Abstract: Based on wavelet package transform technology, this paper proposes a method for analyzing the frequency characteristics of nonsteadystate signals in an energyratio and frequencyband coordinate system. Result of this method is compared with that of common frequency spectrum method. The correctness and feasibility of this method is verified. Meanwhile, taking the fan's noise signal of a fuel cell vehicle as an example with a lowfrequency sinusoidal interference signal chosen as the representation of periodic signals, the periodic signals hidden in the nonsteadystate signals are identified. Finally, the fan's noise signals are analyzed. The noise due to low and medium frequency vibration and the noise due to high frequency vibration in the signals are separated. Their information is extracted in timedomain. The measures for further noise reduction are proposed.

Key words: vibration and wave, nonsteadystate noise signal, wavelet package transform, signal analysis, fuel cell vehicle

摘要： 基于小波包变换理论，提出一种在能量比-频带二维坐标空间内分析非稳态信号频率特性的方法，并与常用频谱分析方法结果做了对比，说明该方法的正确性和可行性。接着以燃料电池汽车的风机噪声信号为例，选取一个低频正弦干扰信号作为其他各种周期信号的代表，证明该方法可以识别出非稳态信号中掺杂的周期信号。最后重点分析风机噪声信号，成功把信号中低频振动噪声与高频气动噪声进行分离，分别提取时域信息，分析并提出进一步降噪的措施。

关键词: 振动与波, 非稳态信号, 小波包变换, 信号分析, 燃料电池汽车