Machine learning algorithms for evaluating concrete strength using marble powder

Abstract

Concrete is made with various industrial byproducts, and to check the effectiveness of the concrete constituents with waste marble powder, Artificial neural network, Random Forest, Support vector machines, and Adaptive neuro-fuzzy inference systems models were created. Six parameters were used to predict concrete compressive strength: cement, fine and coarse aggregate, water-to-cement ratio, waste marble powder, and curing days. The outcomes demonstrate that artificial neural networks are more accurate at predicting the compressive strength of concrete including waste marble powder. The ANN-obtained model has also undergone sensitivity analysis to determine input parameter effects on output. Following marble powder and curing days, the water-cement ratio has the greatest influence on predicting the compressive strength of concrete using a model based on an artificial neural network.