小样本条件下一种旋转机械振动状态阈值的确定方法及应用

摘要

针对旋转机械振动状态阈值设定时数据样本不足的情况,给出了一种基于Bayes-随机加权法的阈值确定方法.该方法综合运用了随机加权法、Bayes理论和验前信息融合技术,实现小样本条件下的振动特征值统计参数估计.通过与经典统计法和随机加权法的仿真对比研究,证明了其在小样本统计参数估计应用中的优越性.对离心泵机组故障模拟平台进行了整机振动状态阈值确定实验,能有效的实现设备振动状态评估.%The insufficient data sample causes the difficulty when setting the rotating machinary vibration threshold. A threshold-setting ways based on the Bayes-random weighted method was presented, which is combining the random weighting method, Bayes theory and pre-test information fusion technology for estimated the statistic parameters of the vibration eigenvalues on small sample. By comparing with the classical statistical method and the random weighted method, the simulating study indicated the advantage of applying the Bayes-random weighted method to estimating the statistic parameters on small sample. At the end of the paper, the vibration threshold-setting experiment on the centrifugal pump fault-simulating platform was done and the results showed that the Bayes-random weighted method can evaluate the vibration status of the rotating machinery effectively.