基于BP神经网络的人防工程结构损伤诊断研究

摘要

为了有效地对地下人防工程结构进行损伤诊断,提出了以结构前三阶振型变化率绝对值之和作为损伤诊断标识量,结合BP神经网络技术对某浅埋地下人防工程结构进行损伤诊断.实例计算分析表明:通过计算给定人防工程结构的损伤诊断标识量,经过BP神经网络的学习、联想、记忆和分类,能较准确地识别浅埋地下人防工程结构的损伤位置和损伤程度,识别精度较高.%In order to solve the problem of damage diagnosis upon the civil defense engineering.The sum of the variation rate of the former three mede' absolute value as a signature for damage detection is proposed in the article, combined with the Backpropagation Neural Network to detect the damage of a fleet-covered civil defense engineering.According to the example it is shown that by giving the signature for damage,through the Backpropagafion Neural Network's self-study, association,memory and classifying,it is proved that the Backpropagafion Neural Network can detect the damage of the fleet buried civil air defense engineering more precisely.