批次定检设备故障诊断的贝叶斯方法研究

针对机载设备故障诊断定位率低的问题,提出充分利用先验故障信息指导批次定检的思想.通过研究批次定检机载无线电设备的失效规律,运用贝叶斯方法对小子样情况下的故障概率进行估计,建立了以各模块的故障概率来指导下一步故障检测的机制.针对不同情况提出3种指导方法并进行了仿真计算,结果表明可显著提高故障的一次定位准确率,减少故障查找时间,具有较高的工程应用价值. 

Bayesian Method for Fault Diagnosis of Equipment at Lot Assure Tests

To resolve the problem of low locating rate in fault diagnosis, efforts were made by using preceding fault information fully in lot assure tests. By studying the failure rules of air-bone radio equipment in lot assure tests, the fault probability under extreme-small-sample conditions was estimated using Bayesian method. An approach for direct fault detecting in next step was established by using the fault probabilities of different modules. Three methods of direction were adopted to simulate those probabilities, the results showed that the strategy of the present paper can effectively improve the rate of locating faults in only one detecting progress and reduce fault|finding time.