基于贝叶斯方法的在线监测I/O卡件故障率研究  

Based on Bayesian Method Study on Failure Rate of On-line Monitoring I/O Card

作　　者：赵先浩 刘飞 ZHAO Xianhao;LIU Fei(Anhui NARI Jiyuan Electric Power System Technology Co.,Ltd.,Hefei 230088,China)

机构地区：[1]安徽南瑞继远电网技术有限公司,安徽合肥230088

出　　处：《山东电力技术》2019年第1期36-38,51,共4页Shandong Electric Power

摘　　要：针对DCS可靠性数据少,样本少,难以获得的特点,基于研发的DCS各系统可靠性监测系统,获取设备运行的特有样本数据,对寿命服从指数分布的I/O卡件,运用贝叶斯方法分析其故障率。利用共轭分布法选取先验分布;采用先验矩的方法估计先验分布的超参数;利用伽玛分布性质进行指数分布故障率的估计,修正I/O卡件的故障率。算例分析表明,该方法求取I/O卡件故障率具有可行性。DCS reliability is not fully studied while the data for relevant evaluation is hard to obtain.To solve this problem,the DCS system reliability monitoring system is developed and introduced to access the unique sample data during equipment operation.The failure rate of the I/O cards whose life expectancy is exponential distributed,is analyzed using Bayesian method.The prior distribution is selected using the conjugate distribution method.The prior moment method is used to estimate the hyper-parameters of the prior distribution.The failure rate of the exponential distribution is plotted using gamma distribution property.Eventually,the fault rate of the I/O card is modified.The example shows that this method is feasible for obtaining the failure rate of the I/O cards.

关 键 词：指数分布 I/O卡件 贝叶斯分析 故障率 

分 类 号：TM621[电气工程—电力系统及自动化]