论文部分内容阅读
为了解决直升机动部件疲劳损伤类型识别问题,提出了一种基于谐波小波包特征提取和层次支持向量多分类器的声发射源类型识别方法.声发射信号经过4层谐波小波包分解后,提取各个频段的能量特征用于声发射源类型识别,克服了传统小波包分析能量泄露、频带选取不灵活、不同层频率分辨率不同的缺点.首先,利用已知声发射源类型的试验数据训练层次支持向量多分类器,然后,利用其余试验数据进行测试.碳纤维材料试件压断试验结果表明:该方法有效地实现了声发射源多类识别,并且在计算效率和识别精度上都优于小波包特征提取方法.
In order to solve the problem of identifying the type of fatigue damage of helicopter moving parts, an acoustic emission source type identification method based on harmonic wavelet packet feature extraction and hierarchical support vector multi-classifier is proposed. After the AE signal is decomposed by the 4th harmonic wavelet packet, The energy features of each frequency band are extracted and used to identify the types of acoustic emission sources, which overcomes the shortcomings of traditional wavelet packet analysis such as energy leakage, inflexible band selection and different frequency resolution at different layers.Firstly, using the experimental data of known acoustic emission sources Level support vector multi-classifier, and then use the remaining test data to test.Carbon fiber material specimen compression test results show that: This method effectively achieve a variety of acoustic emission source identification, and in terms of computational efficiency and recognition accuracy are better than Wavelet packet feature extraction method.