Session: 12-03-03: Data-Enabled Predictive Modeling, Scientific Machine Learning, and Uncertainty Quantification in Computational Mechanics

Paper Number: 113604

113604 - Material Model Parameters Optimization in Liquid Mercury Target Dynamics Simulation With Machine Learning Surrogates 

A pulsed spallation target is subjected to very short (~0.7μs) but intense loads (23.3 kJ) from repeated proton pulses, which knock away neutrons from the mercury atoms’ nuclei for a wide range application in physics, engineering, medicine, petroleum exploration, biology, chemistry, etc. The effect of this pulsed loading on the stainless-steel target module which contains the flowing mercury target material is difficult to predict not only due to its short but intense explosive-like physical reaction, but also the nonlinear material behavior of the liquid mercury in the structure. Different simulation approaches, material models and experimental validations for the mercury under extensive proton loads have been tried for decades, expecting to predict the structural response correctly therefore to improve the target vessel design and extend its service lifetime. The best match of simulation results to experimental data was obtained by an equation of state (EOS) material model with a specified tensile cutoff pressure that simulates the cavitation threshold when the liquid mercury is under tensile state. The inclusion of a threshold to represent cavitation was a key parameter in achieving successful predictions of stress waves triggered by the high energy pulse striking on the mercury and vessel when the total injected proton power is lower, e.g., < 1.0 megawatts (MW) and no gas injection in mercury is needed to alleviate the pitting damage on inner vessel wall. Surrogate models have been developed, along with machine methods to optimize the key model parameters for the liquid mercury material without gas bubbles included when the proton pulses are in lower power stage. The increasing proton beam power level, such as up to 1.4 MW or even more in the proton power upgrade plan, will cause stronger pressure waves that lead to further vessel damage and shorter vessel lifetime due to the severe cavitation and pitting effect on vessel wall. Injecting small helium bubbles in the mercury has been an efficient method of mitigating the pressure wave at high power level stage. However, prediction of the resultant loading on the target is more difficult when helium gas is intentionally injected into the mercury. A 2-phase material model that incorporates the Rayleigh-Plesset (R-P) model is expected to address this complex multi-physics dynamics problem by including the bubble dynamics in the liquid mercury. Verification and validation results show that machine learning method and surrogate models can help optimize the uncertain parameters in the complex 2-phase material model. This approach is expected to fill the knowledge gap between unknown liquid-gas mixture material model and measured vessel strain responses.

Presenting Author: Lianshan Lin Oak Ridge National Laboratory

Presenting Author Biography: Dr. Lianshan Lin of Oak Ridge National Laboratory holds a Ph.D. degree in Mechanical Engineering from the Manchester University. Dr. Lin works in the areas of structural dynamics, user material subroutine development and structural materials in nuclear engineering.

Authors:

Lianshan Lin Oak Ridge National Laboratory
Hoang Tran Oak Ridge National Laboratory
Majdi Radaideh University of Michigan
Sarma Gorti Oak Ridge National Laboratory
Srdjan Simunovic Oak Ridge National Laboratory
Hao Jiang Oak Ridge National Laboratory
Drew Winder Oak Ridge National Laboratory
Sarah Cousineau Oak Ridge National Laboratory