基于时频分析的汽油机爆震特征的提取  

Knock Feature Extraction for a Gasoline Engine Based on Time-Frequency Analysis

作　　者：杨建国[1] 张建峰[1] 刘晓峰[1] 

机构地区：[1]哈尔滨工业大学汽车工程学院,威海264209

出　　处：《内燃机工程》2003年第3期33-37,共5页Chinese Internal Combustion Engine Engineering

摘　　要：分析了短时傅立叶变换 (STFT)、维格纳 瑞利分布 (WVD)及小波变换 (WT)等时频分析方法各自的特点。研究了利用三种方法分别从汽油机缸内压力信号和缸盖上的振动信号中提取爆震特征的可行性和有效性。提出了相应的提取爆震特征的方法。分析和应用结果表明 ,STFT和WVD直接提取爆震特征效果差 ,但能够有效地从滤波后的信号中提取爆震特征。离散小波变换 (DWT)具有明显的优势 ,能够直接从压力信号和振动信号中有效地提取爆震特征。The characters of short time Fourier transform (STFT), Wigner Ville distribution (WVD) and wavelet transform (WT) are analyzed. The feasibility and effectiveness of using STFT, WVD and discrete wavelet transform (DWT) to extract knock feature from a pressure signal measured from a gasoline engine cylinder and a vibration signal measured from a gasoline engine cylinder head are studi. A related approach to extract knock feature based on STFT, WVD and DWT is presented. It is shown that knock feature cannot be directly extracted well from a measured signal by STFT and WVD, but it can be effectively extracted from the filtered signal; DWT is more effective, knock feature can be extracted well directly from a filtered signal by it.

关 键 词：内燃机 爆震 时频分析 

分 类 号：TK411.22[动力工程及工程热物理—动力机械及工程]