主办：陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊：1976年
汽车实用技术

汽车实用技术 ›› 2026, Vol. 51 ›› Issue (12): 53-57.DOI: 10.16638/j.cnki.1671-7988.2026.012.010

• 设计研究 • 上一篇    

基于岭回归算法的吸能盒最大压溃力训推 数字化方法研究

郝维，何金龙*   

  1. 北京国家新能源汽车技术创新中心有限公司
  • 发布日期:2026-06-23
  • 通讯作者: 何金龙
  • 作者简介:郝维（1986－），男，硕士，研究方向为汽车仿真与仿真数字化技术 通信作者：何金龙（1994－），男，硕士，工程师，研究方向为汽车仿真与仿真数字化技术

Research on a Digital Training and Inference Method for Maximum Crushing Force of Energy Absorption Box Based on Ridge Regression Algorithm

HAO Wei, HE Jinlong*   

  1. Beijing National New Energy Vehicle Technology Innovation Center Company Limited
  • Published:2026-06-23
  • Contact: HE Jinlong

PDF

摘要： 为解决车企产品研发仿真验证阶段中工作流程重复、仿真模型计算时间长等痛点，文 章以汽车吸能盒为研究对象，通过 Python、TCL、VB 等脚本语言，实现参数化建模、仿真流 程和吸能盒样本数据自动化。基于小样本数据，采用岭回归算法并进行优化，最终实现代理 模型平均绝对百分比误差（MAPE）为 3.23%、决定系数 R 2 为 0.955 3，可代替传统仿真完成 压溃力快速预测。文章所述方法提升了产品设计迭代效率，为汽车行业仿真与人工智能（AI） 融合应用提供了一种可参考的数字化训推方法。

关键词: 代理模型；仿真平台；模型训练；模型推理；岭回归算法

Abstract: To address the pain points of repetitive workflows and long calculation times of simulation models during the simulation verification phase of automotive product development, the automotive energy absorption box is selected as the research object in this study. By means of script languages such as Python, TCL, and VB, the automation of parametric modeling, simulation processes, and sample data of the energy absorption box is realized. Based on small-sample data, the ridge regression algorithm is adopted and optimized. Eventually, a mean absolute percentage error (MAPE) of 3.23% and a coefficient of determination R 2 of 0.955 3 were achieved for the surrogate model, which could replace traditional simulations to accomplish rapid prediction of crushing force. The method proposed in this paper improves the efficiency of product design iteration and provided a referable digital training and inference method for the integration application of simulation and artificial intelligence (AI) in the automotive industry.

Key words: surrogate model; simulation platform; model training; model inference; ridge regression algorithm