Identification Of Truss Structural Condition Based On Separated Damage

Some damage may occur in civil engineering structures due to environmental factors such as corrosion and aging during the service life.It is significantly important to accurately and effectively identify the location and extent of structural damage for the purpose of maintenance in late stage.Based on the concept of separated nonlinearity and Sherman-Morrison-Woodbury(SMW),a damage-separation method(DSM)is proposed in this study,which is able to rapidly compute structural response and further identify damage in truss structures.The main work of this thesis is as follows:Based on the separated nonlinearity and the formula of SMW,the DSM is proposed to quickly compute structural response.The principle of DSM is that the total strain in a cross section is divided into original strain and damage strain,and the material damage is evaluated from stiffness degeneration.A damage-separation governing equation is deduced by using both the principle of virtual work and the finite element theory,and further solved by SMW formula.The response of a damaged structure can be considered to be a linear superposition of responses caused by both structure itself and structural damage."Separation" means that the damage is separated in the model.The underlying mechanism of the method is that the stiffness matrix related to the damage model is isolated from the global stiffness matrix,and only a small-scale stiffness matrix representing local damage is updated and factorized during the calculation process.Thus,the efficiency of calculation is greatly improved to identify damage of structures.Two operation are applied to improve the optimization efficiency.The first one is to prejudge the damage area in advance based on experience,structural analysis or signal processing,and then the structure is calculated by blocking,which improves the efficiency to identify the local damage in truss structures.The second one is the sparse processing of the corresponding matrix,in which the amount of calculation is reduced and the global damage of structure can be effectively identified when the number of damaged elements is great than certain values compared with the total number of elements according to Saint-Venant’s principle.However,its accuracy is reduced.In addition,the proposed DSM is further used to identify the condition of truss structures.The damage is identified from both static DSM and dynamic DSM.The model updating method is used to identify the structure damage,and the objective function is established based on the structural response calculated by DSM and the measured structure response.The objective function is solved by iteration using sequential quadratic programming(SQP).When the objective function converges,the damage index α of the truss structure unit is solved,and the damage identification of the structure is realized.The DSM avoids the re-formation of the finite element model in each iteration,such that the updating and decomposition of structural stiffness matrix is avoided,and high computation efficiency can be achieved.Finally,case study was performed for a steel truss model considering different damage levels based on DSM.The accuracy and efficiency of this method are compared with conventional finite element method.Results show that the DSM can effectively identify the location and degree of structural damage.In addition,the structural damage identification based on static information has a high accuracy but contains less structural information,and the damage identification based on dynamic information contains more structural information but the computation efficiency is large and the accuracy is low.