Research On The Dynanic Fingerprint Methodology In Damage Detection Of Prestressed Concrete Beam

Due to its energy consumption, effects of long-term loading, chemical erosion, firedamage, natural disasters, man-made damage and other factors, building structures areeasily affected by different degree of damage. Establishing and improving relevantearly-warning mechanism can minimize structural damages and even destructions asto protect people’s lives and property. Structural damage detection is the key to theearly-warning mechanism, and how to accurately test building structure for damageidentification, damage localization and quantification of the degree of injury is alsoone of the focus in the world. Above the background, this article carries out dynamictests of self-compacting prestressed concrete beams,11simply-supported and4free,then conducts simulating modal analysis by ABAQUS and kinetic theory analysis.Via dynanic fingerprint methodology studies, it shows that:(1) Dynamic tests of free beams are less affected by the external environment, andthe vibration curves are smooth. In comparison, simply-supported beams are greatlyinfluenced, of which modes have much more mutations, particularly it is difficult tochoose poles of LMS modal analysis.(2) Via comprehensive analyses of the first four natural frequencies and modes, thesimple supports can not achieve the ideal confinement effect for dynamic tests.Especially the tested high frequencies and modes are obviously different fromcorresponding theoretical values, instead the same as elastic-supported beams. Thesupports plays the role of elastic ones. The conclusions will provide meaningfulreferences to further researches on prestressed concrete beams such as improving theaccuracy of tests, dynamic analyses and theoretical frequency calculation.(3) It is clarified that natural frequency, modal assurance and modal stiffness canreflect the real faulted condition of prestressed beams, manifested through relevantsimulation analysis. The BP and PNN neural net approach are proposed, one withthird and fourth-order frequencies and modal assurance as damage indicators, of which the results show that it is appropriate for damage identification of the testedbeams; the other with third and fouth modal stiffness rate of change as indexes,indicating that it also can identify damage effectively.(4) It is expounded, employing simulation analysis, that the variation rate and theratio of mode, DSI index, and curvature modal are of great significance to damagelocation of prestressed beams. Also it is found that the method to orientate damageusing low-level modal is superior to higher one, and the variation rate and the ratio ofmode, DSI index are basically the same, but curvature modal works better than thefirst three dynamic signatures. In addition, the results of using the above methods toidentify defects of tested free-free beams well, show that it was hard to utilize eachapproach to locate damage accurately and effectively, which needs to be furtherimproved and perfected.