基于非齐次泊松过程和统计仿真的故障样本模拟生成  

Fault Sample Generation Based on Nonhomogeneous Poisson Process and Statistical Simulation

作　　者：张勇[1,2] 邱静[1,2] 刘冠军[1,2] 陈循[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院,长沙410073 [2]国防科学技术大学装备综合保障技术重点实验室,长沙410073

出　　处：《机械工程学报》2012年第15期75-82,共8页Journal of Mechanical Engineering

基　　金：国家自然科学基金(51105369);国家级基础科研计划重点资助项目

摘　　要：由于测试性虚拟试验具有成本低、效率高、风险小、故障注入受限少等优点,故障样本量几乎不受限制,可有效弥补测试性实物试验的不足,但同时也对故障样本生成提出新的要求。为此提出一种适用于测试性虚拟试验的基于非齐次泊松过程和统计仿真的故障样本模拟生成方法。分析指出可修系统的故障发生过程是随机过程,并用非齐次泊松过程及其参数化模型对其进行数学描述。给出故障样本模拟生成流程,建立故障事件发生间隔时间的概率分布函数,通过随机数生成和逆变换法,实现故障样本的模拟生成,仿真获得故障发生次数及其相继发生时间。以某型地平仪为案例进行试验和应用研究。试验结果表明,采用所提方法进行故障样本模拟生成是有效的,能科学指导可修系统测试性虚拟试验中的故障注入。Virtual testability test has many advantages, such as low cost, high efficiency, small risk, and little restricted. The fault sample size is almost unlimited in virtual testability test. It can effectively compensate for the shortcomings of physical testability test. However, fault sample generation method needs to be changed and improved. A novel approach for fault sample generation based on nonhomogeneous Poisson process and statistical simulation is presented. It is pointed out that faults occurrence process is stochastic process. This process is mathematical described by parametric nonhomogeneous Poisson process, such as linear model, power law model and log-linear model. The procedure Of fault sample generation is put forward. Probability distribution function of the interarrival time is established. Fault sample is generated by random number generation and inverse transformation method. The proposed method is applied to a gyro. The example indicates that the fault samples generated by the proposed method are valid. The proposed method can be applied to guide fault injection in virtual testability test.

关 键 词：测试性虚拟试验 故障样本 非齐次泊松过程 统计仿真 

分 类 号：TH17[机械工程—机械制造及自动化] O211[理学—概率论与数理统计]