UHMWPE/LDPE层合板损伤声发射信号聚类分析研究

Cluster Analysis on AE Signal during Damage Process of UHMWPE/LDPE Laminates

为揭示UHMWPE/LDPE层合板损伤机理,运用声发射技术对源于基体损伤、纤维断裂、纤维/基体界面损伤的声发射信号进行聚类分析。以距离为相似性测度,按分层聚类方法提取幅度、持续时间和峰值频率为模式特征,并按k-means聚类算法实现UHMWPE/LDPE层合板声发射信号源机制的分类。实验结果表明,基体损伤、纤维断裂和界面损伤等不同损伤机制产生的声发射信号,在幅度、持续时间和峰值频率等参数上不同,研究结果对UHMWPE/LDPE层合板损伤模式识别具有参考价值。

To reveal damage mechanism of UHMWPE/LDPE laminates, this study investigated cluster analysis on acoustic emission（AE） signals generated from matrix damage, fiber breakage and interface damage by use of AE technology. Pattern features such as amplitude, duration time and peak frequency could be extracted by hierarchical clustering with distance measure, then the AE signals resulted from different damage mechanisms of UHMWPE/LDPE laminates could be identified by k-means cluster analysis. The result revealed parameters of AE signal such as amplitude, duration time and peak frequency resulted from different damage mechanisms were different. This result can be used to investigate pattern recognition on damage mechanisms of UHMWPE/LDPE laminates.