[论文解读] Multi-physics Preconditioning for Thermally Activated Batteries

Malachi Phillips|arXiv (Cornell University)|Feb 13, 2026

Advanced Battery Technologies Research被引用 0

一句话总结

该论文在 TABS 中开发了一个物理感知的分层块 Gauss-Seidel 预条件器，以求解热电池中强耦合的电化学系统，实现可扩展的性能，在 2048 个处理器上达到 51.3M DOF，并且组装和求解时间接近亚秒级。

ABSTRACT

Thermal batteries, also known as molten-salt batteries, are single-use reserve power systems activated by pyrotechnic heat generation, which transitions the solid electrolyte into a molten state. The simulation of these batteries relies on multiphysics modeling to evaluate performance and behavior under various conditions. This paper presents advancements in scalable preconditioning strategies for the Thermally Activated Battery Simulator (TABS) tool, enabling efficient solutions to the coupled electrochemical systems that dominate computational costs in thermal battery simulations. We propose a hierarchical block Gauss-Seidel preconditioner implemented through the Teko package in Trilinos, which effectively addresses the challenges posed by tightly coupled physics, including charge transport, porous flow, and species diffusion. The preconditioner leverages scalable subblock solvers, including smoothed aggregation algebraic multigrid (SA-AMG) methods and domain-decomposition techniques, to achieve robust convergence and parallel scalability. Strong and weak scaling studies demonstrate the solver's ability to handle problem sizes up to 51.3 million degrees of freedom on 2048 processors, achieving near sub-second setup and solve times for the end-to-end electrochemical solve. These advancements significantly improve the computational efficiency and turnaround time of thermal battery simulations, paving the way for higher-resolution models and enabling the transition from 2D axisymmetric to full 3D simulations.

研究动机与目标

• Motivate high-fidelity multiphysics modeling of thermally activated (molten-salt) batteries and the need for scalable linear solvers.
• Develop a scalable preconditioning strategy tailored to the coupled electrochemical system in TABS.
• Implement a hierarchical block Gauss-Seidel preconditioner to address strong inter-physics coupling.
• Demonstrate strong and weak scalability for large problem sizes and discuss implications for 3D simulations.

提出的方法

• Formulate the coupled electrochemical system as a 2x2 block structure with voltage and non-voltage blocks.
• Propose a hierarchical block Gauss-Seidel preconditioner (Equation 19) with inner solves via GMRES(30) and physics-specific subblock solvers.
• Use SA-AMG (smoothed aggregation AMG) with Chebyshev smoothing for liquid/solid voltage subblocks.
• Apply block Gauss-Seidel preconditioning for the non-voltage block, using DD(0)-ILU(0) for subblocks and SA-AMG for liquid-phase pressure.
• Employ flexible GMRES to handle inner solves and enable efficient preconditioner evaluations.
• Leverage the Teko package in Trilinos to construct the block preconditioner and SA-AMG via MueLu.

实验结果

研究问题

• RQ1 Can a physics-aware block preconditioner provide robust, scalable convergence for the monolithic electrochemical system in thermal batteries?
• RQ2 How do strong and weak scalability behave for the proposed block preconditioner as problem size grows and on high-performance hardware?
• RQ3 What is the impact of using domain-decomposition vs. physics-based block preconditioning on convergence and parallel efficiency?
• RQ4 Can the approach enable transition from 2D axisymmetric to full 3D simulations by improving solver feasibility?

主要发现

• The block-based, physics-aware preconditioner converges in very few iterations for the coupled voltage system and the non-voltage block when using GMRES with inner SA-AMG and ILU preconditioners.
• Domain-decomposition one-level RAS preconditioners fail to converge the monolithic electrochemical system at moderate scales, motivating the block-based approach.
• SA-AMG with Chebyshev smoothing provides robust multigrid performance for liquid/solid voltage subblocks, enabling scalable solves.
• Results show strong and weak scalability up to 51.3 million degrees of freedom on 2048 processors, with near sub-second setup and solve times for the end-to-end electrochemical solve.
• The study demonstrates that the physics-aware block preconditioner supports higher-resolution models and facilitates moving toward full 3D simulations.