Structural Damage Detection Using Convolutional Neural Networks Based on Modal Strain Energy and Population of Structures

A convolutional neural network (CNN)-based structural damage detection (SDD) method using populations of structures and modal strain energy (MSE) is proposed. In this study, sufficient samples of the CNN are provided by numerical simulations, and the size of the model can be changed by modifying the coordinates of some nodes, thereby establishing a series of numerical models (i.e., a population). Finally, three groups are investigated, the effects of multiple indices on damage detection based on population are compared. The results demonstrate that the MSE as a damage index is superior to the other indices.

Automated summary

A CNN-based SDD method using populations of structures and MSE is proposed. Sufficient samples are provided by numerical simulations and a series of numerical models are established. The effects of multiple indices on damage detection based on population are compared and the results show that MSE is superior to other indices.