METHOD OF PREDICTING THE COMPRESSIVE STRENGTH OF CONCRETE

摘要

A method for predicting the compressive strength of concrete is provided to improve accuracy and efficiency of prediction by using different probability density function for each variable and determining the probability density function with automatically using the standard deviation for each variable. A method for predicting the compressive strength of concrete comprises the steps of: defining the combination design data of concrete as an input pattern of training pattern and the increasing concrete compressive strength as a class; calculating the standard deviation for each training pattern variable of concrete combination design data and calculating the probability density function F(X) of the probability nerve network by using a mathematical equation(1); and comparing the obtained probability density function with the class to determine the concrete compressive strength of a sample, wherein X_j is a jth input variable; W_i,j is a jth input variable of ith training pattern; b is a bias, sigma_j is the standard deviation of jth input of training pattern; and p is a dimension of the test pattern.