Font Size: a A A

Numerical Analysis Of Structural Damage Identification Based On Improved Levenberg-Marquardt Algorithm

Posted on:2020-12-01Degree:MasterType:Thesis
Country:ChinaCandidate:Y ZhangFull Text:PDF
GTID:2392330602983136Subject:Road and Railway Engineering
Abstract/Summary:PDF Full Text Request
With the development of science and technology,infrastructure construction such as bridges,buildings,tunnels and underground projects is advancing rapidly,and the engineering structure has become increasingly complex and diversified.During the service period,different degrees of damage are usually generated.These problems seriously affect people's daily life safety.Accurate damage detection is one of the most concerned research topics in engineering and academia.The identification method of model updating based on time domain information has become a hot research direction of damage identification method because of its simple calculation process and easy analysis.However,due to the large amount of calculation and the difficulty of sensitivity matrix calculation,this kind of algorithm often fails to achieve the desired effect in the process of identifying structural damage.In view of the above problems,this paper uses L-M algorithm theory and Complex-variable-differentiation method to build the numerical analysis model of damage identification.By analyzing simple-supported beam,frame and tunnel lining model,the related factors such as identification rate and accuracy are discussed.The main contents are as follows:(1)A numerical model for damage identification is constructed.Based on the basic theories of L-M algorithm,improved L-M algorithm and Complex-variable-differentiation method,the numerical model of damage identification was constructed,and the corresponding numerical analysis program was written by MATLAB.By analyzing the numerical models of several typical civil engineering structures,the feasibility and applicability of the identification models are studied.(2)The identification results and influencing factors of various structural models are compared and analyzed.The identification accuracy of various engineering structures under mixed damages of different positions and degrees is analyzed comprehensively,and the influence of noise condition on the identification effect is discussed.In this paper,the effects of different forms of force on the identification of beams are compared and analyzed.The damage identification of beams and columns in the frame structure is carried out and the effect of changing the layout of measuring points on the identification accuracy of linings is studied.The feasibility of using unit set method to identify structural damage is explored.(3)Discuss the recognition effect of different improved algorithms in numerical model.For the identification rate,a variety of improved L-M algorithms are applied to the numerical damage identification analysis,and the influence of different noise conditions on the identification accuracy of various algorithms is discussed.By analyzing the recognition process of different models,the recognition rates and stability of various algorithms are calculated and compared.The numerical results show that the proposed identification model can be used for damage identification of several common engineering structures,and has a fast convergence rate and good noise resistance.The improved L-M algorithm can improve the identification effect to a certain extent and can be further applied to the damage identification analysis of practical engineering.
Keywords/Search Tags:Damage identification, Complex-variable-differentiation, Improved L-M algorithm, Tunnel structure
PDF Full Text Request
Related items