comparison study on neural networks in damage detection of steel truss bridge

this paper presents the application of three main artificial neural networks (anns) in damage detection of steel bridges. this method has the ability to indicate damage in structural elements due to a localized change of stiffness called damage zone. the changes in structural response is used to identify the states of structural damage. to circumvent the difficulty arising from the non-linear nature of the inverse problem, three neural networks, multi-layer perceptron neural network (mlpnn), radial basis function neural network (rbfnn) and general regression neural network (grnn) are employed to simulate damage states of steel bridges. it was observed that the performance of all three networks is well and they have good agreement with actual results performed with finite element analysis. the efficiency of grnn in structural identification is so good, although rbfnn has results close to grnn and mlpnn results are satisfactory. all networks have good results while there is a little damage in structural members. generally, results would have more error when damages in structural members extend. the engineering importance of the whole exercise can be appreciated once we realize that the measured input at only a few locations in the structure is needed in the identification process using neural networks.