Data-informed Emulators for Multi-Physics Simulations

Abstract

Machine learning techniques are powerful tools for construction of emulators for complex systems. We explore different machine learning methods and conceptual methodologies, ranging from functional approximations to dynamical approximations, to build such emulators for coupled thermal, hydrological, mechanical and chemical processes that occur near an engineered barrier system in the nuclear waste repository. Two nonlinear approximators, random forests and neural networks, are deployed to capture the complexity of the physics-based model and to identify its most significant hydrological and geochemical parameters. Our emulators capture the temporal evolution of the Uranium distribution coefficient of the clay buffer, and identify its functional dependence on these key parameters. The emulators' accuracy is further enhanced by assimilating relevant simulated predictors and clustering strategy. The relative performance of random forests and neural networks shows the advantage of ensemble learning in random forests algorithm, especially for highly nonlinear problems with limited data.

References in corpus (8)

• Scikit-learn: Machine Learning in Python
• Popular Ensemble Methods: An Empirical Study
• Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
• Deep UQ: Learning deep neural network surrogate models for high dimensional uncertainty quantification
• Iterative Random Forests to detect predictive and stable high-order interactions
• Data Driven Governing Equations Approximation Using Deep Neural Networks
• Averaging Weights Leads to Wider Optima and Better Generalization
• Data-driven learning of non-autonomous systems