论文部分内容阅读
本论文主要对基于神经网络技术的结构损伤检测理论进行了研究,并确定了结构损伤的位置与程度。对一个选悬臂板模型的损伤状况进行了数值模拟分析,并采取合适的方法构造改进型BP神经网络的输入参数,应用训练后的神经网络对结构进行了损伤检测。结果表明,该方法有效,可靠。
In this dissertation, the theory of structural damage detection based on neural network technology is studied and the position and degree of structural damage are determined. The damage condition of a cantilever plate model was numerically simulated and the input parameters of the improved BP neural network were constructed by a suitable method. The damaged neural network was used to detect the structure damage. The results show that the method is effective and reliable.