大型旋转机械升降速过程故障诊断HMM-AR方法研究

STUDY ON HMM AR METHOD IN FAULT DIAGNOSING IN SPEEDING UP OR DOWN PROCESS OF LARGE ROTATING MACHINERY

由于大型旋转机械升降速过程振动的复杂多样性，对其进行监测能获得较平稳运行更多的机组状态信息。本文通过提取 Ａ Ｒ 系数反映振动特征，并将隐 Ｍ ａｒｋｏｖ 模型引入升降速过程故障诊断研究中，实验证明应用 Ｈ Ｍ Ｍ  Ａ Ｒ 法能较好地达到机组故障诊断目的。

The message of running state of a large rotating machinery induded in the monitoring signals in speeding up or down process is usually more abundand than that in stationary process. In the paper, the Hidden Markov Model is employed and the AR coefficents are taken out to describe the vibration characters of a rotating machine during speeding up or down for fault diagnosing. By experiments, the method of HMM AR is proved to be a good measure for machinery trouble shooting in that case.