基于离散余弦变换的IMF特征频率提取

摘要

针对经验模态分解(EMD)处理后,固有模态函数(IMF)分量通常会受到其他频率的干扰,导致物理意义不明确,提出应用离散余弦变换提取IMF中轴承故障特征信息的方法.该方法利用离散余弦变换对能量的集中性质,通过对离散余弦变换系数进行硬阈值处理重构信号,能够在时域有效凸显IMF的主要规律,在频域准确提取IMF的故障频率.%The intrinsic mode function (IMF) is usually interfered by other frequencies after empirical mode decomposition (EMD) analyzing, which leads physical significance not definitude. The discrete cosine transform is applied to extract fault characteristic information from bearings. By using the characters of discrete cosine transform being centralized for energy and processing discrete cosine transform coefficient by hard threshold, the law of IMF in time domain and the fault frequency of IMF in frequency domain could be identified. The effectiveness of this method is confirmed by using simulated signal and real time signal.