基于最小样本空间的Johnson分布拟合方法

中国机械工程

基于最小样本空间的Johnson分布拟合方法

张琪1;吴义忠2;刘鑫2;乔平2

1.华中科技大学国家数控系统工程技术研究中心，武汉，430074
2.华中科技大学国家企业信息化应用支撑软件工程技术研究中心，武汉，430074

出版日期:2019-11-10 发布日期:2019-11-08
基金资助:
国家自然科学基金资助项目（61575205）

Johnson Distribution Fitting Method Based on Minimum Sample Space

ZHANG Qi1;WU Yizhong2;LIU Xin2;QIAO Ping2

1.National NC System Engineering Research Center of Huazhong University of Science & Technology,Wuhan,430074
2.Research Center for National CAD Support Software Engineering and Technology of Huazhong University of Science & Technology,Wuhan,430074

Online:2019-11-10 Published:2019-11-08

摘要/Abstract

摘要： 基于应力-强度干涉模型计算可靠度指标时，常使用统计方法获得应力和强度的概率分布，然而，在复杂机电产品的设计过程中，每一个样本值的获取都是一次昂贵估值，大量样本值的获取面临耗时严重等问题。针对该问题，采用参数化的Johnson分布拟合概率分布，研究基于最小样本空间的Johnson分布拟合方法。采用百分位数配比法求解Johnson分布族函数的参数，并基于假设检验原理确定满足拟合精度要求的最小样本空间，同时通过K-S拟合检验法对总体真实分布与拟合分布进行一致性分析，保证拟合精度符合要求。工程应用结果表明，该方法基于少量的昂贵估值样本可快速得到样本的总体分布，为可靠度指标的快速计算奠定了基础。

关键词: Johnson分布族函数, 假设检验, 最小样本空间, 一致性分析

Abstract: When calculating the reliability index based on the stress-strength interference model, statistical methods were used to obtain the probability distribution of stress and strength. However, in the design processes of complex electromechanical products, the acquisition of each sample values was an expensive evaluation, so, the acquisition of a large number of sample values would be confronted with serious problems such as time consuming. Aiming at this problem, the parameterized Johnson distribution family function was introduced to fit the probability distribution, and the Johnson distribution fitting method was studied based on the minimum sample space. The parameters of the Johnson distribution family function were solved by percentile matching method and the minimum sample space was determined based on the hypothesis testing principle. To ensure the fitting accuracy, K-S test was introduced to analyze the consistency between the real distribution and the fitted distribution. The results of engineering applications show that the method may quickly obtain the distribution of samples based on a small number of expensive evaluation samples, which lays a foundation for the rapid calculation of reliability index.

Key words: Johnson distribution family function, hypothesis testing, minimum sample space, conformance analysis

中图分类号:

TH122

张琪1;吴义忠2;刘鑫2;乔平2. 基于最小样本空间的Johnson分布拟合方法[J]. 中国机械工程.

ZHANG Qi1;WU Yizhong2;LIU Xin2;QIAO Ping2. Johnson Distribution Fitting Method Based on Minimum Sample Space[J]. China Mechanical Engineering.

参考文献

[1]陈小前,姚雯,欧阳琦. 飞行器不确定性多学科设计优化理论与应用[M].北京: 科学出版社，2013：1-10.
CHEN Xiaoqian, YAO Wen, OUYANG Qi. Theory and Application of Uncertainty-based Multidisciplinary Design Optimization for Flight Vehicles[M]. Beijing: Science Press, 2013:1-10.
[2]XIE M, TANG Y, GOH T N. A Modified Weibull Extension with Bathtub-shaped Failure Rate Function[J]. Reliability Engineering & System Safety, 2002, 76(3): 279-285.
[3]丛伟，陈晓阳，王志坚，等. Weibull分布产品小样本定时截尾试验方案下的可靠性评估[J].中国机械工程，2013,24(14):1891-1896
CONG Wei, CHEN Xiaoyang, WANG Zhijian, et al. Reliability Evaluation of Products with Life Following Weibull Distribution under Type I Censoring Test Plan with Small Sample Size[J]. China Mechanical Engineering, 2013, 24(14): 1891-1896.
[4]高顺川，冯静，孙权，等.基于威布尔分布的动态故障树定量分析方法[J]. 质量与可靠性，2005(5):32-35.
GAO Shunchuan, FENG Jing, SUN Quan, et al. Dynamic Fault Tree Quantitative Analysis Method Based on Weibull Distribution[J]. Quality and Reliability, 2005(5):32-35.
[5]郑锐. 三参数威布尔分布参数估计及在可靠性分析中的应用[J]. 振动与冲击, 2015, 34(5):78-81.
ZHENG Rui. Parameter Estimation of Three Parameter Weibull Distribution and Its Application in Reliability Analysis[J]. Journal of Vibration and Shock, 2015, 34(5):78-81.
[6]凌丹. 威布尔分布模型及其在机械可靠性中的应用研究[D]. 成都：电子科技大学，2011.
LING Dan. Research on Weibull Distribution and Its Applications in Mechanical Reliability Engineering[D]. Chengdu: University of Electronic Science and Technology of China, 2011.
[7]JOHNSON N L. Systems of Frequency Curves Generated, by Methods of Translation[J]. Biometrika, 1949, 36:149-76.
[8]ZAMAN K, RANGAVAJHALA S, MCDONALD M, et al. A Probabilistic Approach for Representation of Interval Uncertainty[J]. Reliability Engineering and System Safety, 2011, 96(1):117-130.
[9]DEBROTA D J, ROBERTS S D， SWAIN J J， et al. Input Modeling with the Johnson System of Distributions[C]∥Proceedings of the 1988 Winter Simulation Conference. San Diego, 1988:165-179.
[10]盛骤，谢式千，潘承毅. 概率论与数理统[M].4版. 北京: 高等教育出版社，2008.
SHENG Zhou, XIE Shiqian, PAN Chengyi. Probability Theory & Mathematical Statistics[M]. 4th ed. Beijing: Higher Education Press, 2008.
[11]SLIFKER J F, SHAPIRO S S. The Johnson System: Selection and Parameter Estimation[J]. Technometrics, 1980, 22: 239-246.
[12]陈开颜，张鹏，邓高明，等. 相关性电磁分析攻击及最小样本量分析[J]. 华中科技大学学报(自然科学版)，2011, 39(1):32-35.
CHEN Kaiyan, ZHANG Peng, DENG Gaoming, et al. Electromagnetic Correlation Attacks and Minimum Number and Attack Traces Analyzing[J]. Journal of Huazhong University of Science and Technology (Natural Science Edition), 2011, 39(1):32-35.
[13]倪延延，张晋昕. 假设检验时样本含量估计中容许误差δ的合理选取[J]. 循证医学，2011, 11(6):370-372.
NI Yanyan, ZHANG Jinxin. How to Determine Permissible Error δ Value Properly When Computing Sample Sizes in Hypothesis Tests[J]. The Journal of Evidence-based Medicine,2011, 11(6):370-372.
[14]张耀，曹小平. 材料力学[M]. 北京：清华大学出版社，2015.
ZHANG Yao, CAO Xiaoping. Mechanics of Materials[M]. Beijing: Tsinghua University Press, 2015.