Theoretical And Experimental Reasearch On Damage Detection Of Latticed Shells Considering Random Defects

Because of the material properties and construction errors,spatial structure has theinitial random defects, which led to the geometric characteristics of the structure,mechanism and load carrying capacity of the damage evolution differ with designmodel, result in the structure unsafe, and prone to injury. The problem, to ensure thenormal construction and use of the structure and also give accurate judgment tohidden danger timely,by accurate modal analysis technology and the rapid andaccurate identification algorithm, to extracte small damage information from thedynamic response of the structure and random defects with normal initial phasedifference, is needed to be solved.Firstly, the article count statistical characteristics information of the spatialstructure, and summarize the distribution,mean,coefficient of variation, and thevariance of the five parameters, elastic modulus, diameter, thickness, Poisson’s ratioand the initial coordinates of the defect, for further studylaid the foundation. In orderto solve the problem of calculating the large number of samples considering therandom case, find software connectiong to carry out the large number of structuralcalculations,by using mathematical software to control the finite element software.Use the stochastic finite element method based on the Monte-Carlo variancereduction techniques to achieve random defects in the selection of parameters andfinite element analysis, and carry out new sensitivity analysis on the a frequency closespatial structure with an initial random defects, to identify sensitive parameters of thestructure and weak areas.Substantially overlap of the corresponding frequencydistribution exists when considering the case of random distribution structure, it isalmost impossible to identify damage using the frequency directly, simultaneouslyinvestigating the data of the vibration type, there is less level of overlapping dataphenomenon and its own characteristics of different rod damage modal case.Invstigate the modal analysis based on Zernike moments on this basis. Thedeterministic structure analysi once again proves that Zernike moments is a goodindicator of damage identification, but the methods is unable to achieve goodrecognition accuracy when the case is considering random variety of identification,mainly due to data overlap,the recognition classifier confusion, and judgmenterrors.Finally, based on experimental tests, a certain number of information representativeof a large number of samples, the experiment procedure accurately identify themodal information.The modal analysis of the information obtains a similar analysiswith the simulation results, the overlap of real data still exists and has an importantimpact damage recognition accuracy.This project invstigates structural damage identification when taking into accountthe structural randomness, compared it with deterministic structure damage identification, and carries out the experiment of randomness structure, and hasimportant theoretical and practical significance to promote research and application ofstructural health monitoring.