| Structural finite element model updating has now become one of the hot research subjects in the field of structural health monitoring,which is a process of adjusting certain parameters of the finite element model to make the numerical calculation agree with the measurements collected from field monitoring.In this paper,a further study is carried out about how to make full use of the limite measurements and update the local or small parameter using overall responses in large complex structural model updating.The main research works are as follows:(1)A test platform of laboratory space truss structrue is established.The space truss structrue is designed with reference to a bailey steel truss,which is a scale structure with the purpose of model updating and damage identification.Static and dynamic characteristics of the truss is carried out to verify its reasonableness and feasibility using the ANSYS software package.The space truss structrue is made and assembled with steel bar and ball joint,and an integrated model experiment system is established.Based on the space truss structrue,static and dynamic test is conduct in standard state(without damage).A Benchmark finite element model is achieved with results of benchmark test by the technology of model updating,which lays the foundation of model updating and damage identification in the further study.Meawhile some preliminary studies on damage experiment of space truss structrue are also made.(2)Baesd on the space truss model and cable-stayed bridge model,some factors are selected to investigate the effect on model updating,such as the amount and sensitivity of design parameters,the amount and type of characteristic responses.The updating results show that the updating parameters can not be corrected effectively without enough characteristic responses and the updating error reduces with the increasing of characteristic responses.Whether to update successfully is related to the sensitivity of parameters and the type of characteristic responses using enough or redundant measurments.An identification index of finite element model updating based on enough or redundant measurments is put forward.The results indicate that the finite element can be updated successfully with the identification index less than or equal to 1.54.When the identification index is greater than or equal to 1.60,the finite element fails to update.Model updating based on experimental data collected from the space truss structure,cable-stayed bridge scale model and other reaschers is conduct to verify the validity of the identification index.(3)The method of adjusting sample size based on response surface is put forward with the purpose of updating the design parameters which have great difference in sensitivity.Firstly,The principles of the method are briefly described,and the processing flow is presented.The method is applied to correcting the design parameters without updated successfully using enough or redundant measurments in the third chapter.The results show that the error under each work condition reduces considerably,which validates the feasibility of the method.And then the influence of the adjustment coefficient on the model updating is investigated,and the method to choice adjustment coefficien is given.The results show that an appropriate adjustment coefficient should be selected,and too large or too small may still result in the failure of the model updating.when the identification index is less than or equal to 1.54 after adjusting a sample size with an appropriate adjustment coefficient,the finite element can be corrected successfully.The adjustment based on some sample sizes is more efficient to reduce identification index and gain a better result.Finally,model updating based on experimental data collected from the space truss structure and cable-stayed bridge scale model is conduct to verify the validity of this method. |
PDF Full Text Request