Two Time-Dependent Reliability Analysis Methods Based On Surrogate Model

Research on the life and reliability of mechanical structures and products plays a key role for improving the performance stability and product competitiveness of major equipments,and has received more and more attention in recent years.However,from understanding,research,development of reliability to its application,these are mostly based on static mechanical models,which belong to the fields of static reliability.Due to the influence of material properties,environmental conditions,time-varying load and other various factors during the use of the engineering structure,the reliability of engineering structures usually decreases with time,which is no longer a single value under the traditional structural reliability model.Therefore,in order to ensure the performance of the structure throughout the service period,establishing time-dependent reliability models more conform to practical conditions,and it has important theoretical and practical.From the above points of view,we systematically studied the time-dependent reliability analysis method and completed the following works as follows:In order to improve the efficiency of time-varying reliability analysis method,an analysis method based on sparse grid numerical integration is proposed,we can use the method to solve some time-varying engineering problems efficiently.Fristly,the input random process of time-variant system is expanded into linear superposition of multiple independent random variables by expansion optimal linear estimation method.Then,the sparse grid numerical integration is adopted to obtain the first four moments of the minimum.In this process,the efficient global optimization method is performed to search the minimum of the time-varivant function of the integral nodes.Finally,the saddlepoint approximation is used to obtain the distribution of the minimum on the basis of the first four moments,and then the time-varying reliability analysis of the structure is completed.This method introduces the sparse grid numerical integration method and uses the extreme value method to solve the time-variant reliability of the structure,which greatly improves the efficiency and precision of time-variant reliability analysis method.In order to provides an effective analytical tool for assessing design reliability of the structure over the whole lifecycle,a time-variant reliability analysis method based on polynomial chaos expansion method is proposed.This method introduces the polynomial chaos expansion method and uses the time domain discretize method to solve the time-variant reliability of the structure,which can obtain the time-variant reliability of arbitrary time interval and arbitrary failure threshold.In the method,a time-dependent polynomial chaos expansion method(t-PCE)is proposed to simulate the time-dependent fuction,and it will be simulated as a function in terms of a standard normal variable and time.Then,Monte Carlo simulation is adopted to sample the standard normal variable to obtain the time-variant reliability of arbitrary failure threshold and time interval.