摘要
旋转机械的连续监测日益成为重要问题。由于传感技术的发展和计算能力的提高 ,使设备的在线监测与故障诊断成为可能。把线性AR系数作为暂态振动信号的特征 ,利用有限状态隐马尔可夫模型 (简称HMM )来模拟旋转机械的运行过程中动态行为的变化 ,从而提出了一种新的故障诊断方法。HMM的参数从监测数据的统计关系中导出。
Continuous monitoring of rotating machine is an increasingly important issue. Recent advances in both sensor technology and computational capabilities have made on line monitoring and fault diagnosis possible. Exacted AR coefficients for the features of temporal vibration signal and used finite states Hidden Markov chain to model changing behavior of rotating machine in running process. Therefore, proposed a new method for faults diagnosis. The parameters of the Hidden Markov Model (HMM) are derived statistically from monitoring data. It is verified that this method is feasible.
出处
《汽轮机技术》
北大核心
2002年第5期301-303,共3页
Turbine Technology
基金
国家自然科学基金项目 (项目编号 :5 0 0 75 0 79)
