| The results of the structural damage identification based on vibration have obvious uncertainties because of the influences of unknown excitation, measurement noise and modeling error. In this paper, the vibration transmissibility function and statistical hypothesis testing are utilized to identify the structural damage. The main contents are as follows:(1) In this paper, the research background and significance of structural damage identification are introduced. The development and current research of structural damage identification based on vibration are summarized.(2) The basic theory of the transmissibility function, principal component analysis(PCA) and the statistical hypothesis testing are presented.(3) Structural damage identification based on the vibration transmissibility function and t hypothesis testing. Firstly, the vibration transmissibility function before and after the structural damage is divided into groups, and which are used to construct vibration transmissibility functions matrix. Then, the PCA is used to compress the transmissibility function matrix and extract the principal components, the Mahalanobis distances of the principal component before and after the structural damage are calculated, which are used as the samples of the statistical analysis. Finally, t hypothesis testing is employed to identify the damage with the constructed statistics. Results of the numerical simulation and the shaking table model experiment of offshore platform show that the presented method is effective.(4) Structural damage identification based on the vibration transmissibility function and multivariate statistical hypothesis testing. The first principal component of vibration transmissibility function before and after the structural damage is extracted with PCA, which is treated as a sample of normal population, and then the F-test of multivariate statistical analysis is employed to identify damage with the constructed statistics. The results of numerical simulation of an offshore platform structure and the shaking table model test show that the proposed method is feasible.(5) The results of the proposed two methods are compared, some conclusions are got. |