Ⅱ型主梁断面颤振导数的神经网络预测

Neural Network Prediction of Flutter Derivative of Type Ⅱ Girder Cross Section

为了提高桥梁颤振临界风速以及颤振导数在初步设计阶段的预估工作效率,本文在风洞试验和CFD计算的基础上,结合神经网络技术,建立一种基于神经网络的快速预测Ⅱ型断面颤振导数的方法。研究结果表明,预测结果具有高精度,与数值模拟结果相近。

In order to improve the critical wind velocity of the bridge flutter and the efficiency of the estimation of the flutter derivative in the preliminary stage of bridge design,based on wind tunnel test and CFD calculation,a neural network-based,fast prediction method of flutter derivative of typeⅡgirder cross section was established by combining neural network technology.The results show that the predicted results are of high accuracy and close to the numerical simulation results.