航天电子元器件可靠性评价的Poisson回归分析  

Poisson regression analysis for reliability evaluation of aerospace electronic components

作　　者：高化猛 薛勇 郝星兆 王洋 陈信波 GAO Huameng;XUE Yong;HAO Xingzhao;WANG Yang;CHEN Xinbo(Space Engineering University,Beijing 101416,China;Taiyuan Satellite Launch Center,Taiyuan Shanxi 030000,China)

机构地区：[1]航天工程大学,北京101416 [2]太原卫星发射中心,山西太原030000

出　　处：《航天工程大学学报》2025年第2期40-44,共5页

摘　　要：航天高密度发射对航天电子元器件的可靠性要求越来越高,但很多航天电子元器件质量问题呈现偶发、散发、时间和次数规律性不强的特点,给可靠性评价带来困难。为实现航天电子元器件可靠性评价,获取可靠性评价数据,建立可靠性评价Poisson回归模型。在统计分析系统软件(Statistical Angalysis System,SAS)环境下编写Poisson回归分析程序,计算Lagrange乘子统计量和拟合优度,得到回归参数估计值、参数标准误估计值、置信区间、参数检验卡方值和P值。计算参数似然比统计量和置信区间,并对元器件的可靠性进行预测。结果表明,Poisson回归能够针对航天电子元器件质量问题特点规律,基于已有数据,实现航天电子元器件可靠性的评价与预测。研究成果对航天可靠性评价和质量管理具有理论意义和实践价值。The increasing demand for high-density aerospace launches imposes higher requirements on the reliability of aerospace electronic components.Many aerospace electronic components quality problems are characterized by episodic,scattered,and irregular in time and number,which bring difficulties in reliability evaluation.To realize reliability evaluation of aerospace electronic components,the data for reliability evaluation is gained,and the Poisson regression model for reliability evaluation is constructed.The Poisson regress program is written by Statistical Angalysis System(SAS).The Lagrange multiplier statistic and fit goodness are calculated.The regress parameter estimation,standard error estimates of parameters,confidence intervals,Chi-square value and P value for parameter tests are gained.The likelihood statistic and confidence interval are calculated,and the components reliability is predicted.The results present that Poisson regress can fit the character and law of aerospace electronic components reliability and can be used to realize aerospace electronic components reliability evaluation and prediction based on existing data.The research results have theoretical significance and practical value for aerospace reliability evaluation and quality management.

关 键 词：航天 电子元器件 可靠性 质量 评价 回归分析 

分 类 号：V557[航空宇航科学与技术—人机与环境工程]