详情

原文传递 Quantifying Uncertainties in Parameterizations of Strength Models of Rolled Homogeneous Armor: Part 1, Overview.

题名：	Quantifying Uncertainties in Parameterizations of Strength Models of Rolled Homogeneous Armor: Part 1, Overview.
作者：	Ramsey, J. J.
关键词：	Models, Strength (mechanics), Rolled homogeneous armor, Computer programs, Python programming language, Temperature, Stresses, Strain (mechanics), Probability density functions, Monte carlo method, Uncertainty quantification, Bayesian analysis, Johnson-cook model, Zerilli-armstrong model, Stan, Pymc3, Ipm (interval predictor model), Rha (rolled homogeneous armor), Flow stress, Plastic strain, Midas (material implementation database and analysis source), Ppd (posterior predictive distribution), Pfp (pushed forward posterior), Hdi (highest density interval), Bcc (body-centered cubic)
摘要：	Guidance is provided on how to obtain uncertainties in parameters of material strength models of rolled homogeneous armor,along with point estimates for those parameters, using existing software tools to implement two different approaches: Bayesianregression and the interval predictor model (IPM) approach. This report shows how to mathematically describe a Bayesianmodel associated with a material strength model and related experimental data, and how to express this Bayesian model in formsthat the aforementioned tools can process. It also describes how the IPM approach can be implemented in Python. The reportshows how the model parameter uncertainties can be visualized and how they may be presented in a form suitable for input tosoftware tools for uncertainty propagation analysis, such as Dakota. Finally, the report shows how Bayesian analysis may beused to evaluate the quality of the fit of a strength model to experimental data.
报告类型：	科技报告