Existing reliability evaluation methods rely on the availability of accuraterncomponent states data. They will become ineffective when the states themselves arernuncertain or unknown, which usually happens during the early stages of the development ofrnnew systems. In such cases it is important to understand how uncertainties will affect thernsystem reliability measures. Another drawback of current methods studying reliability ofrnMulti-State System (MSS) is that they only considered the systems whose componentsrnhave several discrete states. For those whose components have continuous states, thesernmethods are not effective either. This paper considered the continuous distribution ofrncomponents states during the approximation of Multi-State System (MSS) reliability andrnproposed a method to assess the reliability of this kind of system using Monte-Carlornsimulation. This method will also be useful when we have no enough data to know thernexact discrete states and related probability, and can only estimate components statesrndistribution types and related parameters. Two examples were employed to illustrate thernmethod. Comparison of the two examples shows that component state uncertainty hasrnsignificant influence on the assessment of system reliability. Our effort will make thernreliability approximation more realistic compared with existing methods.
展开▼
