage life prediction of highly reliable equipments.To meet the engineering requirement for the identification of mechanism variation,rubber elements are chosen as the object of study by comprehensive consideration of identification effect and maneuverability.Physics of failure and statistical analysis are applied to the model in this dissertation,and a systemeric research is made on the feature extraction method and construction method of decision rules.A set of methods of characterization and identification are developed for the mechanism variation during accelerated storage testing,which provides technical support for the test profile design and storage life prediction of rubber elements.The main contributuions are summarized as follows.1.Characterization and identification methods based on physics of failure are developed for the failure mechanism variation of rubbers.The thermal oxidative degradation mechanism and degradation process are analyzed in detail.The kinetic equations for describing the reaction process are presented.According to the theory of diffusion-limited oxidation,the study of feature extraction method for failure mechanism variation is emphasized.Then identification methods and accelerated degradation models are established based on physics of failure,which provides theoretical support for the quantitative analysis of failure mechanism variation before the accelerated storage testing.2.Molecular dynamics simulations of gas transport in rubbers are investigated.In allusion to the problem that gas duffsion coefficients at high temperatures are difficult to be measured,computer simulations are used as assistant tools and analog computation methods of duffsion coefficients are discussed.The calculation flow chart of diffusion coefficients is constructed.Furthermore,the effect of simulation parameter settings on calculation performance is analyzed and simulation methods of gas diffusion are systematically studied,which offers data for the physics-based reliability modelling of rubber elements.3.Characterization and identification methods based on statistical analysis are developed for the failure mechanism variation of rubbers,which provides theoretical support for the quantitative analysis of failure mechanism variation after the accelerated storage testing.(1)Degradation data are transformed into pseudo lifetime data for statistical analysis.For data following the lognormal distribution and Weibull distribution,the extraction methods of variation characteristics are investigated respectively based on the principle of likelihood ratio test.The relationship between failure mechanism variations and acceleration model variations is analyzed.And the statistical analysis methods of pseudo lifetime data are put forward for constant-stress and step-stress accelerated degradation testing.In addition,decision rules are established for identifications of mechanism variation.The proposed methods are validated by numerical examples.(2)Degradation data are directly applied for statistical analysis.For data modeling with the degradation distribution and degradation path,the extraction methods of variation characteristics are investigated respectively based on the principle of likelihood ratio test.The statistical analysis methods of degradation data are presented based on the degradation distribution and degradation path,which involves the linear and nonlinear acceleration models under constant-stress and step-stress accelerated degradation testing.Similarly,decision rules are established for identifications of mechanism variation and validated by numerical examples.4.Optimal design methods of test plan are presented for characterization on the failure mechanism variation of rubbers.The difference between two optimal designs is analyzed,which includes plan designs for consistent mechanism and inconsistent mechanism.Through minimizing the risk of misclassification,the test plan with mechanism variation is optimized.Based on the basic optimization theories of test plans for accelerated life testing and accelerated degradation testing,the objective function as well as plan variables and constraints are proposed.Optimal design methods of test plan are researched on the condition of constant accelerated stress level,which offers key technical support for the accuracy improvement of identification result.5.The proposed methods are applied to the characterization and identification of failure mechanism variation in the accelerated testing of nitrile rubber material and elements.According to the requirements of storage testing for nitrile rubbers,the basic test plans are designed on the condition of constant accelerated stress level.Before and after the testing,physics of failure and statistical analysis are applied to the identification of mechanism variation respectively.The comprehensive analysis on the mechanism consitency of nitrile rubbers is subsequently obtained.In summary,under the presearch foundation of China ministries and commissions,a systemeric research is made on the characterization and identification methods of failure mechanism variation of rubber elements,which takes advantage of both physics of failure and statistical analysis.The research findings have general guiding significance for the research breakthrough in the mechanism consistency of equipments,and have major theoretical meaning and engineering value for the investigation and application of accelerated testing technique in the storage life prediction of equipments.