Classification of Machine Learning Frameworks for Data-Driven Thermal Fluid Models

Abstract

Thermal fluid processes are inherently multi-physics and multi-scale, involving mass-momentum-energy transport phenomena. Thermal fluid simulation (TFS) is based on solving conservative equations, for which - except for "first-principle" direct numerical simulation - closure relations (CRs) are required to provide microscopic interactions. In practice, TFS is realized through reduced-order modeling, and its CRs can be informed by observations and data from relevant and adequately evaluated experiments and high-fidelity simulations. This paper focuses on data-driven TFS models, specifically on the development using machine learning (ML). Five ML frameworks, dubbed Type I to V, are introduced. The frameworks vary in their performance for different applications depending on the level of knowledge of governing physics, the source, type, amount and quality of available data for training. Notably, outlined for the first time in this paper, Type III models present stringent requirements on modeling, substantial computing resources for training, and high potential in extracting value from "big data" in thermal fluid research. The paper demonstrates the ML frameworks by three examples. First, we utilize the heat diffusion equation with nonlinear conductivity formulated by convolutional neural networks to illustrate the applications of Type I, II, and III ML. The results indicate a preference for Type II ML under deficient data support. Type III ML can effectively utilize field data, potentially generating more robust predictions than Type I and II ML. Second, we illustrate how to employ Type I ML and Type II ML frameworks for data-driven turbulence modeling using reference works. Third, we demonstrate Type I ML by building a deep learning (DL)-based slip closure for two-phase flow modeling. The results showe that the DL-based closure exhibits a bounded error in prediction domain.

References in corpus (3)

• Sensitivity and Generalization in Neural Networks: an Empirical Study
• Deep learning and the Schrödinger equation
• A Data-driven Approach for Turbulence Modeling

Cited by in corpus (16)

• Physics-guided deep learning framework for predictive modeling of the Reynolds stress anisotropy
• A Data-driven Framework for Error Estimation and Mesh-Model Optimization in System-level Thermal-Hydraulic Simulation
• Neural network representability of fully ionized plasma fluid model closures
• Multimodal Imaging-based Material Mass Density Estimation for Proton Therapy Using Physics-Constrained Deep Learning
• Physics-integrated machine learning: embedding a neural network in the Navier-Stokes equations. Part II
• Reynolds-Averaged Turbulence Modeling Using Type I and Type II Machine Learning Frameworks with Deep Learning
• Physics-integrated machine learning: embedding a neural network in the Navier-Stokes equations. Part I
• MRI-based Material Mass Density and Relative Stopping Power Estimation via Deep Learning for Proton Therapy
• Review of Physics-based and Data-driven Multiscale Simulation Methods for Computational Fluid Dynamics and Nuclear Thermal Hydraulics
• Early in vivo Radiation Damage Quantification for Pediatric Craniospinal Irradiation Using Longitudinal MRI for Intensity Modulated Proton Therapy
• Data-driven Hi2Lo for Coarse-grid System Thermal Hydraulic Modeling
• Uncertainty Quantification and Software Risk Analysis for Digital Twins in the Nearly Autonomous Management and Control Systems: A Review
• Quantification of Model Uncertainty in RANS Simulations: A Review
• Deep learning-based Fast Volumetric Image Generation for Image-guided Proton FLASH Radiotherapy
• Development of a Data-driven Turbulence Model for 3d Thermal Stratification Simulation during Reactor Transients
• Extending deep learning U-Net architecture for predicting unsteady fluid flows in textured microchannels