• CN：11-2187/TH
• ISSN：0577-6686

• 仪器科学与技术 •

### 基于数据驱动建模的钣金装配过程误差分析

1. 徐州工程学院江苏省工程机械检测与控制重点实验室 徐州 221018
• 收稿日期:2018-07-15 修回日期:2019-01-25 出版日期:2019-05-20 发布日期:2019-05-20
• 通讯作者: 张磊(通信作者),男,1978年出生,博士,副教授。主要研究方向为精密制造,机电产品制造与装配过程中的质量控制。E-mail:triple-stone@163.com
• 基金资助:
江苏省重点研发计划（BE2016047）、江苏省高校自然科学基金（15KJB460016）和徐州市工业科技计划重点研发（KC16GZ015）资助项目

### Data-driven Modeling-based Variation Analysis for the Compliant Sheet Metal Assemblies

ZHANG Lei, HUANG Chuanhui, ZHU Enxu, WANG Lei, DONG Yan

1. Jiangsu Key Laboratory of Construction Machinery Detection and Control, Xuzhou University of Technology, Xuzhou 221018
• Received:2018-07-15 Revised:2019-01-25 Online:2019-05-20 Published:2019-05-20

Abstract: Variation analysis in a compliant sheet metal assembly process is of great significance for eliminating assembly quality faults. Existing analytical modeling methods are not competent because they are limited by material, geometry and assembly process of the compliant sheet metal parts. Different from the analytical modeling method, a data-driven modeling-based method for variation analysis in a compliant sheet metal assembly process is proposed based on the measured historical data of key product characteristics. Through engineering experience and mathematical deduction, the multivariate first-order autoregressive model and the multivariate partial linear model are established. The maximum likelihood estimation method and the least squares kernel smoothing estimation method are employed to calculate the parameters and nonparametric estimates of the established models. A typical four-step hood assembly example with four key product characteristics demonstrates the effectiveness of the proposed method in the process of variation analysis. The data-driven modeling method is easy to model and the analysis results are accurate and reliable, which provides a new path for variation analysis of sheet metal assembly process.