基于随机子空间和统计模型的输电塔损伤识别

Damage Detection of Transmission Tower Based on Stochastic Subspace and Statistic Model

针对输电塔结构,给出一种基于随机子空间结合统计模型的损伤诊断方法,利用输电塔振动响应信号挖掘数据统计特征并构建高灵敏损伤诊断指标实现输电塔结构局部构件的损伤检测。首先,获取输电塔结构的瞬态动力响应数据;其次,利用随机子空间法识别模态参数,构造名义模态参数,并定义一个与参数识别过程相关的残差向量表征结构动力参数变化;最后,计算该残差对结构参数的灵敏度,并构造残差灵敏度的χ^(2)统计量作为结构损伤诊断指标。通过数值模拟及现场试验对某足尺输电塔结构进行分析,结果表明,上述方法可有效对输电塔结构局部构件的螺栓松动、杆件破坏等损伤进行识别。

Aiming at transmission tower structure,a damage detection method based on statistical model combined stochastic subspace identification algorithm is given,where vibration responses of transmission tower are used to obtain statistical features and construct highly sensitive damage index in order to diagnosis damage of local components.Firstly,transient dynamic response signals are collected.Secondly,modal parameters are identified by stochastic subspace identification algorithm,nominal modal parameters are built accordingly and the residual vector is defined,which is related to the structural parameters variation.Finally,the sensitivity of the residual to structural parameters is calculated,and then the χ^(2) statistic value of the residual sensitivity is constructed as the damage detection index.Through numerical simulation and field test,a full-scale transmission tower Is analyzed,the results show that the proposed method can effectively identify local components damage of transmission tower,such as bolt loosening and structural member failure.