Investigation of the Flexural Strength of Steel Fiber Reinforced Geopolymer Concrete Using Machine Learning Techniques

Geopolymer composites have been studied and applied in many construction fields because of good performance in mechanical properties, workability as well as durability after long using time. The development of artificial intelligence proposes some methods which can predict and determine efficiently the performance of concrete structures through experimental data. The prediction and validation of the performance of fiber reinforced geopolymer composites by machine learning is evaluated in this research. The proposed models use artificial neural network ANN, deep neural network DNN, and 245 experimental datasets with 9 input variables. The validation of machine learning approaches shows the effectiveness of predictive methods with 90%, and 85% in ANN, and DNN respectively. The proposed models can be applied for designing the standard mix for steel fiber reinforced geopolymer concrete.