基于SBAS-InSAR技术钢桁梁结构桥梁形变信息提取与分析

Extraction and analysis of deformation information of steel truss girder structure bridge based on SBAS-InSAR technology

针对传统桥梁监测技术监测时间不连续、只能依靠现有监测数据分析桥梁形变的问题,以南京大胜关长江大桥为研究对象,利用SBAS-InSAR技术获取2018—2019年的桥梁形变结果,以桥梁形变结果中的平均形变速率和LOS向时间序列形变量为输入集,分别构建BP神经网络时间序列模型来预测桥梁的形变量。结果表明:平均形变速率预测模型和LOS向时间序列形变预测模型的预测值与InSAR观测值之间的平均绝对误差分别为1.54、1.28 mm,均方误差分别为1.81、1.34 mm,均方根误差分别为1.81、1.53 mm,表明时间序列形变预测模型的可行性,为未来的桥梁形变预测提供了有力支撑。

To address the problem that the traditional bridge monitoring technology is not continuous in monitoring time and can only rely on the available monitoring data for bridge deformation analysis,this paper takes Nanjing Dashengguan Yangtze River Bridge as the research object.SBAS-InSAR technology was used to obtain the bridge deformation results from 2018 to 2019.BP neural network time series models were constructed to predict the deformation quantity of the bridge by taking the mean deformation rate and LOS time series deformation quantity of the bridge deformation results as input sets.The mean absolute errors between the predicted values and the InSAR values of the two models were 1.54 and 1.28 mm respectively,the mean squared errors were 1.81 and 1.34 mm,and the root mean squared errors were 1.81 and 1.53 mm.The results indicate the feasibility of the time series deformation prediction model,which provides strong support for future bridge deformation prediction.