Research On Reconstruction Method Of Overall Deformation Monitoring Of Prefabricated Subway Station Section Based On Inertial Measurement

With the increasingly prominent advantages of prefabricated buildings,many prefabricated components have been used in subway construction.However,the long-term ground load compression and environmental erosion lead to various imperceptible deformation of the overall structure of the prefabricated subway station,resulting in potential safety hazards for train operation.Therefore,the long-term deformation monitoring of the overall structure of the prefabricated subway station must be carried out.The traditional monitoring technology uses only the limited points arranged according to experience for monitoring,and the correlation between the monitoring points is not strong,which can not meet the needs of long-term deformation monitoring of the whole structure.Therefore,this thesis introduces the compressed sensing theory into the problem of structural deformation reconstruction at finite points,and proposes a long-term deformation reconstruction method of large underground space structures based on compressed sensing.In the process of using sparse matrix,this method takes into account the type characteristics of the overall deformation signal of the prefabricated subway station,so the advantage of structural deformation reconstruction using finite points is obvious.The research results are as follows:（1）According to the actual structural characteristics,the finite element numerical model of the assembled subway station is established to determine the signal analysis form of the overall structural deformation,and meet the key requirements in compressed sensing:the signal is sparse represented based on different dictionaries,and it is clear that the original signal can be sparse and maximized according to the DFT sparse matrix;Design the corresponding observation system,compare different reconstruction methods,and determine the combination form of unit observation matrix and OMP reconstruction method,which can better restore the overall structural deformation curve.（2）The evaluation indexes of signal reconstruction are determined as compression ratio,reconstructed signal-to-noise ratio,similarity and signal Frobenius norm.The number and location of limited observations are fixed through multiple experiments,and finally the original and complete structural deformation signal is reconstructed,and the research on finite point deformation reconstruction under different loads is continued,The proposed method is verified by using the simulation data under symmetrical load conditions and typical conditions.The reconstruction effects of X,y components and two-dimensional contour signals are evaluated and analyzed respectively.When 25 fixed finite points are determined,the reconstructed signal-to-noise ratio is more than 9,the similarity is more than 90%,and the signal Frobenius norm is 1.03×10^-3 mm,the finite point deformation reconstruction method based on compressive sensing can effectively restore the structural deformation trend.（3）During the model test,the layout positions of 12 sensors are determined and the experimental data are collected.The actual monitoring data are transformed by using two mathematical models suitable for the structural deformation,and then the overall deformation curve of the structure is reconstructed.Finally,the reconstruction effect is compared and evaluated through the simulation results of the overall deformation curve in the finite element model and the deformation data at five marked points in the high-speed camera.The results show that the reconstruction effect of the arc model is better than that of the linear model,and the structural deformation trend can be effectively recovered in the process of model test.