对数正态分布下未来区间内故障次数的单样预测  

One-sample prediction for the failure number of lognormal distribution in the future time interval

作　　者：孙永全[1] 郭建英[1] 王凯[1] 陈洪科[2] 

机构地区：[1]哈尔滨理工大学传感器与可靠性工程研究所,哈尔滨150080 [2]闽南理工学院电子与电气工程系,石狮362700

出　　处：《系统工程理论与实践》2013年第3期785-790,共6页Systems Engineering-Theory & Practice

基　　金：国家高技术研究发展计划(863计划)(2009AA042407);黑龙江省教育厅项目(12511093)

摘　　要：基于产品寿命评估与预测理论,研究了对数正态分布下未来区间内故障次数的单样预测问题.引入胜算比,给出了对数正态分布下未来故障次数的预测子及预测区间,分析了预测精度的影响因素.结合风电场对风电机组关键部件合理储备的工程背景,利用工程实例验证了该方法的有效性.据此,对多台同时运行的设备合理规划关键部件的储备、优化维修保障策略具有明显的指导意义.One-sample prediction for the failure number of lognormal distribution in the future time interval was studied based on product; life estimation and prediction theories. Odd ratio was introduced, and then predictor and prediction interval of the failure number in the future under lognormal distribution were obtained, finally, prediction accuracy was discussed through analyzing the influence factors of the prediction results. According to the engineering background that the large wind farm demanded for reserves of the wind turbine key components, an example was presented to show the validity. Failure number prediction is useful to plan key components reserves for multiple devices running simultaneously and optimize maintenance support strategy rationally.

关 键 词：对数正态分布 故障次数预测 胜算比 风力发电机组 维修保障 

分 类 号：TB114.3[理学—概率论与数理统计]