Structural Damage Identification Based on Support Vector Machine and Relative Change Value of Modal Flexibility

Support vector machine (SVM) as a new method of statistic theory is introduced to identify the structure damage in the paper. The relative variety of modal flexibility is the sensitive identification parameter of structural damage and is input as the characteristic parameter to train SVM network. The results of the numerical simulation and experiment indicate that the method of SVM combined with the relative change value index of modal flexibility can reflect the character of structural damages. On a certain noise level, SVM has good classified identification performance and the results of damage identification are very accurate. It shows that the method has a good prospect in structure damage diagnosis field.