201 Crack identification using Mahalanobis discriminate analysis : 2^<nd> Report, Improvement of identification using multiple sensors

The crack identification in a structure is the most important problem in securing safety. We have succeeded in identifying the position and the length of the crack in a plate using natural frequency changes and the Mahalanobis distance method. Mahalanobis distance method is a useful way of determining similarity of an unknown sample set to a known one. In this paper, to improve the identification performance, we employ both higher order natural frequencies and multiple sensors and then identify the crack by using the Mahalanobis distance method. We demonstrate that the presented method is very useful through experiments.