基于HMM的核动力旋转机械故障诊断的研究

Research on Nuclear Power Rotating Machinery Fault Diagnosis Based on HMM

隐Markov模型(HMM)已经证明是学习动态时间序列的概率模型的最广泛应用的工具之一,它可以使用一个隐变量来模拟系统的动态行为的变化。核动力旋转机械升速过程具有信息量大、信号非平稳、重复再现性不佳等特点,HMM很适合处理此类信号。将HMM引人到核动力旋转机械的故障诊断中,提出了一种基于HMM的故障诊断方法。

Hidden Markov models （ HMM ） have learning probability models of dynamics time series. proven to be one of the most widely used tools for HMM can model dynamical behavior variation existing in the system through a latent variable. There are a large amount non-statistical, worse reappearance signal in the nuclear power rotor run-up process. HMM is suitable to deal with these signals. A new fault diagnosis strategy based HMM was proposed in this paper.