Reliability Analysis Of Multi-state System Considering Multiple Failure Modes

Along with the advance of the modern engineering technology,the inaccuracy of the practical engineering systems analyzed by traditional reliability method was prominent.The main reason is the increasing complexity of the system structure led to by system structure,operating state,performance characterization,operating environment,etc.The traditional reliability theory had many shortcomings in solving practical engineering problems,which was mainly reflected in that it could neither describe the characteristics of multi-states of complex system in detail,nor describe the relationship between system performance and component performance truly.What's more,it could not express the correlation between failure states of component clearly.The traditional reliability theory was generally established based on the statistical independency of the invalid problems among the components of the system.However,in practical engineering,there is always coupling existing in the failure of components,which includes common cause failure,subordinate failure,product degradation failure and competitive failure among impact failure.At the same time,the systems and components that the traditional reliability theory aimed at were usually assumed to be two-states components.However,the components and system in practical engineering often have multiple performances.Based on the above problems,the following aspects were studied.(1)The common cause failure problem relevant to unit failure statistics which generally exist in two-state system is studied.The reliability modeling method based on Bayesian network is used to model the two-state system with common cause failure and probabilistic common cause failure respectively,and two examples are considered to verify the availability of the proposed method.(2)The considered system is extended from two-state system to multi-state system.The reliability model of the multi-state system is characterized by continuous time Markov chain based on Markov stochastic process and general generating function.At the mean time the influence of common cause failure in multi-state system shall be considered,a modified general generating function method will be proposed to analyze the reliability of multi-state system under the influence of common cause failure.and to give an example for verification.(3)The reliability analysis on multi-state system with correlation competition failure process is conducted.When the system is affected by both degradation and random impact,the failure process can be divided into degradation failure and impact failure according to different failure mechanisms.When components that construct the system have a mechanical property degradation,the reliability of product exceeds the critical threshold.While hard failures such as strong impact or short impact interval will lead to sudden failure of the system.In this paper,the degradation process is modeled taking advantage of a random variable which represents the degradation time of the internal capability of the product.It is assumed that the random impact leads to sudden change of one-step transition probability of the product,the influence degree of which is determined by the impact size and the vulnerability of the system to the exposed failure process.In addition,an example is given to verify the proposed method.