针对碳纤维复合材料层合板弯曲失效过程开展了声发射监测试验，并对试验采集的声发射信号进行了K-均值聚类分析，提取了不同损伤类型的声发射信号，对每种损伤类型的信号利用小波包特征熵的分析方法，选取能反映不同损伤类型的特征参数，实现对碳纤维复合材料不同损伤信号的有效识别，为碳纤维复合材料的损伤监测提供了理论依据。

In this paper, the acoustic emission monitoring test was used for bending failure process of carbon fiber composite laminated plate. K-means cluster analysis was applied to processing the acoustic emission signals that were collected from experiments. Acoustic emission signals of different damage types were extracted. The analysis method of wavelet packet feature entropy was used to select the characteristic parameter that can reflect different damage types in signal for each damage types. The different damage signals of carbon fiber composites can be identified effectively. It provides the oretical basis for damage monitoring of carbon fiber composite.