Damage detection based on pattern-matching method of response signal

Using response data as the feature to extract the damage, this study proposes an analytical method to detect damage by matching the response pattern between a damage-expected beam system and the database. Response pattern database is composed of the response variation before and after the damage, prescribed by various damage scenarios. Pattern-matching is evaluated by the correlation coefficients between two data sets and the damage corresponds to high correlation coefficient. Response variations at a full set of DOFs are estimated by the expansion of a few measured response data. Numerical applications illustrate the validity and applicability of the proposed method.