Artificial Neural Network Analysis Method of Concrete Performance Based on Virtual Data Samples

摘要

The relationship between concrete performance and influence factors is uncertain and nonlinear. Accordingly, present BP neural network and virtual samples are presented to predict concrete performance in this paper. At first neural network and matters which need attention are introduced, And frost resistance forecasting model and impermeability model are built up, which are three-tier BP neural network of 6-13-2,4-9-1. The results show that the predicted values are ideal, and artificial neural network as one of the methods to forecast performance of concrete is appropriate.