Bootstrap重采样Bayesian方法在白车身制造尺寸不合格率估计中的应用

工业工程 ›› 2012, Vol. 15 ›› Issue (6): 15-19.

Bootstrap重采样Bayesian方法在白车身制造尺寸不合格率估计中的应用

1.北京信息科技大学机电学院，北京 100192；2.西安理工大学管理学院，陕西西安 710054

出版日期:2012-12-31 发布日期:2013-01-15
作者简介:王海燕（1979-）女，山东省人，讲师，硕士，主要研究方向为系统优化及制造业信息化.
基金资助:
北京市中青年骨干人才资助项目（71A1111143）

Application of Bayesian Method with Bootstrap Resampling to Failure Rate Estimation of White Auto Body

1. School of Electromechanical Engineering, Beijing Information Science and Technology University, Beijing 100192, China; 2. School of Management, Xi′an University of Technology, Shannxi 710054, China

Online:2012-12-31 Published:2013-01-15

摘要/Abstract

摘要： 白车身质量是汽车整车质量控制中的重要环节，针对白车身制造尺寸质量控制中检测数据属于小样本数据、数据处理分析不能采用一般大样本条件下统计分析方法的问题引入Bootstrap重采样Bayesian方法。通过对白车身尺寸质量的不合格率进行定义，分析简单计算、滑动计算、β分布Bayesian计算等3种估计不合格率的方法，引入Bootstrap重采样技术结合Bayesian方法进行不合格率的估计，并通过Matlab软件对4种算法进行仿真比较。仿真结果表明，Bootstrap重采样Bayesian方法的预测精度高于其余3种方法，适用于小样本情况下白车身制造尺寸不合格率的估计。最后通过一个实例演示了Bootstrap重采样Bayesian方法在白车身制造尺寸不合格率估计中的应用流程。

关键词: Bayesian方法, Bootstrap方法, 白车身, 质量控制

Abstract: The manufacturing accuracy of white auto body takes an important part in quality control of the whole auto manufacturing. In the white auto body manufacturing process, only small number of inspection samples can be obtained. Hence, quality control methods that require large number of samples are not applicable. To solve this problem, Bayesian method combined with Bootstrap resampling technique is adopted. After defining the failure rate of white body size, for comparison, simple calculation analysis, sliding calculation analysis, and Bayesian method based on β distribution are used to analyze the failure rate, respectively. Then, Bayesian method combined with Bootstrap resampling technique is applied to estimate the failure rate. Simulation based on Matlab is used to compare the four methods. Result shows that Bootstrap resampling Bayesian method outperforms the others and it can be applied to estimate the failure rate with small number of samples. An example is given to show how the application of the proposed method.

Key words: Bayesian method, Bootstrap method, white automotive vehicle body, quality control

王海燕1，侯琳娜2. Bootstrap重采样Bayesian方法在白车身制造尺寸不合格率估计中的应用[J]. 工业工程, 2012, 15(6): 15-19.

Wang Hai-yan1, Hou Lin-na2. Application of Bayesian Method with Bootstrap Resampling to Failure Rate Estimation of White Auto Body[J]. Industrial Engineering Journal , 2012, 15(6): 15-19.

参考文献

［1］金隼, 赵挺,来新民. 基于贝叶斯理论的白车身测量数据相关性分析［J］.机械设计与研究, 2003, 19(3):56-58. Jin Sun, Zhao Ting,Lai Xin-min. The relativity ahalysisof BIW checking data based on bayesian［J］. Machine Design and Research, 2003, 19(3):56-58. ［2］王华. 白车身制造尺寸质量控制中的软测量方法［J］. 计量学报, 2006, 27(4): 309-312. Wang Hua. Soft-sensing in body-in-white assembly quality control［J］. ACTA Metrologica Sinica, 2006,27(4): 309-312. ［3］林忠钦. 汽车车身制造质量控制技术［M］. 北京: 机械工业出版社, 2005.  ［4］Paez T L, Hunter N F. Fundamental concepts of the Bootstrap for statistical analysis of mechanical systems［J］. Experimental Techniques,1998, 22 (3): 35-38. ［5］连军, 姚福, 林忠钦,等. Bootstrap法在车身质量评定中的应用［J］. 中国机械工程, 2003, 14（4）:347-349 . Lian Jun, Yao Fu, Lin Zhong-qin, et al. Application of Bootstrap on the quality evaluation of auto-body［J］. China Mechanical Engineering,2003,14(4):347-349. ［6］谢益辉，朱钰. Bootstrap方法的历史发展和前沿研究［J］. 统计与信息论坛, 2008, 23(2) : 90-96. Xie Yi-hui, Zhu Yu. Bootstrap methods: developments and frontiers［J］. Statistics & Information Forum, 2008, 23(2): 90-96. ［7］Berger J.The case for objective Bayesian analysis［J］.Bayesian Analysis,2006,1(3):385-402. ［8］Efron B.Microarrys,empirical Bayes and the twogroups model［J］.Statistical Science,2008,23(1):1-22. ［9］熊道权. 汽车白车身质量信息管理系统［D］. 北京：清华大学, 2002. Xiong Dao-quan. The quality information system for the manufacturing process of body in white［D］.Beijing: Tsinghua University, 2002. ［10］Efron B. The Bootstrap and modern satistics ［J］. Journal of the American Statistical Association, 2000, 95 (452) : 1293-1296. ［11］Lee S, Bolic M. Confidence interval estimation for oscillometric blood pressure measurements using Bootstrap approaches ［J］. IEEE Transactions on Instrumentation and Measurement, 2011, 60(10): 3405-3415. ［12］Alexopoulos C, Kang K, Lilegdon W R. Bootstrap methods in computer simulation experiments［EB/OL］.(1995-12-03).http:∥ieeexp/ore.ieee.org/xpls/abs_all.jsp?arnumber:478720_tag=1. ［13］Lahiri S N. Validity of a block Bootstrap method for irregularly spaced spatial data under non-uniform stochastic designs［R］.Ames, IA: Iowa State University, 2003.［14］连军, 林忠钦, 来新民,等. 一种用于小样本合格率动态估计的Bayes方法［J］. 上海交通大学学报, 2002, 36(8): 10641067,1074. Lian Jun, Lin Zhong-qin, Lai Xin-min, et al. A Bayesian method for pass-rate estimation based on small sample［J］. Journal of Shanghai Jiaotong University, 2002, 36(8): 1064-1067,1074. ［15］孟庆华, 周晓军, 庞茂. Bootstrap 法在车辆故障检测中的研究及应用［J］. 汽车工程, 2005, 27(4): 498-501.

[1]	赵慧娟, 姜盼松, 黄文聪. 基于图像识别算法的图像传感器质量检测方法改善研究[J]. 工业工程, 2020, 23(3): 115-122.
[2]	范定祥, 李重莲. 网购供应链质量控制的策略研究[J]. 工业工程, 2020, 23(1): 104-111.
[3]	娄璐, 李艳婷. 面向未知自相关过程的Bootstrap控制图设计[J]. 工业工程, 2018, 21(4): 23-33.
[4]	赵会珍, 杨亚楠. 卫浴产品生产过程质量控制研究[J]. 工业工程, 2015, 18(1): 60-63.
[5]	赵会珍, 赵虎. 常闭阀罩生产过程质量控制研究[J]. 工业工程, 2014, 17(5): 29-34.
[6]	赵永强. 维修管理模式对制造企业质量控制影响的实证研究[J]. 工业工程, 2014, 17(4): 7-12.
[7]	樊树海1, Amanda Elizabeth2. 多元质量损失模型的系数确定[J]. 工业工程, 2012, 15(3): 80-82.
[8]	李尽法. 不完美检验和顾客抱怨的供应链质量控制模型[J]. 工业工程, 2011, 14(4): 68-71.