Adjoint Sensitivity Analysis of Dynamic Reliability Models Based on Markov Chains-Ⅱ: Application to IFMIF Reliability Assessment

摘要

In Part Ⅱ of this work, the adjoint sensitivity analysis procedure developed in Part Ⅰ is applied to perform sensitivity analysis of several dynamic reliability models of systems of increasing complexity, culminating with the consideration of the International Fusion Materials Irradiation Facility (IFMIF) accelerator system. Section Ⅱ presents the main steps of a procedure for the automated generation of Markov chains for reliability analysis, including the abstraction of the physical system, construction of the Markov chain, and the generation and solution of the ensuing set of differential equations; all of these steps have been implemented in a stand-alone computer code system called QUEFT/MARKOMAG-S/ MCADJSEN. This code system has been applied to sensitivity analysis of dynamic reliability measures for a paradigm "2-out-of-3" system comprising five components and also to a comprehensive dynamic reliability analysis of the IFMIF accelerator system facilities for the average availability and, respectively, the system's availability at the final mission time. The QUEFT/MARKOMAG-S/MCADJSEN has been used to efficiently compute sensitivities to 186 failure and repair rates characterizing components and subsystems of the first-level fault tree of the IFMIF accelerator system.