基于隐Markov模型的汽轮机运行状态识别

Application of hidden markov model in turbine state recognition

本文首先介绍HMM的基本理论。给出汽轮机振动幅值谱的标量量化方法,并在此基础上提出基于DHMM的火电厂汽轮机升负荷模型的状态识别方法。然后利用汽轮机升负荷的轴振数据验证DHMM状态识别方法的有效性。最后,总结本文研究内容并分析了HMM在汽轮机运行状态识别方面存在的问题。

In this paper,the theories of HMM were studied.Based on the scalar quantization theory,an approach to recognize the running conditions of turbine bused on DHMM was proposed.The experimental data from the turbine under the turbine speed increment process was used to verify the effect of this method.Finally,all the work of this paper was summarized and problems in this field were analyzed.