Scholarlink

<p>Sensitivity analysis has great value in optimization of multibody system dynamics. In this paper, the automatic differentiation method (ADM) is applied to the reduced multibody system transfer matrix method (RMSTMM) for sensitivity analysis. The proposed method combines the advantages of the ADM and the RMSTMM perfectly, including fast computational speed, no need for global dynamics and sensitivity equations, and deriving expressions of derivatives automatically. Compared to the ADM based on the Lagrange method, the computational efficiency of the proposed method is higher for sensitivity analysis of large complex multibody systems. Application to three different multibody systems, including planar pendulum systems, a spatial three pendulum system, and a spatial tree system verifies the reliability of the corresponding code implementation of the proposed method.</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

Automatic differentiation for sensitivity analysis based on reduced multibody system transfer matrix method

  • 期刊论文

  • 发表日期：

    2025年12月3日

查看全文
  • Youyu Wang,
  • Guoping Wang,
  • Yutian Sun,
  • Jinxin Tang,
  • Genyang Wu,
  • Lilin Gu

摘要

Sensitivity analysis has great value in optimization of multibody system dynamics. In this paper, the automatic differentiation method (ADM) is applied to the reduced multibody system transfer matrix method (RMSTMM) for sensitivity analysis. The proposed method combines the advantages of the ADM and the RMSTMM perfectly, including fast computational speed, no need for global dynamics and sensitivity equations, and deriving expressions of derivatives automatically. Compared to the ADM based on the Lagrange method, the computational efficiency of the proposed method is higher for sensitivity analysis of large complex multibody systems. Application to three different multibody systems, including planar pendulum systems, a spatial three pendulum system, and a spatial tree system verifies the reliability of the corresponding code implementation of the proposed method.