Structural Damage Identification Based on the Transmissibility Function and Support Vector Machine

摘要: A novel damage identification method based on transmissibility function and support vector machine is proposed outlined in this paper. Basically, the calculated with acceleration responses from damaged structure. Then two features, namely, wavelet packet energy low order principal components, are constructed by analyzing amplitude of decomposition component analysis separately. Finally, classification algorithm regression employed to identify location severity respectively. The numerical simulation shaking table model test an offshore platform under white noise excitation conducted verify method. results show that does not need information data undamaged structure, needs only small size samples, has certain antinoise ability. detection accuracy feature superior decomposition.