Research On System Reliability Analysis Method Under Multiple Time-Varying Failure Modes

ABSTRACTFor complex products,multiple failure modes are inherent property.Most of the traditional reliability analysis methods are based on a large number of failure data.Because of the long life and high reliability of the complex systems,the traditional reliability analysis methods are obviously not practical.However,compared with traditional reliability analysis,time-varying reliability analysis is more complex and challenging due to its dynamic characteristics.At present,domestic and foreign scholars have developed a variety of time-varying reliability analysis methods from the perspective of analysis and numerical simulation.However,for modern complex systems,affected by a large number of random factors and failure modes during the service,timevarying reliability presents the coupling effect of multi-time-varying parameters and multiple failure modes,which makes the time-varying reliability analysis method to face great challenges,especially the situation that precision and efficiency cannot be satisfied at the same time.Time-varying system reliability estimation is a challenge issue for engineering systems with multiple time-varying failure modes.In this paper,a time-varying copulas method is proposed for the time-varying reliability analysis of systems with two timevarying failure modes.The specific research work of this paper are as follows.(1)The calculation model of time-varying reliability of the system with two failure modes is constructed.Using time-varying copula to calculate the joint failure probability among failure modes is proposed,and the challenges encountered in the calculation process are analyzed.(2)The time-varying reliability analysis method based on time-varying copula function considering two time-varying failure modes is introduced in detail.The key point of the method is to use the time-varying copula function to describe the time-varying correlation between failure modes.Firstly,the time series model is established based on time-varying limit state functions.Then six types of time-varying copula models are used to estimate the time-varying dependence between two time-varying limit state functions.Finally,based on the kernel density estimation method,the joint failure probability and time-varying reliability of the system are calculated by using the estimated time-varying parameters.(3)In this paper,the method is elaborated and verified by a numerical example in detail.The method is then applied to the calculation of time-varying reliability of the lower extremity exoskeleton.Further analysis of the problems and estimation results of the method in the actual application process is carried out.Six different time-varying copula models and two different marginal distribution estimation methods are then compared,and also the influence on the estimation results of the changes for time step.