论文部分内容阅读
针对大跨度空间网格结构健康监测系统中的传感器布置以及损伤识别问题,以凯威特型单层球面网壳为例,首先,定义基于变形能的适应度函数,采用粒子群优化算法布置加速度传感器;其次,根据加速度响应信号,建立时间序列模型(AR模型),根据模型系数的变化,判断损伤存在与否;最后,以AR模型系数为输入,损伤位置为输出,建立BP神经网络,通过BP网络的训练和测试,判定损伤存在的位置.数值模拟结果表明:基于粒子群算法的传感器优化布置方法能够准确获取网壳结构中的关键信息点,有效节省传感器布置数目;基于时间序列分析和神经网络的损伤识别方法可以准确识别网壳结构的损伤及损伤位置.
In order to solve the problem of sensor placement and damage identification in large-span space grid health monitoring system, a new type of Kweite-type single-layer reticulated dome is used. Firstly, the fitness function based on deformation energy is defined and the particle swarm optimization Secondly, the time series model (AR model) is established based on the acceleration response signal, and the presence or absence of the damage is judged according to the change of the model coefficient. Finally, the AR model coefficient is taken as the input, the damage location is output, BP neural network is established, BP network training and testing to determine the location of damage.Numerical simulation results show that the particle swarm optimization based sensor placement method can accurately obtain the key information points in the reticulated shell structure and save the number of sensor arrangements effectively.Based on the time series analysis and Neural network damage identification method can accurately identify the damage of the reticular shell structure and damage location.