|
| 基于降阶模型和机器学习的汽车防撞梁优化设计 |
| Optimization Design of Automobile Anti-collision Beam Based on Reduced-order Model and Machine Learning |
| 投稿时间：2026-06-02 |
| DOI：10.16018/j.cnki.cn32-1650/n.202602007 |
| 中文关键词: 机器学习 本征正交分解 结构优化 防撞梁 |
| 英文关键词: machine learning proper orthogonal decomposition structural optimization anti-collision beam |
| 基金项目:盐城工学院科研项目(YG20240624005) |
|
| 摘要点击次数: 3 |
| 全文下载次数: 2 |
| 中文摘要: |
| 以汽车防撞梁有限元仿真分析结果为依据,采用拉丁超立方采样方法选取样本点,将其结果作为数据集,并验证其满足降阶要求;应用MATLAB软件对数据集进行降维,再采用克里金法机器学习得到预测结果;根据预测结果采用序列最小二乘规划法对防撞梁进行优化设计。结果表明,采用预测模型时,各样本的决策系数R2均保持在0.96以上,RMSE的预测误差均小于0.7,都在工程允许范围之内;采用优化模型时,优化后总位移最大值降低28.99%,X方向最大值侵入量降低86.14%,最大应力降低19.42%。 |
| 英文摘要: |
| Based on the finite element simulation analysis results of the automobile anti-collision beam, the Latin hypercube sampling method is used to select the sample points, and the results are used as the data set, and it is verified that it meets the requirements of dimensionality reduction. Using MATLAB software, the data set is reduced in dimension, and Kriging machine learning is used to obtain the prediction results. Using the predicted results, the sequential least squares programming method is used to optimize the anti-collision beam. The results show that the decision coefficient R2 of each sample is above 0.96, and the prediction error of RMSE is less than 0.7, which is within the allowable range of the engineering. After optimization, the maximum total displacement is reduced by 28.99 %, the maximum X-direction intrusion is reduced by 86.14 %, and the maximum stress is reduced by 19.42 %. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|