基于双树复小波包的柴油机时频奇异谱特征提取研究

Study on diesel engine feature extraction of time-frequency singular value spectrum based on dual-tree complex wavelet packet transform

针对传统时频分布计算量大、分析速度慢和特征提取复杂的缺点,提出一种基于双树复小波包时频奇异谱的柴油机特征提取方法,研究了基于双树复小波包变换的基本原理及其离散时频分布,运用奇异值分解方法对所得的离散时频分布进行特征提取,获得离散时频奇异谱。结合柴油机不同工况测试试验,对比分析了该离散时频特征提取方法的计算效率、不同工况下的特征提取效果,试验结果表明:该方法分析速度快,能够有效表征柴油机不同的运行工况。

In view of the limitation of mass calculation, slow analysis velocity and complicated feature extraction of traditional time-frequency, the feature extraction method of diesel engine based on time-frequency distribution singular value spectrum was proposed. The discrete Dual-Tree Complex Wavelet Packet Transform （DT-CWPT） transform method and the discrete time-fre- quency distribution was studied ,then the feature was extracted by singular value decomposition. The comparison of calculation ef- ficiency and feature extraction effect was analyzed based on different working condition test of diesel engine. The results indicate that the method has high computation efficiency and good performance of accuracy.