调幅-调频信号的经验模态分解包络技术和模态混叠

发布时间：2018-04-30 22:32

本文选题：经验模态分解 + 包络算法　；参考：《机械工程学报》2017年02期

【摘要】：经验模态分解(Empirical mode decomposition,EMD)是一种数据驱动的自适应非线性非平稳数据处理方法。包络技术和模态混叠问题是EMD研究的重要课题。将非线性非平稳信号定义为多成分调幅-调频(Amplitude modulation-frequency modulation,AM-FM)信号模型,而EMD分解的每一个固有模态函数为单一的AM-FM信号。通过研究单成分AM-FM信号的包络以及多成分AM-FM信号EMD分解引起的模态混叠问题,提出新的EMD包络条件,并给出新包络算法的数值计算方法。基于新条件包络算法,提出单成分AM-FM信号相位和瞬时频率的新估计算法。提出解决多成分AM-FM信号EMD分解的模态混叠问题的新方法,并通过几组仿真信号和一组实测的转子碰摩数据验证了方法的有效性。
[Abstract]:Empirical mode decomposition (EMD) is a data-driven adaptive nonlinear non-stationary data processing method. Envelope technique and modal aliasing are important subjects in EMD research. The nonlinear non-stationary signal is defined as a multi-component amplitude modulation-frequency modulation modulation-frequency modulationAM-FM signal model, and each inherent mode function of EMD decomposition is a single AM-FM signal. By studying the envelope of single-component AM-FM signal and the modal aliasing caused by EMD decomposition of multi-component AM-FM signal, a new EMD envelope condition is proposed, and the numerical calculation method of the new envelope algorithm is given. Based on the new conditional envelope algorithm, a new estimation algorithm for phase and instantaneous frequency of single component AM-FM signal is proposed. A new method to solve the modal aliasing problem of multi-component AM-FM signal EMD decomposition is proposed. The validity of the method is verified by several sets of simulation signals and a set of measured data of rotor-rub-impact.
【作者单位】：西南交通大学牵引动力国家重点实验室;
【基金】：国家自然科学基金(61134002,51305358)资助项目
【分类号】：TN911.7

【参考文献】