单层网壳损伤识别理论与试验研究

Theoretical and experimental research on damage detection for single-layer lattice shell

结构损伤识别可以归结为结构损伤参数的模式识别问题.对结构响应信号进行小波包分解可以获得各频带的信号能量,将此特征向量作为输入,利用支持向量机强大的模式分类功能,可以实现结构的损伤识别.在环境振动下,对1/10比例的单层网壳模型进行损伤识别试验,将不同的杆件沿径向进行相应程度的截面切割用以模拟不同程度的损伤状态.对不同损伤情况的加速度样本进行三层小波包分解,以相应频带的信号能量作为输入建立支持向量机,利用支持向量机对未训练样本的信号能量进行损伤分类.试验结果表明该方法简便准确,验证了小波包和支持向量机方法用于损失识别的有效性.

The problem on structural damage detection can be viewed as the study on pattern recognition of damage parameters. The signals of structural responses can be decomposed by wavelet packet and the energy of different frequency bands is gained. The energy vector can be used as the input of support vector machine, which has the strong ability on pattern classification, and the damage detection could be realized. A damage detection test for a single layer bielliptical spherical lattice shell model with a scale of 1/10 is carried out under ambient vibration. The damage cases are constructed by incising the sections of different elements with three degrees. Acceleration signals are decomposed by three-layer wavelet packet, and energy vectors are produced and used for training and classification as the inputs of the support vector machine. The validity of this damage detection method is proved by the experimental results.