Damage Detection and Identification

Marine, aerospace, ground and civil structures can receive unexpected loading thatmay compromise integrity during their life span. Therefore, improvement in detectingdamage can save revenue and lives depending upon the application. The prognosticcapability is usually a function of the examiner’s experience, background and datacollection during the evaluation. Nondestructive evaluation (NDE) methods are variedand specific to a given type of system (material, damage type, loading and environmentalscenarios). As a result, one method of damage detection alone cannot examine allpossible conditions and may even give false readings. In other words, by using morethan one NDE technique, the probability of ensuring a more accurate detection increases.This work examined various existing NDE techniques to assess damage in sandwichcomposites structures including: vibration modal analysis, transient thermal response, andacoustic emission. Sandwich composites consisting of two carbon fiber/epoxy matrixface sheets laminated onto a urethane foam core were experimentally and analyticallycharacterized using vibration, and thermal response to detect the presence of varioustypes of damages. A neural network (NN) approach that uses vibration and thermalsignatures to determine the condition of a composite sandwich structure is purposed. Thedata used to train a probabilistic neural network (PNN) were provided by numericalsimulations. Literature offers substantial evidence of the validity of each of the chosendamage detection schemes separately. However, we will show that these methods canwork jointly to complement each other in detecting the state of a sandwich compositestructure.