Scholarlink

<p>This study investigates the dynamic and structural behavior of a precision milling machine using finite-element analysis (FEA) and experimental modal analysis (EMA), and further predicts results using artificial neural networks (ANNs). Modal analysis is performed to identify natural frequencies and mode shapes, thereby highlighting potential resonance issues. The natural frequencies predicted by FEA are 79.6, 86.7, 114, 128.9, and 166.6&#xa0;Hz, while experimental modal testing yields corresponding values of 83.8, 90.7, 104, 120, and 166&#xa0;Hz, with percentage errors ranging from −5 to 9.6%. Static structural analysis is conducted to assess mechanical stability under operational loading conditions, revealing a maximum total deformation of 0.12&#xa0;mm and directional deformations of 0.02&#xa0;mm along the X-axis, 0.11&#xa0;mm along the Y-axis, and 0.05&#xa0;mm along the Z-axis. Harmonic analysis is employed to evaluate the vibrational response of the structure under periodic excitation forces. Experimental validation is performed using impact hammer testing, and the results are further assessed using the modal assurance criterion (MAC) and the coordinate modal assurance criterion (Co-MAC) to ensure accuracy. ANN models trained using both FEA and experimental data demonstrate high predictive capability, achieving correlation coefficients (<i>R</i>-values) of 0.944 for training, 0.925 for testing, and 0.947 for validation. The ANN effectively captures complex relationships and hidden patterns in the data that are not directly observable through conventional FEA. Overall, the findings enhance predictive accuracy and computational efficiency, thereby improving machine stability, reducing vibration levels, and optimizing precision machining performance.</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

Multiphysics simulation of a precision manufacturing machine with experimental validation and prediction

  • 期刊论文

  • 发表日期：

    2026年4月4日

查看全文
  • Ratnakar Behera,
  • Tzu-Chi Chan

摘要

This study investigates the dynamic and structural behavior of a precision milling machine using finite-element analysis (FEA) and experimental modal analysis (EMA), and further predicts results using artificial neural networks (ANNs). Modal analysis is performed to identify natural frequencies and mode shapes, thereby highlighting potential resonance issues. The natural frequencies predicted by FEA are 79.6, 86.7, 114, 128.9, and 166.6 Hz, while experimental modal testing yields corresponding values of 83.8, 90.7, 104, 120, and 166 Hz, with percentage errors ranging from −5 to 9.6%. Static structural analysis is conducted to assess mechanical stability under operational loading conditions, revealing a maximum total deformation of 0.12 mm and directional deformations of 0.02 mm along the X-axis, 0.11 mm along the Y-axis, and 0.05 mm along the Z-axis. Harmonic analysis is employed to evaluate the vibrational response of the structure under periodic excitation forces. Experimental validation is performed using impact hammer testing, and the results are further assessed using the modal assurance criterion (MAC) and the coordinate modal assurance criterion (Co-MAC) to ensure accuracy. ANN models trained using both FEA and experimental data demonstrate high predictive capability, achieving correlation coefficients (R-values) of 0.944 for training, 0.925 for testing, and 0.947 for validation. The ANN effectively captures complex relationships and hidden patterns in the data that are not directly observable through conventional FEA. Overall, the findings enhance predictive accuracy and computational efficiency, thereby improving machine stability, reducing vibration levels, and optimizing precision machining performance.