Prediction of wind-induced pressures on long-span roofs with complex shape using artificial neural networks

Abstract

The backpropagation neural network(BPNN) and the radial basis function neural network(RBF) are widely employed to simulate many kinds of nonlinear relationships, and have received increasing interests in recent years. This paper is concerned with the above two artificial neural networks for the prediction of mean wind-induced pressures of two long-span roof structures, the Shenzhen Citizen Center(SCC) and the Guangzhou International Exhibition Center(GIEC). In this study, simultaneous pressure measurements are made on two long-span roof structure models in a boundary layer wind tunnel and parts of the model test data are used as the training sets for the two ANN models to recognize the input-output patterns. Comparisons of the prediction results by the two ANN approaches and those from the wind tunnel test are made to examine the performance of the two ANN models, which demonstrates that the two ANN approaches can successfully predict the pressures on the corner surfaces of the long-span roof structure except the approaching wind direction, and if more experimental data are involved in network training, the network may provide more accurate prediction. It is also indicated in this paper that both the BPNN and the RBF prediction illustrate better performance in an interpolation than that in an extrapolation case.