【作者】 王慧；

【导师】 李亚东；

【作者基本信息】 广州大学 ， 结构工程， 2021， 硕士

【摘要】 纤维增强复合材料(FRP)由于轻质高强和耐腐蚀性等优异的性能,在航空、海洋和汽车工业等各种工程学科中得到了广泛的应用。然而,由于其基体的层间断裂韧性低,容易发生分层或层间损伤,这可能会显著降低FRP材料的刚度,影响结构完整性进而导致人员安全危险和经济的严重损失,且分层损伤通常从表面难以观察到。因此,有必要研究一种可靠有效的分层检测方法。本文以FRP复合材料层合悬臂梁为研究对象,提出了一种基于振动模态数据的逆向检测方法,利用模态数据的变化比和基于经验的学习算法来量化复合材料中分层损伤的位置、尺寸和界面层。对文中提出的用于分层损伤识别的方法进行了数值模拟案例和试验实测案例的双重验证:(1)数值案例中研究并比较了三种类型的目标函数,包括只使用频率变化比的函数(F-based方法)、只使用模态形状变化比的函数(MS-based方法)以及它们的组合函数(F&MS-based方法);建立了无损和有损分层悬臂梁的三维有限元模型并提取了其模态数据,比较了分别使用4、6、8、10阶模态数据对损伤识别结果的影响;对比了使用其他三种不同的算法(CMFOA、PSO和SSA算法)识别悬臂层合梁的损伤效果;研究并对比了F-based方法和F&MS-based方法的鲁棒性。(2)设计制作了无损伤和含分层损伤的悬臂梁试件并对其进行模态测试获得了试件的实测频率值,再将实测数据用于F-based方法和F&MS-based方法以预测FRP悬臂梁试件中的分层损伤(实验验证)。数值结果表明,所提出的F&MS-based方法在复合材料的三个分层变量(位置、尺寸和界面层)中都得到了优异的结果,并且确定了模态振型信息的添加能大大提高对离散分层变量z(界面层)的识别准确度;收敛速度和识别精度都不一定随着模态数据的增加而提高,这与实测中的噪声水平有关;F&MS-based方法对固有频率高达7%的噪声水平和模态振型高达10%的噪声水平都不敏感,具有很强的鲁棒性。使用EBL对层合梁进行损伤识别的方法比使用CMFOA、PSO和SSA优化算法等其他三种较常用的优化算法有着更具竞争力的识别结果。实验结果再次验证了基于模态数据的分层检测方法(F&MS-based方法)的识别精度。更多还原

【Abstract】 Fiber-reinforced polymer(FRP)composite laminates have been widely and popularly used in various engineering disciplines,such as aeronautical,marine,and automotive industries,due to their outstanding mechanical properties,such as lightweight and ease of manufacture.However,delamination or interlaminar damage is susceptible to occur because of the low inter-laminar fracture toughness of the matrix,which may significantly reduce the stiffness of the FRP laminates and then lead to a severe loss of structural integrity.Meanwhile,delamination is usually not visible from the surface.Therefore,it is necessary to develop a reliable and efficient method for delamination detection.Present work focuses on assess delamination in composite laminated cantilever beams by using the change ratio of modal data and an experience-based learning algorithm.A reverse detection method based on vibration modal data is proposed to quantify the location,size and interface layer of delamination damage in FRP composite laminated cantilever beams.The methods for delamination identification proposed in this paper are validated both numerically and experimentally.(1)In the numerical case,three types of objective functions are studied and compared,including the function using frequency change ratio only(F-based method),the function using modal shape change ratio only(MS-based method)and their combination function(F&MS-based method);Three-dimensional finite element models of intact and delaminated cantilever beams were established and their modal data were extracted.The effects of modal data of order 4,6,8 and 10 on delamination identification structures were compared;Three other algorithms(CMFOA,PSO and SSA)are compared for delamination identification of cantilever laminated beams;The robustness of F-based method and F&MS-based method is studied and compared.(2)The cantilever beam specimens without damage and with delamination damage were designed and manufactured,and the measured frequency and mode shape values of the specimens were obtained through modal tests.Then the measured data were used in the F-based and F&MS-based method to predict the delamination damage in the FRP cantilever beam specimens(experimental verification).Numerical results show that the proposed F&MS-based method can obtain excellent results in the three delamination variables(position,size and interface layer)of the composites.It is confirmed that the addition of mode shape information can significcantly improve the accuracy of interface layer prediction;The convergence speed and recognition accuracy may not improved with the increase of modal datasince they are generally related to the measured noise level;The F&MS-based method is not sensitive to the measurement noise even under the noise as high as 7% for frequency data 10% for mode shape data;In comparison to other three state-of-the-art optimization algorithms including CMFOA,PSO,and SSA,the proposed EBL is able to obtain more competitive results in delamination detection of laminated beams.The experimental results further indicated that the F&MS-based method can accurately detect the delamination in composite laminates.更多还原