基于神经网络方法的高速列车车外气动噪声预测

摘要

The neural network method was used to predict exterior aerodynamic noise of high-speed trains. Based on Lighthill’s acoustic analogy theory, an aerodynamic noise computation model of the high-speed train was built. Then, a neural network model for aerodynamic noise prediction was built up using Levenberg-Marquardt (LM) algorithm. The prediction model was trained by the sample data of the external aerodynamic noise signal, and the trained neural network model was used to predict the external aerodynamic noise. The results show that the neural network method for aerodynamic noise prediction is a quite accurate algorithm and can be used for exterior aerodynamic noise prediction of high-speed trains.%利用神经网络进行高速列车车外气动噪声预测研究。基于Lighthill声学类比理论，建立高速列车气动噪声计算模型。在此基础上采用Levenberg-Marquardt (LM)算法建立车外气动噪声的神经网络预测模型，选取车外气动噪声样本点对预测模型进行训练，用训练好的神经网络预测模型预测车外气动噪声。结果表明，建立的神经网络模型对车外噪声具有较好的预测效果，可以用来进行高速列车车外噪声预测。