在分析1(1/2)维谱的定义与其计算方法的基础上,结合信息熵理论,提出了一种1(1/2)维谱熵的特征提取方法。通过建立评价发动机状态的特征熵指标,真实反映了机械故障信号的复杂性程度。应用结果表明,该方法对不同状态的信号具有较好的分美能力,实现了对发动机状态的有效识别。
Based on analyzing the concept of 1 (1/2) dimensional spectrum and its calculation method, the feature extraction method of 1 (1/2) dimensional spectrum entropy was proposed hy combining with the theory of formation entropy. Then the feature entropy index to estimate the engine state was formed, which reflected truly the complexity of mechanical fault signal. The results indicate that 1 (1/2) dimension spectrum entropy has a good recognition of different state signals can recognize engine state effectively.