Investigation on damage self-diagnosis of fiber smart structures based on LS-SVM

The self-diagnosis function is one of main research contents of smart structures. And it is the foundation of other functions realization of smart structures. Aiming at the localization of present structural damage detection methods and the virtue of Least Square Support Vector Machine arithmetic, Least Square Support Vector Machine (LS-SVM) used to detect damages in fiber smart structures was proposed in this paper and was compared with the improved BP neural network. The experimental research results show that this proposed method is feasible and effective for detecting damages in smart structures. Least Square Support Vector Machine provides the more advanced method for realizing the self-diagnosis function in fiber smart structures.