机构地区: 浙江大学航空航天学院
出 处: 《振动工程学报》 2003年第1期41-45,共5页
摘 要: 旋转机械启动过程中的振动信号包含有重要的参考信息 ,研究该过程中的旋转机械运转状况 ,有助于发现存在于系统中的早期故障。根据离散隐 Markov模型 (DHMM)的建模理论 ,对旋转机械启动过程的振动谱矢量序列进行标量量化处理 ,并建立了各种模拟故障的 DHMM进行故障分类尝试。实验证明 。 Vibration signals of a rotating machine in run up stage contain reference information of equipment. Therefore the information of the running status in this stage is helpful to detect early latent faults in the system. Based on the Discrete Hidden Markov Model (DHMM) theory, a series of FFT spectrum vectors of rotor vibration signals in the run up stage are quantized scalar and DHMMs are constructed to classify the modeling faults. Experiment results show that the method is very effective.