The method of dynamic systems reliability modeling based on hybrid theory and interval analysis

Abstract

In engineering, the process of a dynamic system is formulated by a set of continuous variables and a set of discrete states and failure can be seen as the unexpected state in the process. Moreover, the discrete events and the continuous variables influence each other while the dynamic system is processing. On the other hand, the distributions of some system's parameters are hardly to obtain and it is also time-consumed to simulate the model even if the distribution is obtained due to the traditional sampling algorithms. These problems bring great difficulties in reliability modeling and analysis. In this paper, a method of non-probabilistic reliability modeling and analysis based on hybrid theory and interval analysis is proposed to solve the problems described above. Hybrid theory, which is proposed to solve the interaction between continuous and discrete and mainly applied in the field of control and computer, is utilized to build the reliability and performance integrated model. Interval analysis is utilized to describe the uncertainties. Combining with the advantages of hybrid theory and interval analysis, the method can be used to verify the reliability of system which is influenced by its continuous variable and discrete state. When the bounds on the uncertain information can be obtained readily, the method is particularly useful.