Session Y02: Uncertainty Quantification and Error Analysis

Uncertainty Quantification for Equation of State using Heterogeneous Data Sources

Uncertainty quantification (UQ) for the Equations of State (EOSs) and various other properties of constituent materials (e.g., strength) is crucial to discovering the range of outcomes that are consistent with small-scale data within a multi-physics hydrodynamic simulation. Depending on the quantity and quality of data available, a mix of data sources (i.e., both calculated and experimental data) may be included in the calibration data. The type of data available, as well as the types of uncertainty present in that data, will influence the UQ workflow. This talk will focus on combining calculated data with experimental data for two examples: 1) reactant and product EOSs for a variety of high explosives, and 2) a multiphase Aluminum EOS. The use of statistical distances as a mechanism to compare the results of the UQ is considered.   

Progress Towards Gaussian Process Emulation for Improving Multiphase EoS Optimisation

Multiphase equation of state (EoS) uncertainty quantification (UQ) is limited by the dimensionality of the EoS input vector, the EoS model calculation time, and phase boundaries that are sensitive to input parameter changes. For multiphase materials the number of input parameters is high which makes the optimisation space large. Therefore, optimisation of EoS parameters takes many EoS calculations to complete. Gaussian Process Emulation (GPE) with Bayesian Optimisation (B-Opt) provides an efficient methodology for mapping the multiphase EoSs cost function. Confidence set contours from GPE with B-Opt applied to single phases of Tin are utilised to aid the optimisation procedure for a multiphase EoS model. The confidence set contours alter the optimisation procedure by: providing non-random starting locations, and creating a non-euclidean optimisation space. This reduces the total number of multiphase EoS calculations required to find the optimum answer.

UK Ministry of Defence © Crown owned copyright 2023/AWE   

Quantifying Motion Blur by Imaging Shock Front Propagation with Broadband and Narrowband X-ray Sources

Time-integrated radiography using MeV Bremsstrahlung X-ray sources is the norm for imaging during system-level testing of components and structures under dynamic condition. One source of error in the analysis of the dynamic radiography data sets stems from motion blur, which smears out sharp interfaces at a greater degree with longer exposure times, which become necessary to provide sufficient X-ray penetration of objects of interest. To investigate and quantify motion blur, a 1D shock wave through PMMA was investigated experimentally at The Dynamic Compression Sector at The Advanced Photon Source with tapered broadband and 25.46 ± 1.056 keV narrowband X-rays to test our state-of-the-art hydrodynamic and radiography simulation tools and compare the effect of the X-ray source type and exposure time on the radiography image. Four cameras with different exposure times were used for each experiment to compare the effect that exposure time has on motion blur, signal-to-noise ratio, and contrast level. There is a high level of agreement between the experimental and simulation results across the range of data sets investigated in this study.