Damage quantification in truss structures by limited sensor-based surrogate model

Abstract

In this study, we have explored the structural damage detection of truss structures using the state-of-the-art deep learning techniques. The surrogate models, deep neural networks, are used to train the knowledge of the patterns in the response of the undamaged and the damaged structures. The limited sensors are then used to collect the response from the truss structures. Most previous studies on structural damage detection by using the conventional neural networks have been limited by the lack of a technique that determines an optimum learning rate in the training process. Recent advances in deep learning techniques can provide a more suitable solution to the problems and the process of feature engineering. A 31-bar planar truss is considered to show the capabilities of the deep learning techniques for identifying the single or multiple-structural damage. The frequency responses and the elasticity moduli of individual elements are used as input and output data sets, respectively. The results showed that, in all cases considered, the proposed surrogate model was possible to detect damaged states with very good accuracy.