Research on the damage detection in the structure

在之前的研究中，我们提出了利用图像处理技术识别两种悬臂梁有限元模型参数（即密度和弹性模量）的新技术。其中一根是组合梁，另一根是有小裂缝的破损梁。采用对称样条平滑和加权Tikhonov正则化技术对组合梁的一小段参数进行识别。此外，为了检测小裂纹的大小和位置，提出了一种新型的模拟退火方法。在本研究中，我们开发了一种检测二维结构损伤的技术，并减少了SA方法的计算时间。我们得到以下结果。(1)采用动刚度矩阵代替CM型有限元模型，大大缩短了计算时间。(2)采用上述方法，可以以较少的计算量精确地估计出小裂纹的位置和大小。(3)提出用半转动刚度系数作为表征结构中节点破坏破坏程度的指标。在消除图像数据中噪声的影响后，估计出宝贵的半旋转刚度系数。

In our previous researches, we proposed new techniques to identify parameters of a finite element model, i.e. the density and the modulus of elasticity, of two types of cantilever beam, by using an image processing technique. One beam was a combined beam, the other a damaged beam with a small crack. To identify the parameters of a small segment of the combined beam, spline smoothing with symmetrization and the weighted Tikhonov regularization technique were used. In addition, to detect the size and the location of a small crack, a new type of Simulated Annealing method was proposed.In this research, we develop our technique to detect a damage in the two dimensional structure and to reduce the calculation time in the SA method.We get following results.(1) Instead of using the CM type FEM model, we reduced the calculation time by using the dynamic stiffness matrix dramatically.(2) Using above technique, we can precisely estimate the location and magnitude of a small crack with less calculation.(3) The semi-rotational stiffness coefficient was proposed as the index to represent the damage of the joint failure in the structure. After reducing the effect of noise in the image data, the precious semi-rotational stiffness coefficient was estimated.