Non-exponential Distributions in Reliability Modeling of PMS: Approximation and Simulation Approaches
The phased mission systems (PMSs) are systems subject to consecutive, multiple, nonoverlapping mission time durations (phases) and they need to complete different missions. Such a system is usually seen in aerospace, nuclear power, and many other applications . One classic example of the PMSs is the manned spacecraft which involves takeoff, orbital transfer, orbit operation, and back to earth phases. In each phase, the system needs to accomplish different missions, which means that the system structure and failure criteria differ from phase to phase [2–4]. Furthermore, the system may be subject to different stresses in different phases due to different environments, for example, the endo-atmosphere in the takeoff phase and the outer space in the orbital operation phase of the spacecraft. Therefore, a distinct model for each phase is necessary for the modeling and assessment of an accurate system. But with the distinct models, the dependences across the phases pose a great challenge to the reliability modeling and assessment of the whole system; for example, in a non-repairable PMS a component that fails in the former phases will also stay in the failure state in the latter phases, making system modeling and assessment more difficult. Furthermore, many practical systems are subject to dynamic behaviors, such as cold standby or functional standby . Therefore, the dependences across the phases and dynamic behaviors pose great challenges to the existing system modeling and assessment methods.