摘要
提出了基于极值理论的系统最大值指标评定与测算方法,为某些特殊领域(如航空、航天、国防)中的长周期系统的最大值指标评定提供了理论依据。首先,遵循一定的原则对系统输出的误差序列进行区组划分,取得近乎独立的最大值观测序列;其次,对最大值观测序列应用极值理论建模,取得一定重现期下的重现水平和不确定度;然后考查重现水平与最大值指标间的关系,并应用经典统计学理论完成对系统最大值指标的评定与测算;接着,给出了具体的应用实例,实验结果表明,该方法是可行的。最后,从不同角度对时间序列模型方法与极值理论方法建模求解系统最大值指标的特点进行了对比分析,并给出了相关结论。
A testing methodology for the system maximum-error specification based on the extreme value theory is proposed.It an provide a theoretic basis for the testing of maximum-error specification on long-working systems in some special fields,such as aviation,aeronautics and national defenses.Firstly,the error series of system outputs is divided into blocks by obeying some essential principals,and the corresponding maximum error series which are nearly independent with each other are gotten.Secondly,the extreme value models for the maximum error series are built and the return level and its uncertainty are obtained under some return period.Then,the relationships between the return level and the maximum-error specification are researched and the system maximum-error specification is tested and evaluated by using classic statistical theories.After that,a practical sample is given to introduce the methodology more clearly,which shows that this methodology is feasible.Finally,by comparing the proposed methodology with that based on the time series models from different points of view,some conclusions are given.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2012年第5期1073-1084,共12页
Systems Engineering and Electronics
基金
教育部博士点新教师基金(200802881012)资助课题
关键词
质量控制与可靠性管理
评定方法
广义极值
最大值指标
quality control and reliability management
testing methodology
generalized extreme value(GEV)
maximum-error specification