Study for correlations of mechanical properties in CFRP laminates to interlaminar damage by neural network

本文讨论了损伤增加对CFRP层合板静、动力性能的影响，以及作为宏观损伤的孔洞对含孔洞CFRP层合板固有频率的影响。此外，还研究了用神经网络识别CFRP层合板的边界条件。研究结果包括：1.由于损伤面积的变化与CFRP层合板的阻尼力的变化成正比，因此可以通过阻尼力的大小来预测损伤面积。抗弯强度、损伤面积和阻尼力之间存在一定的相关性。阻尼力越大，抗弯强度越小。在压缩过程中，破坏模式取决于阻尼力的大小。在小阻尼量和大阻尼量下，得到了屈曲破坏和纯压缩破坏模式。含孔碳纤维复合材料层合板的数值计算结果与试验结果吻合较好，表明本文提出的方法能较好地反映碳纤维复合材料层合板的固有频率和振型。利用本文提出的神经网络中几个节点的固有频率、振型值的输入数据，可以识别CFRP层合板的边界条件。

In this project, the effect of increase in damage on the static and dynamic properties in CFRP laminates and the effect of a hole, as a macroscopic damage on natural frequencies of CFRP laminated plates with a hole are discussed. In addition the identification of boundary conditions in CFRP laminated plates by neural network is also studied. The results include :1. Since the change of damage area is proportional to the change of damping capacity in the CFRP laminates, it is possible to predict the damage area by the value of damping capacity.2. There are the correlations among the flexural strength, damage area and damping capacity. The lager damping capacity becomes, the smaller flexural strength is bacome.3. In the compression, the failure mode is depend on the value of damping capacity. The failure of bucking and the pure compression failure modes are obtained in small and large damping capacity.4. The good agreement between the numerical and the expertmental results of CFRP laminated plate with a hole shows that the natural frequencies and modes can be presicted by the proposed method in this project.5. The boundary conditions of CFRP laminated plates can be identified using by the input data of natural frequencies, mode values at a few points in the Neural network proposed here.