A Study On Damage Identification For Bridge Structures

In recent years,with the development of computer technology,the research about bridge structural damage identification has also been rapidly developed,followed by a large number of theoretical analysis and rich application results.With the increasing number of large and complicated bridges,the application of bridge health monitoring system which installed on major projects is becoming increasingly widespread.After the installation of the system,the bridge health monitoring is started.Based on the enormous number of measured data,accurate and effective judgement and evaluation of the bridge can be estimated.Damage identification of bridge structures is based on the information collected from the health monitoring system.However,the theory and application of damage identification is lag than the design and installation of health monitoring system.From the point of the existing damage identification methods,most of them are focused on a certain type of simple bridge,if you want to apply the methods on the large complicated bridge structures,a lot of problems must be faced.Based on the current research status of structural damage identification,this paper analyzes and studies some problems involved in the damage identification of bridge structures,the details are as following.1)model updating of structural finite element model(FEM).2)damage identification of bridge structures based on flexibility matrix method.3)assessment of damage state of bridge structures.4)bolt loosening identification based on computer vision technology.In view of the above problems,the main contents of this paper include that:(1)The finite element model updating of bridge structure.The traditional structural matrix updating method has the problem that the updated structural matrix cannot be feedback to the existing general finite element analysis software.Based on Kriging model and swarm intelligent optimization method,this chapter proposed the finite element model updating method,which applied genetic algorithm(GA),bird mating optimization algorithm(BMO),particle swarm optimization algorithm(PSO)and Latin hypercube sampling to update the design parameters of the finite element model.Firstly,one truss structure was taken to verify the effectiveness of the methods.Then,take one large cable-stayed bridge as example,three methods were used to update the design parameters,and the updated results were compared and analyzed.The results shown that the method based on PSO has the highest accuracy,followed by BMO-based algorithm.GA-based algorithm took the least time,followed by BMO-based algorithm.(2)Damage detection of bridge structures based on flexibility matrix method.Among all kinds of structural damage location methods,the method based on flexibility matrix has been widely used and considered as one of the promising methods.Firstly,the uniform load surface parameters of the structure are established,and the curvature difference of the uniform load surface is taken as a damage index to identify the damage of the structure.Through one numerical example,the proposed damage identification method was verified.Secondly,based on the flexibility matrix,the damage locating vectors(DLVs)method proposed by Bernal is applied to do damage identification analysis.Based on the theory,and by using measured data,the method applied to one simple supported beam.(3)Assessment of damage state of bridge structures.Firstly,based on the fuzzy theory,the fuzzy membership function is introduced into the evaluation of the damage state of bridge structures,and the model of the damage state of the bridge structure is constructed.The simply supported beam bridge and cable-stayed bridge are selected as examples,and the various factors are expressed quantitatively in the form of normal distribution function,then the membership degree of each factor is calculated respectively.Third,the overall membership degree of the assembled structure is graded to judge the damage state of the structure.Based on the establishment of the model and the study of different types of bridges,two bridges are selected to carry out the verification of damage state assessment method.Compared with the results from Code for Maintenance of Highway Bridge and Culvers,the bridge damage state results obtained by adopting the method mentioned in the paper are consistent with those determined by the code.Secondly,based on the uniform load surface curvature difference method and fuzzy inference system,the method ULSCD-FIS was proposed to estimate the damage state of the structure.After calculated the ULSCD of the structure,the fuzzy rules were established,then inferenced the damage extent of the structure.By one simple supported beam,the proposed method was verified.(4)Bolt loosening identification based on computer vision technology.Firstly,based on computer vision technology,combined with Shi-Tomasi feature extraction method,KLT feature points tracking method,MLESAC estimated method,one bolt looseness locating method based on tracking feature point movement trajectory was proposed to effectively locate the loosened bolts of the structures.Secondly,after locate the loosened bolt,combined with the corresponding feature point tracking algorithm,feature point matching and geometric transformation matrix method,the method based on feature points trajectory tracking identification method was proposed to estimate the bolt looseness angles.At last,five groups of tests are set up to verify the proposed method.As the angle between the locked camera/smartphone and the component decreases,the effectiveness of the damage degree assessment algorithm decreases gradually.