Generating Correlated Corrosion Life Predictions from Affordable Accelerated Test Data

Impact Technologies LLC, in collaboration with the Ohio State Fontana Corrosion Center, proposes to develop and demonstrate an innovative multi-scale Bayesian-based approach for both correlating probabilistic corrosion model predictions with accelerated test data and for definitively improving the accuracy of models as experience grows. Since the Army is currently challenged with costly and burdensome full-scale accelerated corrosion testing, this program also promises to provide a flexible software architecture with which tuned models, in conjunction with component and/or subassembly tests, can serve to validate vehicle corrosion performance more affordably. In a comprehensive Phase I software demonstration, the development team will highlight automated model correlation processes, based on an accelerated vehicle subassembly corrosion test performed at the OSU Fontana Center. As an HLA-Compliant toolkit for corrosion modeling, we will further demonstrate how varied data sources such as service life histories, expert rulebases, accelerated tests and correlated lifing models can be integrated to generate reliability predictions as a function of expected platform usage, maintenance and environment.