基于EMD功率谱方法的机电作动系统早期故障特征提取

Electro-Mechanical Actuation System Early Fault Feature Extraction Based on EMD Power Spectrum Method

提出一种改进的基于经验模态分解(EMD)功率谱分析的特征提取方法,该方法自适应地将故障信号分解为不同频段上的基本模式分量(IMF)之和,然后只提取对故障敏感的IMF分量的功率谱。由于只有少数IMF对故障敏感,该方法能获得具有较高信噪比的原数据特征信息,这对预测功率变换器早期故障趋势具有极其重要的应用价值。实验结果表明,该方法可有效提取早期隐性故障特征,为飞机机电作动系统的故障预测和健康评估提供必要的前期准备。

An improved feature extraction method based on empirical mode decomposition（EMD） power spec- trum analysis is presented. This method decomposes the fault signal into the sum of intrinsic mode function （IMF） adaptively, then only extracts the power spectrum of IMF which is sensitive to the fault. Thus, this meth- od obtains the high signal-noise ratio characteristic information of raw data, which is value to early fault fore- cast. As simulation result shows that this method can extract the forepart recessive fault feature, and provide the necessary forepart provision for fault prognostics and health evaluation of airplane electro-mechanical actuation svstem.