A Bayesian-Based Graphical Modeling Tool for Probabilistic Reliability Analysis

Impact Technologies proposes to develop a Matlab/SimulinkT based Probabilistic Modeling and Analysis Toolbox (P-MAT) application that provides an intuitive and information-rich modeling environment for probabilistic reliability analysis with a powerful toolset consisting of Bayesian Update, Monte Carlo, 3rd party software integration and other modules. At the conclusion of Phase I, the USAF will be provided with prototype software that fully supports traditional probabilistic reliability analysis as well as uncertainty management via a multi-loop, iterative approach to Bayesian Updating in the presence of various forms of test data or other evidence types. While SimulinkT will act as the primary modeling and simulation environment for most analysis functions, file I/O, plotting, etc., P-MAT will also have the capability to translate some model elements into efficient routines for computationally intensive processes or for integration with 3rd party open source applications as well as perform high level project management in a front-end interface. The experienced Impact development team will work directly with the USAF TPOC to validate that P-MAT meets the USAF functional requirements and verify that it is robust and accurate by exercising it on a simple turbine blade design problem or a specific USAF case study.