【摘要】 提出一种分步识别结构损伤模式的方法,首先根据结构损伤识别的特点,利用小波包分解原理,获得基于响应信号能量的损伤特征向量,然后利用神经网络确定损伤模糊集隶属度函数,通过模糊模式识别方法对结构损伤模式进行识别,最后对桥梁Benchmark模型进行试验验证.结果表明,该方法能较准确地识别结构的损伤模式,在工程应用中具有一定的实用价值.更多还原

【Abstract】 Based on the characteristic of damage identification for engineering structure,a method of damage pattern identification is put forward.First,using wavelet packet decomposition algorithm,damage eigenvector is extracted.Then,using neural network to determine damage fuzzy membership function,and then based on the fuzzy pattern recognition,the pattern of damage in structure is found.A calculation example of the bridge benchmark model proves that the method is much more accurate in identification and practical in engineering.更多还原