High strength concrete modeling by artificial neural networks

• Main Authors: Flores, Arturo C., Ng, Tiffany L., Roxas, Christian Carlo L.
• Format: text
• Language: English
• Published: Animo Repository 2002
• Subjects: Civil Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_honors/172
• Institution: De La Salle University

Abstract. Artificial Neural Networks of the backpropagation type was used to map the strength of High Strength Concrete given the design mix. Several ANN models were trained and simulated using 89 sets of data composed of the amount of cement, water, admixture, slag, silica fume, RHA, fine aggregates, coarse aggregates, fly ash, metakaolin, and the corresponding compressive strength of concrete at 28 days. The ANN models were validated through error metrics (root mean squared error, mean average error), minimum, mean, and maximum errors, sufficiency of number of training data, parametric studies, and statistical analysis (coefficient of regression). The results show that ANN can be used to trace the behavior of HSC and predict its strength.