Scifaro

中文

用于反应流高效预测的全阶、降阶与机器学习模型流水线

数值分析 2022-05-02 v3 数值分析

作者: Pavel Gavrilenko , Bernard Haasdonk , Oleg Iliev , Mario Ohlberger , Felix Schindler , Pavel Toktaliev , Tizian Wenzel , Maha Youssef

在 arXiv 上查看 ↗ PDF ↗ DOI ↗

摘要

我们提出一种集成方法，利用全阶离散的模拟数据以及基于投影的 Reduced Basis（约化基）降阶模型来训练机器学习方法，特别是核方法（Kernel Methods），从而为具有不同输运机制的反应流中的化学转化率实现快速、可靠的预测模型。

关键词

反应流全阶模型降阶模型约化基机器学习核方法化学转化率

引用

@article{arxiv.2104.02800,
  title  = {A full order, reduced order and machine learning model pipeline for efficient prediction of reactive flows},
  author = {Pavel Gavrilenko and Bernard Haasdonk and Oleg Iliev and Mario Ohlberger and Felix Schindler and Pavel Toktaliev and Tizian Wenzel and Maha Youssef},
  journal= {arXiv preprint arXiv:2104.02800},
  year   = {2022}
}

备注

9 pages, 1 table; Previously this version appeared as arXiv:2110.12388v2 which was submitted as a new work by accident