Font Size: a A A

Methods And Research Of Structural Damage Detection Based On Neural Network Technique

Posted on:2005-04-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q ZhangFull Text:PDF
GTID:2132360122491233Subject:Structural engineering
Abstract/Summary:PDF Full Text Request
Research of structure health monitoring is one of the hot issues in the field of civilengineering in recent years. The structural damage detection is the foundation and central coreof structure health monitoring. The technique of vibration modal analysis and neural network isavailably used to solve this difficult problem. This paper mainly focuses on researching anddeveloping the systems of structural damage detection that adapt to different styles of structure.The paper not only introduces the ambient excitation test and the apply in the direction ofstructural dynamical identity but also compares the effects of the traditional frequency-spectrum analysis and the time-frequency analysis based on wavelet analysis aided the aspect ofstructural dynamic data's acquisition. The paper researches and analyses the implementations ofthe damage localization and the extent of the damage accessed and puts forward the way of thedamage detection that used the technique of curvature mode shapes aided the aspect of thestructural damage identification. For illustrating correctness of index discussed above, damageson a cantilever beam are simulated. This paper respectively does the simulations of RC frame structure and highrise buildingstructure by the large-scale finite element analysis software SAP2000. There are 1236 sampleelements constituted by dint of the results of the simulation. On the one hand these indexesincluding natural frequency index, displacement mode shapes index, assembled index andcurvature mode shapes index are used to make damage detection and compared each other. Onthe other hand the applied effect of BP network, PNN network and LVQ network are comparedusing the same index. The paper brings forward a new grading method using different damageindex and different network in the different period of damage detection. The nonlinear dynamicresponse analysis is researched and analyzed to resolve the question of sample elementsaccording to the interstory drift. The seismic appraisal to ancient construction is researched combined the project titled"QUAN ZHOU seismic disaster prevention layout". The paper obtains the integrated methodthat combined the technique of ambient excitation test and finite element modeling.
Keywords/Search Tags:structure health monitoring, damage detection, network technique, time historyanalysis, wavelet analysis, ancient masonry pagoda
PDF Full Text Request
Related items