基于MEMD和ELM的飞机机翼健康状态预测技术

复合材料在现代飞机结构中的应用越来越广泛,为了有效地对飞机机翼健康状态进行预测,提出了基于多元经验模态分解(MEMD)和极限学习机(ELM)的飞机机翼健康状态预测方法。以某型飞机复合材料机翼盒段为具体研究对象,对其进行冲击与疲劳加载试验,利用光纤传感器及其采集系统募集飞机复合材料机翼盒段的原始应变信息,对其健康状态予以表征。对所采集的原始应变信息进行MEMD分解,提取分解后各频带信号的能量熵作为表征飞机复合材料机翼盒段健康状态的特征信息,采用动态主元分析法(DPCA)将所提取的能量熵特征信息进行融合,采用融合后所得到的能量熵构建ELM预测模型,对某型飞机复合材料机翼盒段健康状态进行预测。试验研究表明,本文方法可以有效实现飞机机翼的健康状态预测,具有很好的应用前景。

Health state prediction technique for aircraft wing based on MEMD and ELM

Composite materials are more and more widely used in the modern aircraft structure.In order to effectively forecast the health state of aircraft wing,the multivariate empirical mode decomposition (MEMD) energy entropy and extreme learning machine (ELM) model are introduced into prediction method of aircraft wing health state.A certain type of aircraft's composite wing box section is taken as the specific research object.Then we carried out fatigue load tests after impacting it.The original strain information of aircraft's composite wing box section was obtained by fiber optic sensor acquisition system,and the health state of aircraft's composite wing box section was represented.Then,the original strain information was decomposed by MEMD,and the energy entropy was extracted from the band signal which is decomposed with MEMD as the feature.The energy entropy was then fused by dynamic principal component analysis (DPCA).The energy entropy after fusion was taken to construct the ELM prediction model.And we forecasted the structural damage of a certain aircraft's composite wing box section.Experimental research shows that the method can achieve the prediction of aircraft wing health state effectively and has a very good application prospect.