A sparse, triangle-shaped sensor array for damage orientation and characterization of composite structures

Abstract

Abstract Since numerous sensors are needed to create a sensor array for the structural health monitoring of large-scale structures, the equipment quantity and cost considerably increase. This study proposes a sparse, triangle-shaped sensor array to identify, orient, and assess the degree of structural damage in composite constructions in order to overcome this shortcoming. The damage-scattered Lamb waves are recorded by the sparse sensor array with a variety of features that are then extracted and fed into the support vector machine (SVM) classification method. The location and severity of the damage in composite constructions can be determined by training the SVM model. The principal component analysis technique is used to compress the wave feature vectors while maintaining the majority of the damage information because the high dimension of the wave feature vectors required a significant amount of calculation during the training phase. Proof-of-concept tests show that the trained model, by utilizing the many properties of Lamb wave signals, can orient and define the degree of damage with excellent accuracy. Multiple Lamb wave properties can be used to make up for the triangle sensor array’s loss of damage information. In conjunction with the SVM, the triangle-shaped sensor array that was proposed in this study can efficiently make it easier to identify and characterize damage to large-scale structures while using fewer sensors.