Model-based reliability analysis

Modeling, in conjunction with testing, is a rich source of insight. Model parameters are easily controlled and monitoring can be done unobtrusively. The ability to inject faults without otherwise affecting performance is particularly critical, Many iterations can be done quickly with a model while varying parameters and conditions based on a small number of validation tests. The objective of Model-Based Reliability Analysis (MBRA) is to identify ways to capitalize on the insights gained from modeling to make both qualitative and quantitative statements about product reliability. MBRA will be developed and exercised in the realm of weapon system development and maintenance, where the challenges of severe environmental requirements, limited production quantities, and use of one-shot devices can make testing prohibitively expensive. However, the general principles will also be applicable to other product types. There are many anticipated benefits from MBRA, especially in the context of weapon systems: Development time and required test assets will be reduced. In addition, there will be fewer design iterations necessary. Fewer modifications in the production processes will be needed, leading to a more homogeneous product. Furthermore, MBRA can be used to evaluate the impact of production and part changes if they do become necessary. Typically it has been challenging in the past to determine the generalized impact of an observed anomaly (i.e., how the anomalous behavior may manifest itself under different - but still valid - conditions). Often specific conditions cannot be varied in a controlled fashion in a testing situation. Use of a modeling framework permits one to inject hypothesized behaviors under different conditions and observe the consequences. This enables the effective leveraging of (often limited) test results that provide the initial foundation for model development.