Multi-model Structural Identification Theory And Experiment Research Based On Probability Statistics

Mode updating and parameter identification belongs to the category of structural identification(St-Id),which is conducted by pairing the structural finite element(FE)model with the static and dynamic test results,it is significant to evaluate its response prediction analysis,working conditions and health monitoring.This paper first introduces the concept of the multi-model structural identification(MM St-Id),which uses multiple FE models to anticipate the structural test results,the key is to form a large number of FE model libraries and check assessment,and select the correct models from the model cluster,rather than select the optimized FE model for simple model St-Id identification.In this paper,the issue related to MM St-Id has been researched based on Bayesian theory.Major contents of this research are drawn as follows:(1)The concept and basic framework of MM St-Id method based on Bayesian theory were introduced,the Markov chain-Monte Carlo simulation(MCMC)was utilized to build FE model libraries.(2)Since MCMC is not easy to converge and it has low calculation efficiency when the parameters have high dimensions,three improved MCMC sampling method for MM St-Id are introduced.The Matlab-Strand7 Application Programming Interface(API)strategy can be used to update the parameters of large structural FE model automatically.After the calibrated FE model libraries were established,they can be used to predict the responses based on the posterior probability distribution of the FE models.(3)In order to verify the feasibility and effectiveness of the proposed theory,a numerical example of a simply-supported beam were researched by the proposed improved MH sampling method,the TMCMC sampling method,Latin hyper-cube sampling(LHS)based on Bayesian theory and response prediction,and a simple model St-Id method-genetic algorithm(GA)was used to compare.(4)Using a reinforced concrete composite slab test and an on-site large concrete-steel tubular truss arch bridge and verify that the recognition method based on Bayesian theory and response prediction by an improved MH sampling method,TMCMC sampling method,LHS and error threshold method.It proves that they are effective and scientific in the actual test and engineering application,and a simple model St-Id method-genetic algorithm(GA)was used to compare.The results show that the proposed MM St-Id method is much better in St-Id,structural response prediction and model selection.