Proof-of-Concept Study for Uncertainty Quantification and Sensitivity Analysis using the BRL Shaped-Charge Example

Hughes, Justin Matthew¹  Mississippi State Univ., Mississippi State, MS (United States)] 

[1] Los Alamos National Lab. (LANL), Los Alamos, NM (United States)

关键词: Uncertainty;    Sensitivity;    Surrogate Modeling;   

DOI  :  10.2172/1291182
RP-ID  :  LA-UR-16-25183
PID  :  OSTI ID: 1291182

美国|英语

来源: SciTech Connect

【 摘 要 】

These are the slides for a graduate presentation at Mississippi State University. It covers the following: the BRL Shaped-Charge Geometry in PAGOSA, mesh refinement study, surrogate modeling using a radial basis function network (RBFN), ruling out parameters using sensitivity analysis (equation of state study), uncertainty quantification (UQ) methodology, and sensitivity analysis (SA) methodology. In summary, a mesh convergence study was used to ensure that solutions were numerically stable by comparing PDV data between simulations. A Design of Experiments (DOE) method was used to reduce the simulation space to study the effects of the Jones-Wilkins-Lee (JWL) Parameters for the Composition B main charge. Uncertainty was quantified by computing the 95% data range about the median of simulation output using a brute force Monte Carlo (MC) random sampling method. Parameter sensitivities were quantified using the Fourier Amplitude Sensitivity Test (FAST) spectral analysis method where it was determined that detonation velocity, initial density, C1, and B1 controlled jet tip velocity.