Damage Detection in Structures Based on Feature‐Sensitive Neural Networks

Abstract

Detection of damage in structural systems is formulated as an inverse problem and solved by a new approach utilizing neural networks. Damage is modeled through reduction in the stiffness of structural elements, and manifests itself in the form of variations in observable static displacements under prescribed loads. A modified counterpropagation neural network is used to develop the inverse mapping between a vector of the stiffness of individual structural elements and the vector of the global static displacements under a testing load. It is shown that the network functions as an associative memory device capable of satisfactory diagnostics even in the presence of noisy or incomplete measurements. Numerical examples involving frame and truss structures show that the network approximations are fully acceptable from a practical standpoint.