论文部分内容阅读
提出了基于经验模式分解(EMD)和隐马尔科夫模型(HMM)的故障诊断模型,为通过设备状态监测数据分析进行基于状态维修和维修决策提供了一种新途径.为了消除EMD的端点效应,使用神经网络拟合延拓原始数据序列端点极值,并通过定义序列复杂度来定性地确定延拓极点数.进一步,采用分解所得的固有模态(IMF)能谱熵作为HMM分类系统的输入,得到一种设备故障诊断方案.通过数值仿真和发动机故障诊断验证了该方法的有效性.
A fault diagnosis model based on Empirical Mode Decomposition (EMD) and Hidden Markov Model (HMM) is proposed, which provides a new way for decision-making based on state maintenance and repair through equipment condition monitoring data analysis.In order to eliminate the end effect of EMD , The endpoint extremum of the original data sequence is extended by using the neural network fitting, and the extension pole number is defined qualitatively by defining the sequence complexity.Furthermore, the energy entropy of the intrinsic mode IMF obtained by the decomposition is used as the HMM classification system Input and get a device fault diagnosis scheme.The effectiveness of this method is verified by numerical simulation and engine fault diagnosis.