Structural Damage Detection Algorithm Based on Principal Component Indexes and Embedded on a Real Time Platform

This paper presents the main results obtained by using a structural damage detection algorithm based on Principal Component Analysis (PCA) and piezo-actuation principle. A known high frequency piezoactuated signal is applied on an analyzed structure in order to determine the base-line performance for the undamaged state (undamaged PCA model). Q-statistic and hottelingOs T2 indexes are computed by projecting time data onto the principal component space, and used to identify deviations of the current dynamical responses respect to the undamaged state. The algorithm was embedded in the Beaglebone Black Hardware (platform based on an ARM cortex A8 processor) and tested by using experimental data supplied by the CODALAB group. The obtained results indicate that it is possible to identify and locate structural faults for this kind of structures. Identification capability of the algorithm for 10 damages is tested by adding masses on the surface of an aircraft turbine blade at different positions.