Diagnosis, Based On The Crack Of The Vibration Characteristics Of Engineering Structures

Ship and ocean engineering structures mostly consist of plates and beams. Cracks may occur under the dynamic loading during the process of manufacture and operation. The undiscovered ongoing growth of cracks will lead to collapse of the whole structure, thus causing serious marine accidents. As cracks are mostly invisible to bare eyes, especially those in the underwater structures, it's essential to develop an effective way for detecting cracks.Based on the alterative information of vibration mode and natural frequency in the presence of cracks, the thesis focus on the study of diagnosis index of cracks for plate and intelligent diagnosis of cracks using BP neural network.The major works in this thesis are:The effect of the straight and skew cracks on the vibration properties are studied for the cantilever beam, clamped-clamped rectangular plate, simple-supported rectangular plate and triangular plate. The effect on the natural frequency and vibration mode shapes of the bracket in the presence of cracks is discussedBased on the available index of crack diagnosis, a new index, is proposed. The effect of the crack with differentposition and length on the different index is discussed. The proposed index is demonstrated better than other methods.BP neural network developed by Matlab is used to diagnose the position and length of the crack in the plate. The error of the position and the length between the predicted result and the assumed issues is acceptable. Furthermore,this application verifies that the index proposed in thisthesis is better than others.