General regression neural network model for predicting impermeability of concrete

Concrete permeability have a direct impact on its durability,and the tradition test consume large amount of time surely.According or-thogonal test data,the relationship between the chloride diffusion coefficient and six parameters on the impermeable performance was presented.Based on general regression neural network nonlinear analysis method,a intelligent information analysis of nonlinear general regression neural network method,an artificial intelligence information model for predicting impermeability of concrete was established.Result indicates that pro-jection pursuit regression model is reliable enough to concrete's virtual design.It provides a new way to predict impermeability of concrete.