earticle

영어

The purpose of this study is to present a model that can predict the appropriate level of input value, divided into 40MPa and 60MPa high fluidity concrete by applying neural network theory, confirm the mix design model through test validation, and is expected to minimize material waste and quality variation. Neural network theory was used as a method in this study. Algorithms for the neural network theory was selected as multilayer perceptrons and used back propagation algorithms to increase accuracy and reduce error rates while training. The factors in the neural network theory affect the compressive strength as input value, and collects the mixed design data through references related to the compressive strength characteristics. To validate the model, a compressive strength test was conducted based on the prediction of the mix design factor, comparing between the target compressive strength and the test value in mix design. According to the results of the mix design verification experiment, the lowest error rate in 40MPa high fluidity concrete was 4.3% composed of fly ash and silica fume and in 60MPa high fluidity concrete the lowest error rate was 5.8% composed of silica fume and ground granulated blast slag. Overall, 60MPa high fluidity concrete was higher than 40MPa high fluidity concrete.