基于振动信号的包装件损伤预测技术研究

Research on Damage Prediction Technology of Package Based on Vibration Signal

为了对包装件的疲劳损伤进行预测,基于振动信号,提出一种HHT边际谱特征和相关向量机(RVM)相结合的包装件疲劳损伤预测方法。首先,设计了包装件的振动损伤测试系统并进行试验方案设计,在此基础上采集损伤状态数据;然后采用希尔伯特-黄变换HHT方法及边际谱特征对采集的振动信号进行预处理,采用多距离形态相似度评估方法构建包装件损伤指数;最后,将损伤指数作为相关向量机(RVM)的建模数据,结果表明该模型达到了较好的预测效果,实现了基于振动信号的包装件损伤预测。

In order to predict the fatigue damage of the package,based on the vibration signal,a fatigue damage prediction method of package with HHT marginal spectrum feature and correlation vector machine(RVM)was proposed.Firstly,the vibration damage test system of the package was designed and the test scheme was also designed.Based on this,the damage state data was collected.Then the Hilbert-Huang transform HHT method and the marginal spectral features were used to preprocess the acquired vibration signals,and the multi-distance morphological similarity evaluation method was used to construct the package damage index.Finally,the damage index was used as the modeling data of the correlation vector machine(RVM).The results show that the model achieves a good prediction effect and the damage prediction of package based on vibration signal was realized.