Scholarlink

<p>This work presents an ongoing Sandia National Laboratories initiative aimed at establishing a laboratory-wide capability for projection-based reduced-order models (ROMs) and demonstrates its utility on two applications of national interest. We outline the motivations and needs of the lab and introduce Pressio, the open-source ROMs software ecosystem under active development constituting the foundation of this ROM capability. Packaged as C++ and Python libraries, Pressio mitigates the intrusiveness of ROM integration by providing a modular, extensible framework for developing, analyzing, and deploying ROM methodologies across diverse application domains. Pressio currently supports model reduction techniques for dynamical systems expressible as parameterized ordinary differential equations. Leveraging this expressive formulation, Pressio offers a minimal API that is natural for dynamical systems. After discussing the distinguishing characteristics of Pressio, we outline its key design features, describe how existing applications can use it, and present two large-scale test cases of interest to Sandia: (1) uncertainty quantification and sensitivity analysis of a steady turbulent flow over a hypersonic vehicle and (2) inference of material properties for a hypersonic vehicle thermal protection system. In both cases, Pressio enables ROMs that substantially accelerate the analyses of interest with a minimal loss in accuracy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Projection-based model reduction using Pressio: application to hypersonic aerothermodynamics

  • 期刊论文

  • 发表日期：

    2026年4月27日

查看全文
  • Francesco Rizzi,
  • Eric J. Parish,
  • Patrick J. Blonigan,
  • John T. Tencer,
  • Christopher R. Wentland,
  • Andrew J. Kurzawski,
  • R. Loek Van Heyningen,
  • Caleb Schilly,
  • Victor Brunini

摘要

This work presents an ongoing Sandia National Laboratories initiative aimed at establishing a laboratory-wide capability for projection-based reduced-order models (ROMs) and demonstrates its utility on two applications of national interest. We outline the motivations and needs of the lab and introduce Pressio, the open-source ROMs software ecosystem under active development constituting the foundation of this ROM capability. Packaged as C++ and Python libraries, Pressio mitigates the intrusiveness of ROM integration by providing a modular, extensible framework for developing, analyzing, and deploying ROM methodologies across diverse application domains. Pressio currently supports model reduction techniques for dynamical systems expressible as parameterized ordinary differential equations. Leveraging this expressive formulation, Pressio offers a minimal API that is natural for dynamical systems. After discussing the distinguishing characteristics of Pressio, we outline its key design features, describe how existing applications can use it, and present two large-scale test cases of interest to Sandia: (1) uncertainty quantification and sensitivity analysis of a steady turbulent flow over a hypersonic vehicle and (2) inference of material properties for a hypersonic vehicle thermal protection system. In both cases, Pressio enables ROMs that substantially accelerate the analyses of interest with a minimal loss in accuracy.