In order to improve the efficiency of the prediction of the flutter derivative and the critical flutter wind speed in the preliminary design stage of the bridge, this paper established a fast prediction method for the flutter derivative of the Π-type bridge section. Based on the two-dimensional rigid segment model wind tunnel test, combined with the CFD calculation that using RMS model, two geometric parameters that are aspect ratio and size factor on the flutter derivatives were studied and the database for flutter derivatives of the Π-type bridge section was set. Furthermore, based on this database, the neural network technology was introduced to predict flutter derivatives of the Π-type bridge section, and the prediction results have reached a similar precision with the numerical simulation. Key words: wind resistance of bridge; fast calculation; neural network; flutter derivative

CICTP