This paper explores the potential of using a Multilayer Perceptron (MLP) network trained with the Back-propagation algorithm for damage assessment of a free-free cracked straight steel beam based on vibration measurements. The problem of damage assessment, i.e. detecting, locating and quantifying a damage, is essentially a pattern recognition problem. Since artificial neural networks are proving to be an effective tool for pattern recognition the basic idea is to train a neural network in order to recognize the behaviour of the damaged as well as the undamaged structure. Subjecting this trained neural network to information from vibration tests should imply information about damages states, locations and sizes. The inputs to the network are estimates of the relative changes of the lowest five bending natural frequencies due to damage. During the training these estimates are obtained by an FEM of the beam. A damage in the beam is modelled by a fracture mechanical model. The basic idea of this model is to model the crack zone of a beam by means of a local flexibility matrix found from fracture mechanics. The utility of the neural network approach is demonstrated by a simulation study as well as laboratory tests. The results show that a neural network trained with simulated data is capable for detecting location and size of a damage in a free-free beam when the network is subjected to experimental data.

purchase the full-text of this paper (price £20)

go to the next paper
return to the table of contents
return to the book description
purchase this book (price £34 +P&P)