Hybrid Structured Artificial Network For Compressive Strength Prediction Of HPC Concrete

Abstract

This study aims to develop a surrogate artificial neural network-based technique for predicting the compressive strength of concrete, which is the most significant factor in the life service of concrete and its durability in civil construction projects of civil engineering. For this goal, a data set for high performance concrete is gathered from the literature repository, which includes a different percentage of fly ash, silica fume, and superplasticizer in the mix designs. The data set is applied to train and validate the optimal structured artificial neural network optimized by innovative equilibrium optimization and particle swarm optimization algorithms. The results showed that the EOANN-I and PANN-I model with the R2 and RMSE values of 0.9889,1.771, 0.9881, and 1.8813, stood at the first and second stage of the most capable models in the prediction of HPC compressive strength. Although the PANN-I model’s complexity is lower than the EOANN-I model, the rate of its prediction accuracy covers its lack of complexity

Keywords

• hpc concrete
• compressive strength
• artificial neural network
• equilibrium optimization algorithm
• particle swarm optimization algorithm