Research On Shape Sensing Technology Of Engineering Structures Based On Inverse Finite Element Method

Due to external loads and environmental conditions,civil infrastructures are usually subjected to loss of stiffness,resulting in increased deformation.As an important structural health monitoring(SHM)procedure,deformation monitoring not only guarantees structural safety during the construction and service process but also provides scientific references for further design and maintenance.Therefore,it is of great significance to develop an effective technology for achieving full-field shape reconstruction of civil infrastructures.Just relying on discrete strain measures,the inverse finite element method(i FEM)enables deformed structuralshape reconstruction by simple matrix-vector multiplication,without any material properties and loading information.The i FEM methodology is well-suited for complex structures under complicated boundary conditions because the structural domain is discretized using appropriate‘‘inverse” finite elements.The i FEM framework is based on a least-squares variational principle,determining its computational efficiency,high accuracy and robustness for real-time applications.These exclusive features make i FEM much more powerful than the other shape reconstruction algorithms.However,very few applications of the i FEM to the deformation monitoring of civil structures are present in the literature.Two challenges limit the i FEM applications:(1)i FEM methodology is only applicable to isotropic homogeneous materials,which generally mismatches with material properties used in civil buildings.(2)The non-compatibility between the actual geometry model and the inverse finite element(FE)model may be induced since the structure geometry is updated continuously during construction.Moreover,when the i FEM analysis is used for complex structures,it is inevitable to adopt assumptions of node and boundary to simplify actual structures,increasing the inherent errors of the i FEM solution.To solve the above problems,this dissertation develops a novel fiber-optic shape sensor(FOSS)based on i FEM methodology.The FOSS can estimate the deformation of a structure with different material properties,which enhances the applicability and practical utility of the i FEM methodology.The boundary conditions and node characteristics for the FOSS are optimized to increase the accuracy of i FEM solution.Structural deformation will lead to shape variation in the FOSS.Distributed strain measures along the FOSS span are attained leveraging on surface-mounted fiber Bragg grating(FBG)sensors and then input into the i FEM algorithm to regenerate full-field displacement fields.This dissertation focuses on the development and application of the FOSS.The following are the main studies.(1)Planar and spatial inverse beam elements,i Beam3 and i EBT2,are proposed based on a least-squares variational principle utilizing classical beam theory as its underlying design theory.Numerical simulations and experimental tests regarding beam-like structures with different boundary conditions have been performed and demonstrated the excellent predictive capability of inverse elements.The effects of different factors(such as the discretization of the geometry)are assessed with respect to the solution accuracy.It has been demonstrated that lowfidelity mesh combined with optimal strain-sensor configuration enables superior deformation reconstruction while reducing the monitoring costs.(2)Then,this dissertation develops a temperature-insensitive fiber-optic shape sensor(FOSS)discretized using i Beam3/i EBT2 elements.The FOSS achieves two-and threedimensional bending deformation reconstruction of the structure.To improve the accuracy of the FOSS solution,an optimal sensor placement(OSP)model is established based on the leastsquares(LS)criterion,and its performance has been evaluated by a static loading test.A series of tests,involving FBG calibration tests,FOSS calibration test,and corresponding temperature characteristic validation test,have been performed and demonstrated the good working performance of FOSS and its components.(3)Finally,this dissertation establishes a FOSS-based deformation monitoring system.This system employs a complex multi-sensor interrogator and Lab VIEW programming environment as its hardware and software platforms respectively.The self-designed system is used for monitoring deformation of steel piles and bridges.The FOSS has the ability to regenerate both static and dynamic displacement fields of a pile and the predicted tip displacements agree well with those measured directly from the laser displacement sensor.Moreover,the FOSS enables multi-point deformation estimation of a bridge,and the monitored results serve to assess the health condition of the structure.