Traditionally, system engineering and system reliability estimation have been separate activities. Reliability methods have focused on predicting how a particular inventory of systems will age and for how long a given percent will remain functional. This effort has been predicated on the assumption that the system architecture is defined, and that a sufficient body of testing (whether derived from functional or component sources) exists. In contrast, system engineering has focused on the determination of an architecture that can be validated against customer requirements. Reliability enters into system engineering efforts generally as a requirements specification, however as the system does not exist during the design life-cycle phase, rigorous reliability analysis is not typically an active part of architecture trade-off studies. Similarly, although the central challenge of reliability prediction is to make the most effective use of data gathered from many different sources, information about the system grounded in physical limits gathered during the system engineering phase, is not generally incorporated into reliability calculations.
In this talk, I will discuss these issues, and describe how a variety of different approaches to system representation, concurrent engineering, and system architecture definition, could be included in reliability prediction efforts. Taking advantage of these relatively new system engineering and simulation approaches it is possible to inform reliability models with both design and operational information. At the same time, it is possible to imagine greater use of reliability tools and methods during the design process. The potential developmental opportunities and challenges for these joint reliability-system engineering methods will be described.
