Non-Probabilistic Reliability Analysis for Multidisciplinary Systems

In this paper, the reliability analysis problem of multidisciplinary systems is investigated by using non-probabilistic approach, that is, convex set of uncertainty. This method is a rational way to deal with uncertainties when available information is insufficient to warrant probabilistic analysis. The main aim of this paper is to develop computationally efficient techniques to calculate the non-probabilistic reliability index for multidisciplinary systems. Because of the interactions among various disciplines, the execution of the system level analysis is iterative by nature. Thus the all-in-one reliability analysis method requires a double-loop procedure and therefore is generally very time-consuming. To improve the computing efficiency, a new formulation is constructed according to the idea of SAND(Simultaneous Analysis and Design). In this formulation, the system level analysis is avoided and the interdisciplinary compatibility will be achieved progressively. As a result, only a single loop is required and all the analysis can be conducted concurrently at the individual discipline-level. Both methods are validated and compared in application to an electronic packaging problem. The calculated reliability index values are also compared with the safety frequencies obtained by Monte-Carlo simulation.