Wind Load Time-History Simulation Study Based on Neural Network

Abstract

In this paper, on the basis of the existing research and Matlab, scientific computing software, and through the characteristics of the artificial neural network theory, it presents artificial simulation of wind loads which are set up by BP neural network, and it mainly consists of the preparation of sample set, the certainty of network layer and hidden nodes, and the selection of activation functions and training methods. Based on Matlab and the neural network toolbox, it carries out the constriction, training and testing of neural network. According to the simulation of five different positions of horizontal wind speed time, it proves that the artificial neural network term wind loads simulation models are effective in practical application.