论文部分内容阅读
飞机结构的损伤严重影响着飞机的飞行安全,为了解决飞机复合材料结构损伤难以有效识别问题,本文提出一种基于广义回归神经网络(General regression neural network,GRNN)与极限学习机(Extreme learning machine,ELM)组合的飞机复合材料结构损伤识别新方法。首先对飞机复合材料层合板进行冲击,而后对其进行疲劳拉伸试验,通过优化布局在复合材料层合板上的光纤光栅传感器募集应变信息,并对其进行预处理。采用变分模态分解(Variational mode decomposition,VMD)对募集的应变信息进行自适应分解,得到多个基本模式分量(Intrinsic mode function,IMF)。计算各阶IMF分量的奇异熵,通过核独立主元分析(Kernel independent component analysis,KICA)方法对奇异熵进行特征融合,构建融合特征向量。采用融合特征向量建立基于GRNN-ELM的复合材料结构损伤识别模型,通过试验对损伤识别模型的有效性进行了验证,并分别与所构建的ELM和GRNN损伤识别模型的识别结果进行比较。结果表明,该方法能有效对飞机复合材料结构损伤进行识别,具有很好的工程应用价值。
Aircraft structural damage seriously affects the flight safety of the aircraft. In order to solve the structural damage of the aircraft composite material, it is difficult to effectively identify the problem. In this paper, a general regression neural network (GRNN) and Extreme Learning Machine ELM) combined structural damage identification of aircraft composite materials. The aircraft composite laminates were first subjected to an impact test and then subjected to a fatigue tensile test. The strain information was recruited by optimizing the FBG sensors laid on the composite laminates and pretreated. Variational mode decomposition (VMD) is used to adaptively decompose the collected strain information to obtain a plurality of Intrinsic mode functions (IMFs). The singular entropy of IMF components of each order is calculated, and the singular entropy is fused by kernel independent component analysis (KICA) method to construct fusion eigenvectors. The composite structure damage identification model based on GRNN-ELM is established by using the fusion eigenvector. The effectiveness of the damage identification model is verified through experiments. The recognition results of the damage identification model of ELN and GRNN are compared respectively. The results show that this method can effectively identify the structural damage of aircraft composite materials and has good engineering application value.