自适应BP网络在复合材料脱层损伤信号处理中应用的研究.doc

南京航空航天大学 硕士学位论文 自适应BP网络在复合材料脱层损伤信号处理中的应用研究 姓名：华秋萍 申请学位级别：硕士 专业：测试计量技术及仪器 指导教师：熊克 南京航空航天大学硕士学位论文  摘  要  结构健康监测是当前学术界研究的一个热点问题，也是智能材料与结 构研究领域的一个重要研究方向。复合材料损伤检测技术则是结构健康监 测中的一个重要研究内容。 较全面地评述了神经网络的应用模型、改进算法等，利用节点法确定 了试件的宏观弹性系数及密度，提出了坐标相同、节点号不同模拟脱层损 伤的新思路；本文采用 Lamb 波主动监测技术，通过选择恰当的激励信号， 激励出可进行 Lamb 波分析的主动监测信号，进而对监测信号进行分析及模 式处理，初步实现了损伤定位；制备了纤维增强复合材料试件，并进行实 验模态分析；针对基于实验数据训练神经网络存在样本不足的缺陷，本文 提出了利用有限元方法对含有脱层损伤的复合材料试件进行数值模拟，以 前五阶弯曲模态频率构建训练样本的新思路，并提出了一种初步的计算值 修正方法，以此构建神经网络的训练样本；重点介绍了利用 MATLAB 开发 的自适应 BP 神经网络程序，将实验模态分析的结果送入训练好的神经网络 进行预测，实现了对复合材料梁脱层损伤的定位。 本文工作得到了国家自然科学基金项目；国家自然科学基 金重点项目的资助。 关键字：复合材料，脱层损伤，损伤定位，信号处理，BP 网络 i 自适应 BP 网络在复合材料脱层损伤信号处理中的应用研究 Abstract Structural health monitoring system is an important research direction in the hot area of smart material and structures and Composite damage detection is an important research content of structural health monitoring. A compressive discussion about neural network and its improved algorithm were presented in this paper, using node law to confirm composite’s grand elasticity and consistency, putting forward same coordinate but different node mark to simulate delamination. This study adopts active detection technique based on Lamb Wave to monitor damage in composite materials. The waves emitted interact with discontinuities and experience a change in their propagation characteristics when damage is generated. By comparing sensor signals collected before and after damage happens, damage location can be determined.Two carbon fiber reinforced composite beams were fabricated , and their model frequencies were measured by an experiment method. A novel method combining computational mechanics and neural network was demonstrated for composite health monitoring; The first five flexure model frequencies obtained by FEM were modified by a primary revising approach and were used to train the neural network. A program of self-adaptive BP neural network using MATLAB was finished and the first five flexure exp