Bridge Damage Identification Based On Dynamic Response And Sparse Regularization

Solving structural damage equation is a typical ill-posed problem so that is the regularization means are usually implemented.As L1 regularization method can make a good use of the sparsity of structural damage equations,it has been widely used in the field of bridge monitoring and structural detection.This method usually adopts the modal parameters such as frequencies and mode shapes.However,in the actual test,the order number of frequency that can be measured is limited and the frequency is known not sensitive enough to the damage,while the identified mode shape accuracy is relative low.All these issues seriously affect the effectiveness of structural damage detection.In this thesis,a structural damage detection method based on finite element model updating and L1 regularization is proposed by using the structural response induced by moving load.The proposed method is herein verified by both numerical examples and laboratory experiments.The main work and conclusions of this thesis are as follows:(1)The bridge finite element model(FEM)to calculate the moving load induced responses is established.Numerical examples are provided to compare the frequency change and the area change of moving load induced displacement-time curve caused by the local damage.The results indicate that the area change of moving load induced displacement-time curve is larger than change in frequency,and it is more suitable to be the damage feature for bridge damage identification.(2)A damage identification method based on FEM updating and L1 regularization is presented.Firstly,the elemental stiffness reduction factor in the structural FEM is taken as the parameter to be updated.Then the minimum difference between the area of moving load induced displacement-time curve calculated by the FEM and measurements in the laboratory experiments is taken as the objective function.Finally the L1 regularization optimization is employed to solve the objective function and to identify the structural damages.Numerical examples of simply supported beams and continuous beams with various damage conditions are used to verify the effectiveness of the proposed method.Besides,the effects of the number of measurement points,noise,and velocity of the moving load on the damage identification results are studied.(3)The experimental model of a moving vehicle and a simply supported beam are set up in the laboratory for moving load tests.The dynamic responses are obtained by sensors installed on the beam.Then the area of moving load induced displacement-time curve is implemented to verify the effeteness of the proposed method.The experimental results have shown that the proposed damage detection method in this thesis can effective to identify the location and severity of single damage,but its effectiveness decreases with the increase in the number of damage.