GNSS与加速度计融合超高建筑动态监测数据分析  

Reconstruction of super high-rise buildings dynamics using integrated GNSS and accelerometer

作　　者：吴绍诚 王怀宝[1] 王坚 WU Shaocheng;WANG Huaibao;WANG Jian(School of Surveying and Exploration Engineering,Jilin University of Civil Engineering and Architecture,Changchun 130118,China;School of Geomatics and Urban Spatial Informatics,Beijing University of Civil Engineering and Architecture,Beijing 102616,China)

机构地区：[1]吉林建筑大学测绘与勘查工程学院,长春130118 [2]北京建筑大学测绘与城市空间信息学院,北京102616

出　　处：《测绘科学》2023年第2期45-53,共9页Science of Surveying and Mapping

基　　金：国家自然科学基金面上项目(41874029);北京市自然科学基金项目(8222011)。

摘　　要：针对影响全球导航卫星系统(GNSS)变形监测精度的多路径误差,该文建立基于经验模态分解(EMD)算法的系统趋势分离模型,修正GNSS变形序列。构建多速率卡尔曼滤波Rauch-Tung-Striebel(RTS)平滑模型,融合超高层建筑的GNSS和加速度计监测数据,充分发挥两种传感器各自的优势。针对超高层建筑首次应用能量差值法确定变分模态分解的分量数,进而对分量进行频谱分析以提取超高层建筑的主模态振动频率。模拟数据表明,该文算法能够提高分析精度,融合位移的均方根为4.3 mm,相关系数为0.95,信噪比为12.66 dB。通过长春海容广场大厦采集的监测数据进一步验证得出,与单一传感器相比,该文算法能够提高位移数据的采样率,增加数据的完备性,削弱GNSS高频噪声的影响,提取到超高层建筑前两个主模态振动频率为0.19、0.28 Hz。Global navigation satellite system(GNSS)can monitor structure displacement,but the GNSS positioning result is typically noisy and infected by multipath errors.To address this issue,an empirical mode decomposition(EMD)based systematic error separation and noise filter model was developed and applied to GNSS time series in this paper.The Rauch-Tung-Striebel smooth based on multi-rate Kalman filter fuses high-rise GNSS and accelerometer data,complementing the respective advantages of the two sensors.Aiming at the deformation monitoring direction of super high-rise buildings,the energy difference formula was used for the first time to determine the number of components of variational mode decomposition,and then the frequency spectrum of the components was analyzed to extract the main modal vibration frequency of super high-rise buildings.The simulation data showed that the proposed algorithm could improve the accuracy,the RMSE of fusion displacement was 4.3 mm,the correlation coefficient was 0.95,and the signal-to-noise ratio was 12.66dB.The trial carried out on the Changchun Hairong Building was used to validate the proposed approach,it showed that compared with a single sensor,the proposed algorithm could improve the sampling rate of displacement data,increase the completeness of data,and weaken the influence of high-frequency GNSS noise.The extracted natural frequencies of Changchun Building were 0.19Hz and 0.28Hz.

关 键 词：多速率卡尔曼滤波 RTS平滑 变分模态分解 数据融合 

分 类 号：P228.4[天文地球—大地测量学与测量工程] TU198[天文地球—测绘科学与技术]