Structural Damage Identification And Finite Element Model Updating Based On Distributed Long-Gauge Strain Sensing

Structural health monitoring(SHM)system is crucial for timely and effectively monitor and mastering the structure’s state and avoiding significant damage accidents during the service period of the structure.Among them,developing reliable damage detection methods and establishing an accurate finite element(FE)model is critical for the SHM system to achieve real-time monitoring and real-time evaluation of the structural state.This paper systematically studies the structural health monitoring technology based on distributed long-gauge strain sensing for simple and complex structures using our research group’s long-gauge FBG strain sensor.They are taking advantage of the distributed long-gauge strain sensing technology that can effectively fusion the structure of the whole and local information to realize the whole structural performance monitoring.Several damage detection methods for structural damage locating or degree assessment under random excitations are therefore proposed,and the long-gauge strain sensor placement optimization and the corresponding structural response reconstruction algorithm are presented,and the application of the distributed long-gauge strain monitoring system in model updating is improved.Therefore,this paper provides a theoretical basis and an effective means for structural damage identification based on the distributed long-gauge strain monitoring and supports the development of an optimal monitoring system and the theory of model updating.The main research contents are as follows:(1)A structural damage detection method based on distributed long-gauge strain sensing technology under multi-point excitation is proposed.Based on the previous modal macro strain theory under single-point excitation,the expression of macro strain frequency response function under multi-point excitation is deduced,and the modal macro strain vector of the structure is therefore extracted under wind loads.A new damage index based on the modal macro strain is proposed to identify the structure’s local damages,and the proposed detection results were compared with those of the previous method,which directly uses the modal macro strain as the damage index.The effectiveness and reliability of the proposed method for damage detection of flexible structures are verified by numerical simulation and experimental investigation.Moreover,the applicability of the proposed method for structural damage identification under single-point excitation and multi-point excitation is studied.(2)A quasi-static macro strain extraction and structural damage identification method based on distributed long-gauge strain sensing is put forward.The theory of adaptive noise complete empirical mode decomposition(CEEMDAN)to decompose the dynamic macro strain is derived,and the criterion of quasi-static macro strain is established.The accuracy of the quasi-static long gauge strain extraction method is verified based on the vehicle bridge dynamic model,and its robustness is further studied under different noise levels effects.A damage identification method using the quasi-static macro strain energy ratio with a period of time as the damage index is proposed,and the local damage is detected by comparing the index changes before and after structural damage;The effectiveness and robustness of the method are analyzed and demonstrated through numerical simulations and experimental studies.(3)A data fusion approach for structural damage identification based on distributed long gauge FBG strain sensing is studied.A two-stage damage identification method based on the macro strain flexibility matrix is firstly proposed for the regional monitoring system with long-gauge strain sensors and accelerometers.For this proposed method,its damage locating and damage quantification is divided into two steps to solve one by one.In the first stage,the aim of accurately detecting the potential structural damage locations is realized based on the macro strain flexibility difference index before and after damage.In the second stage,the degree of the damage element is accurately evaluated by the virtual static analysis method.In addition,a damage identification method based on Dempster-Shafer(D-S)evidence theory is proposed for the distributed long gauge strain monitoring system.This method can fuse the information of multiple damage indexes to judge the damage locations of the structure comprehensively.The reliability and practicability of the proposed damage identification methods are verified by experimental tests.(4)A optimized monitoring system of structural response reconstruction based on the optimal long gauge FBG strain sensor placement is presented.An optimal long-gauge strain sensor placement method using the macro strain flexibility to modify the effective independent method(EFI)is proposed to determine the priority installation position of the sensor.The minimum installation number of sensors is determined by setting the average value of the nondiagonal elements of the modal assurance criterion of modal macro strain as the objective function.A macro strain response reconstruction method based on the optimal layout is proposed,and the macro strain response at the unmeasured locations can be therefore reconstructed effectively.Also,this section demonstrates the feasibility of the dynamic displacement estimation through the measured and reconstructed macro strain response using the conjugate beam method.Both the reconstructed macro strains and estimated displacements are in good agreement with actual values in the time and frequency domain.The effectiveness and practicability of the above methods are verified by numerical simulation and experimental research.(5)A structural damage identification method based on long-gauge strain sensing and finite element(FE)model updating is proposed.Set frequencies and modal macro strain vector as the objective function,a systematic damage identification system is established,including parameter selection,establishment and updated of the initial FE model,updated the FE model under damage state and damage assessment.The feasibility and reliability of the proposed method for damage locating and assessment are verified by an actual steel beam under single damage and multiple damage conditions.Finally,the structural health monitoring system of Wanlong bridge based on distributed long gauge strain sensing technology is established,and the dynamic and quasi-static macro strain of Wanlong bridge is separated.Besides,the FE model of Wanlongshan bridge is updated by using the modal parameters extracted from the measured macro strain data.